Residential Segregation in Comparative Perspective: Making Sense of Contextual Diversity (Cities and Society) [1 ed.] 1409418731, 9781409418733

Citation preview

Residential Segregation in Comparative Perspective

This set of city-case-studies across the globe—ranging from Tokyo to São Paolo, from Beijing to Istanbul—confronts different patterns and levels of inequality and segregation with different state and other contexts. Due to this focus it is a “must read” for all aiming at a fuller understanding of segregation. Through their colourful selections the authors convincingly show that context matters. They fill a gap in the literature. Sako Musterd, University of Amsterdam, The Netherlands Residential segregation is common to all cities but the nature, form and scale varies between cities and societies depending on economic and political structure, class and ethnic composition and level of development. With its wide range of case studies this valuable book breaks out of the assumptions of traditional Anglo-American contexts. It will be a “must read” for anyone interested in residential segregation and its variations worldwide. Chris Hamnett, King’s College London, UK

Cities and Society Series Series Editor: Chris Pickvance, Professor of Urban Studies, University of Kent, UK Cities and Society is a series disseminating high quality new research and scholarship which contribute to a sociological understanding of the city. The series promotes scholarly engagement with contemporary issues such as urban access to public and private services; urban governance; urban conflict and protest; residential segregation and its effects; urban infrastructure; privacy, sociability and lifestyles; the city and space; and the sustainable city. Other titles in the series Opportunities and Deprivation in the Urban South Poverty, Segregation and Social Networks in São Paulo Eduardo Cesar Leão Marques Beyond the Resources of Poverty Gecekondu Living in the Turkish Capital Sebnem Eroglu

Residential Segregation in Comparative Perspective Making Sense of Contextual Diversity

Edited by Thomas Maloutas Harokopio University and National Centre for Social Research, Greece Kuniko Fujita Michigan State University, USA

First published 2012 by Ashgate Publishing Published 2016 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN 711 Third Avenue, New York, NY 10017, USA Routledge is an imprint of the Taylor & Francis Group, an informa business Copyright © 2012 Thomas Maloutas and Kuniko Fujita Thomas Maloutas and Kuniko Fujita have asserted their right under the Copyright, Designs and Patents Act, 1988, to be identified as the editors of this work. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing in Publication Data Residential segregation in comparative perspective : making sense of contextual diversity. -- (Cities and society) 1. Discrimination in housing. 2. Marginality, Social. 3. Sociology, Urban. I. Series II. Maloutas, Thomas. III. Fujita, Kuniko. 363.5'1-dc23 Library of Congress Cataloging-in-Publication Data Maloutas, Thomas. Residential segregation in comparative perspective : making sense of contextual diversity / by Thomas Maloutas and Kuniko Fujita. p. cm. -- (Cities and society) Includes bibliographical references and index. ISBN 978-1-4094-1873-3 (hardback) 1. ethnic neighborhoods. 2. Cultural pluralism. 3. social classes. 4. SS, Urban. i. Fujita, Kuniko. ii. Title. HT221.M348 2012 305.8--dc23 2012004147 ISBN 9781409418733 (hbk) ISBN9781315605661 (ebk)

Contents List of Figures and Maps List of Tables   Notes on Contributors  

vii xi xv

1

Introduction: Residential Segregation in Context   Thomas Maloutas

2

Residential Income Inequality in Tokyo and Why it Does Not Translate into Class-based Segregation   Kuniko Fujita and Richard Child Hill

3

The Impact of Housing Tenure on Residential Segregation in Beijing, China   John R. Logan and Limei Li

4

Residential Segregation in an Unequal City: Why are There No Urban Ghettos in Hong Kong?   Ngai-ming Yip

5

A Portrait of Residential Differentiation in Taipei City (1980–2010)   Chia-Huang Wang and Chun-Hao Li

6

Residential Segregation and Social Structure in São Paulo: Continuity and Change since the 1990s   Eduardo Marques, Renata Bichir and Celi Scalon

7

Segregation, Social Mix and Public Policies in Paris   Edmond Préteceille

153

8

The Solidity of Urban Socio-spatial Structures in Copenhagen   Hans Thor Andersen

177

9

Residential Segregation in Budapest before and after Transition   Zoltán Kovács

197

10

The Limits of Segregation as an Expression of Socioeconomic Inequality: The Madrid Case   Marta Domínguez, Jesus Leal and Elena Martínez Goytre



1

37 69

89 111

135

217

Residential Segregation in Comparative Perspective

vi

11

Changing Dynamics of Residential Segregation in Istanbul   Tuna Taşan-Kok

237

12

Social Polarization and De-segregation in Athens   Thomas Maloutas, Vassilis Arapoglou, George Kandylis and John Sayas

257

13

Conclusion: Residential Segregation and Urban Theory   Kuniko Fujita

285

Index  

323

List of Figures and Maps Figures 1.1

A simplified mapping of the causal mechanism in the reproduction process of residential segregation  

12

2.1 2.2

40 Housing and commercial land value per square meter in Tokyo, 2010   Per capita and per household income inequality among Tokyo’s 49 local districts (23 central city wards plus 26 suburban cities), 1971–2005   44 2.3 Per capita income inequality among Tokyo’s 23 central city wards, 26 suburban cities and Tokyo (23 wards plus 26 cities), 1971–2005   46 2.4 Per capita income inequality among 23 central city wards, with and without Tokyo’s four core wards (Chiyoda, Chuo, Minato and Shibuya), 1971–2005   47 2.5 Correlation between per capita income and housing land value per square meter among Tokyo’s 49 local districts (23 central city wards and 26 suburban cities), 2005 48 2.6 Percent distribution of professional and managerial jobs, and production and laborer work among 2970 chomes of Tokyo Central City, 2005   52 2.7 Percent occupational mix of Tokyo Central City residents among higher, middle and lower income wards   55 2.8 Percent occupational mix of Tokyo Central City residents among higher, middle and lower income chos (based on average figures of 2 chos from each ward) 55 2.9 Percent occupational mix of Tokyo Central City residents among higher, middle and lower income chomes (based on average figures of 4 chomes from each ward)   56 2.10 Tax and income inequality among Tokyo’s 49 local districts (23 central city wards plus 26 suburban cities), 1970–2005   61 4.1 4.2

Distribution of household income in Hong Kong, 1996 and 2006   Distribution of selected variables by areas with and without public housing in Hong Kong  

103

5.1 5.2 5.3

A conceptual framework of the residential distribution patterns in Taipei City   Average yearly household income in Taipei City   Occupational distribution of residents in Taipei City  

112 118 119

6.1 6.2 6.3

Total income by social class, São Paulo Metropolitan Region, 1991/2000   Years of schooling by social classes, São Paulo Metropolitan Region, 1991/2000   International Index of Socioeconomic Occupational Status (ISEI) by social classes, São Paulo Metropolitan Region, 1991/2000   Distributions of the average income of the area/average metropolitan income in São Paulo, 1997/2007  

146 147

6.4

95

148 149

viii

8.1 8.2 8.3 8.4 8.5 9.1 9.2

Residential Segregation in Comparative Perspective

Occupational categories in jobs located in City of Copenhagen: a) 1981–1996 and b) 1997–2009   European countries at risk of poverty rate before and after social transfers, 2007   Type of ownership in multi-storey housing, City of Copenhagen, 1981–2010 (percent)   Number of residents in social housing sector in Copenhagen distributed by income deciles, 1986–2001   Number of residents in detached ownership housing in Copenhagen distributed at income deciles, 1986–2001   Transformation of Budapest Metropolitan Area after 1990   Ratio of university and college graduates in different urban zones of Budapest (1990, 2001)  

10.1 Distribution of Madrid’s total and foreign population by household size (2007)   12.1 Census tracts in the metropolitan area of Athens according to the percentage of Large employers and higher grade professionals and managers (ESeC 1) and Lower technical and routine occupations (ESeC 8 and 9) within the active population (2001)   12.2 Census tracts in the metropolitan area of Athens according to the percentage of immigrant and native population (2001)   12.3 Percentage of major occupational categories in the active population in Athens (1961–2001)   12.4 Percentage of households in the Ampelokipi neighborhood by social rank of head of household and floor of residence (1998)   12.5 Percentage distribution of Greek and foreign citizens in the Ampelokipi neighborhood by floor of residence (1998)  

188 190 192 192 193 208 211 233

261 262 263 272 273

Maps 2.1 2.2

Tokyo Metropolitan Government: Central city (23 wards) and suburbs (26 cities, 5 towns and 8 villages)   Tokyo’s 23 central city wards and four core wards  

39 41

3.1 3.2 3.3 3.4 3.5 3.6

Beijing’s central city, inner suburbs, and adjacent outer suburban zones   Distribution of migrants with urban registration in Beijing in 2000   Distribution of migrants with rural registration in Beijing in 2000   Distribution of people with tertiary education attainment in Beijing in 2000   Distribution of government employees in Beijing in 2000   Composition of housing tenure in Beijing in 2000, showing the tenure type with the largest share in each street office  

72 74 75 76 77

4.1

Distribution of the richest and poorest 10 percent of households Hong Kong, 2006   93

5.1

Metropolitan Taipei (Taipei City and Xinbei City)  

78

114

List of Figures and Maps

ix

5.2

Higher vs. non-higher education in Taipei City (1990 and 2000)  

6.1

Average International Index of Socioeconomic Occupational Status (ISEI) in the survey areas, São Paulo Metropolitan Region, 2000   141 Distribution of the high level professional class, São Paulo Metropolitan Region, 2000   142 Distribution of unskilled manual workers, São Paulo Metropolitan Region, 2000   143

6.2 6.3 7.1 7.2 8.1 8.2 8.3 8.4 8.5 9.1 9.2 9.3

Paris metropolis 1999—Cluster analysis of the socioeconomic profile of neighborhoods (IRIS)   Paris metropolis 1999—share of immigrants from Maghreb and South-Sahara Africa in the total population of neighborhoods (TRIRIS)  

115

160 164

Copenhagen’s urban structure (simplified)   Location of high-income persons in Metropolitan Copenhagen (9th and 10th deciles) 2007, location quotient   Relative distribution of unemployed persons in Metropolitan Copenhagen, 2007, location quotient   Spatial distribution of persons with academic degrees (master level) in Metropolitan Copenhagen, 2007, location quotient   Distribution of ethnic minorities in Metropolitan Copenhagen, location quotient  

179

Structure of Budapest   Percentage of university and college graduates in Budapest urban planning districts (1990)   Percentage of university and college graduates in Budapest urban planning districts (2001)  

199

10.1 Socio-spatial inequality in Madrid Metropolitan Area: Census tracts classified according to their distance from the mean socioeconomic profile (2001)   10.2 Proportional distribution of managers and high civil servants in Madrid Metropolitan Area (1991)   10.3 Proportional distribution of managers and high civil servants in Madrid Metropolitan Area (2001)  

181 183 184 185

202 207

227 230 231

11.1 Districts of Istanbul in the historical peninsula and the city center   11.2 Squatter areas (unauthorized settlements) with and without improvement plans   11.3 The distribution of new housing, renewal and transformation projects by HDA, metropolitan municipality and district municipalities  

245 246

12.1 Social types/clusters of residential areas in Athens, census tract level (2001)   12.2 Relative concentration of immigrants in the residential space of Athens by census tract (2001)  

264

251

267

This page has been left blank intentionally

List of Tables 2.1 2.2 2.3 2.4 2.5 2.6 2.7

Per capita and per household income inequality among Tokyo’s 49 local districts (23 central city wards plus 26 suburban cities), 1971–2005 (unit: 10,000 yen)   Per capita income inequality among Tokyo’s 23 central city wards and 26 suburban cities, 1971–2005 (unit: 10,000 yen)   Per capita income among Tokyo’s four central core wards and the rest of 23 central city wards, 1971–2005 (unit: 10,000 yen)   Occupational inequality among Tokyo Central City’s 23 wards, 877 chos, and 2,970 chomes, 2005   Tokyo Central City’s chomes with over 50 percent residents employed in professional, technical and managerial (P&M) jobs, and industry with highest percent P&M, 2005   Ten chomes with highest housing land prices by percent professional, technical and managerial (P&M) and percent employed in finance, insurance and real estate (FIRE) industries, 2005   Occupational mix of residents at Tokyo Central City’s ward, cho and chome levels, 2005  

43 45 46 51 53 53 54

3.1 3.2 3.3

Education and high-status occupations by district, 2000   73 Housing tenure composition in Beijing, 2000   77 Locational attainment model predicting distance (ln) from the city center (N = 9114)  82

4.1

Segregation patterns in Hong Kong 1996–2006 (education, occupation and employment)   Segregation patterns in Hong Kong 1996–2006 (income, family type and housing)  

4.2 5.1 5.2

Housing prices in Taipei City (2008–2010, thousand new Taiwan dollars and percentage)   Housing prices in New Taipei City (2008–2010, thousand new Taiwan dollars and percentage)  

97 99 117 117

6.1 6.2

Global Moran Index and Dissimilarity Index, São Paulo Metropolitan Region, 2000 EPG class distribution (%), São Paulo Metropolitan Region, 1991/2000  

140 144

7.1 7.2 7.3

Paris metropolis—active persons by occupation (CS)   Paris metropolis—segregation index for detailed occupations (CS)   Paris metropolis—dissimilarity index between immigrants and native French born in mainland France  

157 158

8.1

Occupation categories, index of segregation 2009  

189

162

xii

Residential Segregation in Comparative Perspective

8.2

Occupation categories, index of dissimilarity 2009  

189

9.1 9.2

Changing income inequalities in Hungary (1972–2004)   Gini coefficient of income distribution in Hungary by settlement categories (1987–2004)  

204

10.1 Proportional distribution of Socioeconomic Categories (CSE) in Madrid Region 1996–2010   10.2 Percentage salary increases by National Occupation Categories (CON) in Madrid (1995–2006)   10.3 Income inequality indices based on average household expenditure in Madrid (1990/91–2005)   10.4 Correlation coefficients of socioeconomic categories with the first factor of a principal components analysis of their distribution in Madrid’s census tracts (2001)   10.5 Segregation indices for socioeconomic categories in Madrid Metropolitan Area (1991 and 2001)   10.6 Segregation index by immigrant origin in Madrid Metropolitan Area (2001 and 2009)   11.1 Changing dynamics of residential segregation in Istanbul   11.2 The change of the occupational composition in Istanbul from 1980 to 2000   12.1 Social typology of residential space in Athens (2001), cluster composition in terms of major occupational groups (values above average in bold)   12.2 Percentage of the 10 major groups within the immigrant population in Athens (2001)   12.3 Indices of dissimilarity (ID) for occupational categories in the Athens Metropolitan Area; census tract level in 1991 and 2001   12.4 Percentage point difference of dissimilarity indices between occupational categories (1991–2001), positive values (in bold) denote the increase of spatial distance   12.5 Dissimilarity indices (ID) for ESeC classes in Athens (2001), total active population and Greek population   12.6 Dissimilarity indices for selected immigrant groups in the Athens Metropolitan Area (2001) in respect to the Greek population and to each other   13.1 Cities grouped by levels of residential segregation and class inequality and by the forces that intensify or counteract them  

205 220 221 223 226 228 232 240 250 265 266 268 269 270 271 292

To Hartmut

This page has been left blank intentionally

Notes on Contributors Hans Thor Andersen is Director of Research at the Danish Building Research Institute (SBi), Aalborg University. His main areas of interest are social geography, housing and urban politics [[email protected]]. Vassilis Arapoglou is Assistant Professor in Social Inequalities and Social Exclusion at the Department of Sociology—University of Crete, and tutor at the Greek Open University. His teaching and research focuses on critical analysis of urban social problems, migration, inequality and forms of belonging [[email protected]]. Renata Bichir is PhD in Political Science (Iesp/Uerj) and Coordinator of the Evaluation Department of the Ministry of Social Development and Fight against Hunger (MDS) [renatambichir@yahoo. com.br]. Richard Child Hill is an Emeritus Professor of Sociology at Michigan State University. He has taught and published widely in the fields of urban and industrial sociology, international political economy, and East Asian development. He recently co-edited Locating Neoliberalism in East Asia: Neoliberal Spaces in East Asian Developmental States [[email protected]]. Marta Domínguez is Associate Professor at Universidad Complutense de Madrid and Coordinator of the Master of Sociology of Population, Territory and Migration. Her areas of interest are segregation, spatial analysis methods, the city as a cultural reality, urban identity, and urban vulnerable groups [[email protected]]. Kuniko Fujita is a retired Professor of Sociology who has previously taught at Michigan State University, the National University of Singapore, and Hiroshima University. Fujita recently edited “Global Financial Crisis and State Regime Shift,” Environment and Planning A (43–2, 2011), and is currently editing the book, Cities and Financial Crisis: New Critical Urban Theory [fujitak@ msu.edu]. Elena Martínez Goytre, Sociologist, Researcher and PhD candidate at Universidad Complutense de Madrid focuses on recent urban changes in Spain. She has been working on segregation, gentrification, social composition in inner cities, public space, and housing especially related to the emancipation process of new generations [[email protected]]. George Kandylis is Researcher at the Institute of Urban and Rural Sociology at the National Centre for Social Research [[email protected]]. Zoltán Kovács is Professor in Human Geography at the University of Szeged, Hungary he is also scientific advisor at the Geographical Research Institute of the Hungarian Academy of Sciences in Budapest [[email protected]].

xvi

Residential Segregation in Comparative Perspective

Jesus Leal is full Professor at Universidad Complutense de Madrid, teaching Urban sociology and housing sociology. He is the director of a research group on society, Environment and Space. He is the author of several books and articles about housing and urban sociology [[email protected]]. Chun-Hao Li is Assistant Professor in the Department of Social and Policy Sciences at Yuan Ze University, Taiwan. As a sociologist and a demographer, he is interested in human migration, urban segregation, and community research. He is currently working on research projects which are concerned with acculturation and psychological adaptation of immigrant brides in Taiwan [chl@ saturn.yzu.edu.tw]. Limei Li is Associate Professor in the School of Social Development, East China Normal University, Shanghai [[email protected]]. John R. Logan is Professor of Sociology at Brown University, and Director of Brown’s research initiative on Spatial Structures in the Social Sciences. He is also Visiting Research Professor at Hong Kong University [[email protected]]. Thomas Maloutas is Professor of Social Geography at Harokopio University in Athens and Interim Director of the National Centre for Social Research [[email protected]]. Eduardo Marques is livre-docente Professor at the Political Science Department of the University of São Paulo and Researcher at the Center for Metropolitan Studies [[email protected]]. Edmond Préteceille is Senior Researcher Emeritus at Observatoire Sociologique du Changement, Sciences Po–CNRS. His present research is focused on the transformation of social structures, urban inequality and segregation, and social relations in large metropolises [edmond.preteceille@ sciences-po.fr]. John Sayas is Assistant Professor in the Department of Geography and Regional Planning, School of Rural and Surveying Engineering of the National Technical University of Athens [isayas@ central.ntua.gr]. Celi Scalon is full Professor at the Sociology Department of the Federal University of Rio de Janeiro [[email protected]]. Tuna Taşan-Kok is Associate Professor of City and Regional Planning. She teaches Human Geography at the University of Utrecht, Roosevelt Academy, and conducts research at Delft University of Technology, OTB Research Institute for the Built Environment in the Netherlands [[email protected]]. Chia-Huang Wang is Professor in the Department of Social and Policy Sciences at Yuan Ze University, Taiwan. He is interested in urban sociology, political economy of development, sociology of information and communication technology, and critical culture studies [wanghcia@ saturn.yzu.edu.tw]. Ngai Ming Yip, is Associate Professor at the Department of Public and Social Administration, City University of Hong Kong [[email protected]].

Chapter 1

Introduction: Residential Segregation in Context Thomas Maloutas

Introduction This book is about the great variety of patterns and trends of social and ethnic segregation in cities nested in different regions of the world. It is also about the limited impact this contextual variety has had on the dominant explanatory schemes in urban theory and about the shortcomings of the latter in making sense of contextually diverse forms of segregation. Its chapters challenge primarily the vision of the dual and polarized city as a fitting description of current socio-spatial divisions in large metropolitan areas around the world and its projection as their unavoidable future under the pressure of capitalist globalization. They challenge, in fact, the depictions and predictions about increasing segregation and spatial polarization founded on essentially monocausal explanations, such as the social polarization thesis (Sassen 1991), by drawing attention to outcomes and processes that are not in line with, and often contradict, theoretical expectations. By doing so this book brings to the fore the double contextual blindness of such theoretical approaches: blindness in terms of the contradicting empirical evidence from diverse contexts; and blindness due to their implicit attachment to specific contexts. Contextual blindness in the latter sense is not new. It is an issue with early approaches and tools of segregation research as well. Since, the major theoretical assumptions—old and new—about segregation were formulated in the US and, to a much lesser extent, in the UK, the focus of this book lies outside the Anglophone world, seeking to avoid the interpretative limitations and misconstructions resulting especially from universalizing the American experience. Residential segregation no longer attracts interest as an independent issue, but mainly as part of urban social changes related to the post-industrial metropolis and the globalization era. According to Hamnett (2001: 163–4) interest in segregation declined with David Harvey and radical geography and reappeared with William Julius Wilson and the underclass debate and, further, it shifted from segregation patterns to conceptions of duality in world/global cities. The strongest theoretical assumptions involving segregation are certainly related to the world/global city model (Friedman and Wolff 1982, Sassen 1991, Knox and Taylor 1995, and in more nuanced terms Mollenkoprf and Castells 1991, Fainstein et al. 1992) produced by global forces unleashed by neoliberal deregulation. The social polarization thesis (Sassen 1991) is probably the most direct claim about the relation between social and spatial trends: Social polarization is the assumed outcome of economic restructuring for global cities, which become the strategic spaces for global capitalist management; this role entails the rapid development of high-end producer services that generate high profile and highly paid jobs and attract a highly skilled workforce from all over the world. The growth of the upper occupational pole is complemented by the simultaneous growth of menial jobs related to the low level tasks in the expanding sector of producer services, but also in the service of the expanding occupational elite, while the loss of secure and averagely paid jobs in industry completes the polarization trend by depleting the middle of the social hierarchy. According to Sassen (1991: 251) social polarization leads also to spatial polarization: gentrification, supported

2

Residential Segregation in Comparative Perspective

by the housing demand for the new occupational elite, and the appropriation of prime space for corporate use, both lead to increased segregation for the lower social strata. The social polarization thesis endorses the perception of cities as increasingly socio-spatially divided under the changes brought about by globalization that pull away all stops and leave no margin for political intervention. It treats segregation as a simple and homogeneous negative social outcome deriving almost automatically from changes in the socioeconomic structure and does not adequately corroborate its theoretical claims by empirical evidence. I claim that both of these shortcomings are, partly at least, related to the contextual blindness of the polarization thesis, which assumes general validity in spite of the contextual attachments to the Anglophone world— and to US global cities in particular—it implicitly carries. The social polarization thesis has been criticized on many grounds: The lack of corroborating evidence for social polarization in par excellence global cities like London, Paris or Tokyo (Hamnett 1994, Préteceille 1995, Hill and Kim 2000, Fujita 2003, Hill and Fujita 2003); the inadequacy of duality as the essence of socio-spatial division that should be replaced by the more nuanced descriptions and assumption of the quartered or layered city (Marcuse 1989, 2002, Marcuse and van Kempen 2002a: 265–6); the neglected importance of politics and the state, with particular reference to the welfare state (Hamnett 1996, Musterd and Ostendorf 1998, Marcuse and van Kempen 2002, Musterd et al. 2006); its explanatory inadequacy for regional metropolises around the world etc. (Baum 1997, 1999, Wessel 2000, Walks 2001, Vaattovaara and Kortteinen 2003, Maloutas 2007a, Borel-Saladin and Crankshaw 2009). Other approaches to urban socio-spatial processes and outcomes under conditions of capitalist globalization put much more emphasis on contextual causality. Brenner and Theodore (2002: 353) stress the different pathways that lead to different forms of “actually existing neoliberalism” related to the contextual embeddedness of neoliberal restructuring projects within “a historically specific, ongoing, and internally contradictory process of market-driven sociospatial transformation (…).” Hill and Fujita (2003) insist on the nested structure of urban, national and regional systems that reduce the influence of global forces and contribute significantly in shaping socio-spatial outcomes. Following an institutional approach, adapted to his focus on European cities, Kazepov (2005: 6) stresses the open-ended and path-dependent character of socio-spatial outcomes within different contexts, stemming from processes configured as “a set of alternatives made of constraints and enablements within which individual (or collective) actors can or have to choose.” Swyngedouw et al. (2003) emphasize the local crafting of emblematic urban development projects producing a kind of unexpected expectedness in the outcome of the interaction between global forces and local factors. Without denying the existence and the importance of global forces that push toward increasing inequality and segregation, the contributions of this book try to illustrate that in cities around the world there are often alternative outcomes. These outcomes are significantly affected by targeted national and local policies in the North and West-European welfare states or the East Asian developmental states; they follow the dynamic of market forces in the transition economies of Eastern and central Europe; they appear as the unintended outcome of policies related to other issues in the clientelist and family-centered regimes in Southern Europe. They also appear related to private housing production structures, which in some cases are too weak to enhance division and sometimes so powerful and centralized that they tend to mitigate the dividing impact of their product in their own business interest. In most cases, these alternative outcomes are largely influenced by processes and structures originating long before the emergence of new global forces, like the local long-lasting social division patterns and the spatially uneven distribution of quality in the housing stock. Contextually varied situations offer different possibilities for policy intervention, and

Introduction

3

empirical findings show that policy impact in tandem with unintended consequences from policy and business decisions, and inertia in the reproduction of urban structures can seriously impede global forces from leading to “a new spatial order” of increasingly clear socio-spatial division that Marcuse and van Kempen (2002) have not identified across several cities either. Segregation is a context-bound concept. For this book, context is important in two ways: first, in the form of varied urban settings around the world involving multiple versions of segregation that are not amenable to simple and universal explanations regarding their formative processes, their patterns and their impact; second, as the context-bound, and therefore limited, “shared understanding of reality” (Kazepov 2005: 6) which derives from the binding of the concept of segregation to the context of the US metropolis of the first half of the twentieth century that has to be considered when the concept travels worldwide. “Context” is used here in a more mundane manner than in Wittgenstein’s or Frege’s philosophical elaborations concerning the (im)possibility of meaning or truth/falsity claims outside the (contextual) frames of propositions. It is mainly used to remind us that expected outcomes deriving from theoretical claims are often contradicted by outcomes whose understanding entails taking into account contingencies not included in theoretical models. Concepts and theories are always context-dependent and the degree of this dependency varies in relation to their specific object. Urban segregation is context-dependent in the sense that its patterns and social impact are determined by the combined effect of mechanisms and institutions involving the market, the state, civil society and the specific and durable shape of local socio-spatial realities. Theoretical models usually take into account part of this interrelation and, to a large extent, disregard the rest. The market is usually privileged as the focus of theoretical constructions with a particular focus on economic restructuring during the last decades. By “context” this book refers to the specific intertwining of four major spheres: (1) the economic sphere (exchange) that mainly focuses on labor market conditions and on market access to housing; (2) the state sphere (redistribution) that covers housing and public services allocation, and local regulation regimes; (3) the social sphere (reciprocity) that includes social and family networks, churches and other local voluntary organizations. “Context” also extends to (4) the specific and durable shape of local socio-spatial realities, i.e. built environments, social relations inscribed in property patterns, urban histories and ideologies. The three first derive from Polanyi’s (1944) modes of economic integration1 while the fourth involves the physical support of segregation processes and the social relations directly inscribed in it. This understanding of contextual elements is not fundamentally different from the ‘contingencies’ identified by Marcuse and van Kempen (2000a: 266) affecting the impact of global forces on socio-spatial urban forms, and from Hill and Fujita’s (2003) or Kazepov’s (2005: 6–7) elaboration of urban systems’ embeddedness in wider contexts of social, institutional and economic relations. There should be no question by now whether residential segregation can be adequately decontextualized and assumptions about it formulated in market—or state-related mono-causal terms or if a less de-contextualized plural causality should rather be adopted. There are two different, but interconnected, ways to proceed with the construction of such a causal plurality. The first is to elaborate on causal mechanisms and processes using a hypothetico-deductive approach; the second is inductive and could rely on building a large database relating contextual features to specific segregation processes and outcomes on the basis of a number of initial theoretical assumptions. This introduction, as well as the city chapters, are steps in both directions. 1  See the adaptation of Polanyi’s ideas in urban studies by Kesteloot and others (Meert et al. 1997, Kesteloot 1998, Kesteloot et al. 2006).

4

Residential Segregation in Comparative Perspective

Definition and Etymology Segregation indicates the spatial separation of two (or more) population groups; here this separation is understood as residential, but it may also refer to separation in schools, in the workplace, in transportation or in leisure activities. Segregation can vary from complete separation to completely even distribution of population groups in the spatial units of study areas. Highly segregated areas are those where the distribution of population groups is particularly uneven. Although in the recent literature there are attempts to re-focus segregation studies either in terms of effective life experiences (Schnell 2002) or in terms of a simultaneous layering of different activities and functions (Marcuse and van Kempen 2002a: 266) the focus remains on traditional understanding of residential segregation. The term originates from nineteenth-century genetics and refers to the separation of allelic genes that occurs during meiosis (Mendel’s 1st law). In the early twentieth century the Chicago School drew explanatory inspiration from analogies with the vegetable kingdom and segregation was adopted by human ecology as a metaphor for the residential separation of ethno-racial groups (Park 1936 [1957]). This metaphor subsequently became segregation’s dominant meaning. The definition of segregation in The Dictionary of Human Geography is very brief—“The residential separation of subgroups within a wider population” (Johnston et al. 1986: 424)— and is followed by references to the degree of segregation and to its measurement techniques using various segregation indices. However, as I will subsequently stress, despite its apparently simple definition, the social and political content of segregation becomes relatively ambiguous; and this is mainly due to the understanding of segregation as exclusively related to the lower social strata and as an unequivocally negative social condition disregarding the complex relations between spatial and social distance, especially across different contexts. Segregation is imbued with connotations—acquired through its long history as a social and political issue and a research object and practice—that continually add new meaning and make this concept rather imprecise.2 It is therefore imperative to start by elucidating how the definition of segregation is operationalized within the practice of segregation research. The Practice of Segregation Research: Choosing Population Groups and Measurement Techniques The simple definition of segregation leaves a number of important issues to be resolved through practical decisions, sometimes in ad hoc ways. Such decisions involve the choice of population groups who come under scrutiny or the methods that will be used to quantify the level of segregation and reveal its shape.3 The choices made in terms of these issues add further meaning and reshape the definition of segregation in ways that are not always explicit. In this book we consider mainly segregation research that deals with city-wide patterns and trends; therefore, we focus on quantitative and broad urban area research rather than on localized neighborhood studies that may be appropriate to dissect segregation processes, but often project out of proportion the extreme social condition of particular neighborhoods onto the cities they belong to.4 2  See Brun (1994) on the fluid meaning of segregation. 3  See Préteceille (2004) for a thorough presentation of these issues; this section is largely inspired from his text. 4  For a more comprehensive account of segregation research see Boal 1987, Hamnett 2001 and

Introduction

5

Population Groups and Spatial Units Theoretically, all kinds of population groups (ethnic, racial, social, age, …) can be the object of segregation. However, segregation research and literature have focused on groups whose spatial separation created a social and political problem, i.e. on those identified by race or ethnic origin and on social groups, mainly identified by occupational status or income. Racial and ethnic groups have been the primary object for the pioneering segregation studies in the United States, where they continue to constitute the main concern. This focus is related to the context of American cities in the early twentieth century, in which the legacy of the slavery regime and the very important immigrant inflow were regulated through institutionalized discrimination against specific racial and ethnic groups that involved, among other things, their residential segregation. Even though this situation has gradually changed after the Second World War with the high social mobility and desegregation of most immigrant groups, and the progressive abolition of discriminatory legislation, the long established segregation patterns along ethno-racial lines have not been fundamentally reshaped. This is especially striking for hypersegregated 5 metropolises, like Chicago or New York, where the index of segregation for African-Americans remained extremely high (over 0.806) until 2000, even though it had slightly decreased after 1980 (Logan et al. 2004). European cities, on the other hand, are much more homogeneous in terms of ethnic and racial composition (Kazepov 2005, Musterd 2005, van Kempen 2005, Musterd and van Kempen 2009) and have been so during most of the twentieth century; and those that were traditionally varied in terms of ethnicity—especially in Central and Eastern Europe—have usually become homogeneous as a result of wars and ethnic cleansing. Segregation studies in continental European cities, that started developing in the early post-war decades as an export product from the Anglophone world,7 focused on social class as the prime identifier for residential segregation. The UK and the continental European industrial core encouraged immigration toward their Fordist labor markets from former colonies and Southern Europe since the early post-war period. Outside these regions, ethnic and racial minority groups have substantially developed as an important component of cities’ populations during the last decades of the twentieth century, following the strong and lasting wave of international immigration, and only subsequently have they become an important item on the segregation research agenda. Focusing on ethno-racial rather than social segregation (and vice-versa) and on discriminatory rather than market mechanisms of segregation are options related to context, in the sense that research and policy attention is primarily turned to what constitutes a social and potentially a political issue. Within each one there are further options related to the specific designation of groups to be studied; these options are not free from theoretical and methodological assumptions, and are not devoid of consequences when transferred to different contexts. The focus, for instance, on the residential segregation of an oppressed Black minority by a White majority, frames segregation van Kempen 2002; See also Préteceille (Chapter 7 in this volume) and Kesteloot et al. (2006) for the impact of neighborhood focused research on the perception of segregation. 5  The high degree of segregation on five different dimensions (evenness, exposure, clustering, centralization and concentration) in several US cities was defined as hypersegregation by Massey and Denton (1989). The hypersegregated metropolis contains hyperghetto areas where the extreme segregation of mostly African-American population is coupled with the dismantling of local institutions and networks, representing a regression compared to the traditional Black ghetto (Wacquant 2008). 6  See the meaning of segregation indices in the following section on segregation measures. 7  See, for example, McElrath (1962) attempting a social area analysis of Rome in the early 1960s; see also Robson (1969) and a brief overview in Robinson (1998: 137–41).

6

Residential Segregation in Comparative Perspective

as the problematic condition of the former versus the “normality” or mainstream condition of the latter. Eventually, policies designed to redress segregation ills considered under this light are limited to devising ways of socially integrating the groups, or spaces, mostly victimized by segregation. Segregation becomes, thus, a problem at the margin, less visible as a process operating across urban society, which contributes to reproducing social inequality at all levels of the social hierarchy and throughout urban space. Following this contextually specific perception of segregation, the specter of the Black ghetto sets the political and research agenda about segregation even where there is no evidence to justify it (Marcuse and van Kempen 2002b, Kesteloot et al. 2006, Wacquant 2008). The choice of relevant social groups to investigate segregation in Europe and other parts of the world is usually much more complex than the reading of racial segregation, in the sense that racial division is much clearer—even visually—and the categories used for its registration are much simpler and fewer than those necessary to designate social hierarchy. The categories used to register social hierarchy are usually based on occupational positions, which are then aggregated into broader hierarchical classes. More detail can reveal particular category patterns and levels of segregation whose relevance, however, depends on the theoretical importance of such categories as class constructs and on the use of specific spatial scales. The detailed or aggregate grouping of social categories is bound to lead to different findings: for example, professionals and managers related to corporate activities may show quite dissimilar location patterns and segregation indices from the same categories occupied in the public sector (Préteceille, Chapter 7 in this volume). Distinguishing these different patterns would not be possible with the use of categorizations that aggregate the relevant occupations.8 Moreover, census variables are configured in relatively diverse ways in different countries rendering comparative research more complex. They often contain different types of information (as in the case of race which is registered in some countries but not in others) while they may also register the same information using a different protocol. Occupation, for instance, is used as a fundamental index of social hierarchy but with substantial national variations: most countries register occupational information using the ISCO standard (International Standard Classification of Occupations) of the ILO (International Labour Office); often this information is subsequently recoded into socioeconomic class categories according to more elaborate theoretical assumptions. Some countries use their own classification standards that may be quite distant from the standard provided by the ILO—France being the outstanding case with its catégories socioprofessionnelles that rely on different theoretical assumptions (Desrosières and Thévenot 1988). All these national differences in terms of availability and quality of data have an impact on both the analyses that can be carried out locally, and on the reliability and relevance of international comparisons. Choosing a particular model of grouping occupations or an alternative variable—like income or education level—to express social hierarchy is related to methodological and ideological preferences and may have an impact on research results.9 These choices are also related to context: 8  Such aggregations are the outcome either of practical considerations or theoretical choice: in the European Socioeconomic Classes (ESeC) for example—a classification of occupations inscribed in the Weberian tradition (Rose and Harrisson 2007)—the delimitation of categories depends exclusively on the employment relationship (Goldthorpe 2000) for which the distinction between employment in the public or private sector is not relevant. Such classifications are not context-free either: The ESeC are, for instance, much less suitable for Southern or Eastern Europe compared to Western or Northern Europe where they were devised (Maloutas 2007b). 9  The choice of occupational categories instead of income classes to investigate social segregation, frames the problem in terms of class relations (therefore considers important the position within the relations of production) rather than in terms of a mere indicator of individual placement within an internally undifferentiated

Introduction

7

the production of socioeconomic categories like those used in France would be rather astonishing in a country whose political tradition is characterized by class cooperation rather than class struggle. The focused interest of segregation research on lower socioeconomic strata and discriminated ethno-racial groups, even though in most cases it is the higher social groups that are the most segregated (White 1984: 156–8 and Préteceille 1993), seems to be a long-lasting feature of the perception of cities and their problems in countries like the UK, France, Germany and the US throughout the nineteenth and the first half of the twentieth century (Lees 1985). Limiting the focus of segregation research to the lower and discriminated groups is reducing its scope and exonerating broader mechanisms and groups whose choices are far more enabled than constrained and are in fact much more responsible for the constrained choices of others (Pahl 2001). Studying segregation as the uneven spatial distribution of all social groups in the city is therefore a prerequisite to understand socially or spatially more localized phenomena and trends. The choice of spatial units is much less related to context, but the size and nature of spatial entities used for segregation analysis are very important since different options can lead to different results. The optimum spatial units are neither too large nor too small. Large units have the disadvantage of hiding potential segregation pockets within their space; on the other hand, very small units—city blocks, for example—may not be relevant for groups’ effective isolation, since the spatial range of everyday activity and that of social networks largely exceed their confines. The optimum size of spatial units in segregation analysis is that of a neighborhood. In practical terms, the ideal size of spatial units for neighborhood segregation analysis is close to that of census tracts with a population close to 1,000 inhabitants. This should cover the largest part of the local web of social networks and most of the important local services. Moreover, spatial units should be of rather uniform size in order to avoid measurements in different parts of the study area that are in fact incompatible. In this sense, municipalities and other forms of local administrative units should be considered rather unfit for this type of analysis due to their variable size and population, and in spite of the convenience they usually offer in terms of data availability. Segregation Measures: Indices, Multivariate Analysis and Autocorrelation Models There are several ways to quantify segregation. The older and simpler ones consist of segregation indices that calculate segregation levels. The index of dissimilarity (ID) is the best known and most popular segregation index (Duncan and Duncan 1955; see also Robinson 1998, 257–60). It measures the dissimilarity between the spatial distributions of two groups within a study area. In case the second group is the whole population, the index measures the first group’s segregation level. The ID is easy to calculate: it is equal to half the sum of the absolute differences of the percentage of the group under scrutiny from the percentage of the whole population (or of another group) within each spatial unit of the study area. It is expressed as a percentage, with values ranging from 0 to 100 percent, and indicates the percentage of members of the first group that would need to relocate to different spatial units in order to obtain a similar distribution to the reference group. Other indices measure the degree of a group’s isolation or exposure by calculating the possibility of its members having encounters with members of other groups. Indices of isolation come also in percentage form and express the likelihood for a member of the first group to interact with members of another group (Coulter 1989: 149). continuum of socioeconomic hierarchy. However, the practical outcome in terms of identifying segregation patterns and levels may not differ substantially when using variable expressions of social hierarchy, especially if the investigation is not carried out using detailed sets of social categories.

8

Residential Segregation in Comparative Perspective

The indices of segregation have been devised and widely used in the US since the 1950s. They were adequate for measuring the segregation/isolation of the Black from the White population and, at the same time, were simple to understand and easy to communicate. Their simplicity, however, glosses over certain important aspects, like the spatial structure of segregation, as these measures are not affected, for instance, by the spatial aggregation or the dispersion of segregated units within the study area. Moreover, the comparability of segregation indices across cases is undermined to some extent by the effect of unit size—smaller units, and therefore spatially finer analysis, usually yield higher index values—as well as by the effect of group size, especially in respect to very small groups for which even a random distribution can produce significant levels of segregation (Robinson 1998: 260). The more elaborate the assumptions about segregation become, the harder that task faced by simple segregation indices.10 More elaborate measures of segregation, involving multivariate statistical techniques, are suitable to examine its spatial structure and to address more complex segregation aspects, like its multidimensionality (i.e. the different spatial forms and degrees of separation in respect to different groups of variables). Multivariate techniques synthesize large amounts of tabulated data and reveal their underlying structure (if there is one).11 The application of multivariate statistical techniques in segregation research developed under the name of factorial ecology and was inspired by social area analysis (Shevsky and Bell 1955) which described the expected shapes of segregation in respect to social rank, family status and ethnicity. The empiricist foundation of social area analysis led inevitably to crude Americanocentric generalizations. It is not a paradox that these techniques were mainly used in the US to illustrate this particular structure of multidimensionality—sometimes in comparison to other parts of the world where the analysis revealed fewer dimensions12—and much less in Europe, where they were mostly used to reveal the multidimensionality within the social rank dimension. 10  The investigation of the five dimensions of hypersegregation following Massey and Denton (1989) or of the four types of processes (assimilation, pluralism, segregation and polarization) related to ethnic segregation according to Boal (1999, cited in Johnston et al. 2002) needs at least the combined use of a host of different indices. 11  The synthesis is operated on the basis of co-variance patterns between the initial variables (the way that is that their values are distributed in the different spatial units) and takes the form of new variables called factors, components, dimensions or axes. If the initial dataset possesses an underlying structure, a relatively small number of such factors will account for a substantial part of the variance in the initial dataset, and in this sense they may be considered an adequate summary of the information contained in that dataset. Then, each of the few important factors is given content derived from that of the initial variables with which it is highly correlated. Having determined the content of factors, the analysis turns to the factor scores (or coordinates) of the spatial units that were used in the analysis, which position them on each factor. If, for instance, a factor stands for social hierarchy, factor scores will determine the relative position of spatial units in this hierarchy. They can thus be used to produce maps summarizing the relation between spatial units and factors and, therefore, to reveal the spatial structure produced by population groups’ distribution in residential areas according to each factor. Further use of factor scores can lead to the clustering of spatial units according to their scores on all (or on a selected number of) important factors in order to create a typology of residential spaces. This typology can be mapped and reveal the spatial structure produced synthetically from all retained factors, while the different clusters can be studied in terms of their specific population features (social, racial, demographic etc.) and of the change these features present over time. 12  See Abu Lughod’s (1969) analysis of Cairo, where the dimensions of social rank and family status appeared to be collapsed due to Egypt’s lagging position in the modernization process (family status— accounted for by family size—appeared highly correlated with social rank since polygamy was a privilege of wealthy men). Berry and Rees (1969 cited in Robinson 1998: 137) made similar observations for Indian cities.

Introduction

9

Factor and clustering techniques may be used to identify and map social and ethno-racial patterning of segregation and change over time, and they can be powerful tools if they serve theoretically informed inquiries. The fact that they usually lacked a solid connection with social theory, does not diminish their potential usefulness. On the other hand, they may be helpful in revealing the spatial structures characterizing the distribution of different population groups within a study area, but they do not take account of this structure in their algorithms (Sharre 1995). The factorial or clustering algorithms operate independently of contiguity or dispersion between similar types of spatial units (i.e. of spatial autocorrelation). There are, however, other techniques that combine—through the use of GIS—measures of spatial autocorrelation (like Moran’s I13) with multivariate statistical techniques of segregation analysis (Wong 1993, Robinson 1998: 270–80). The main problem with multivariate techniques, and even more so with those that are sensitive to spatial autocorrelation, is that their output is not easily grasped by those unfamiliar with their logic; confusion may easily infiltrate between the characterization of factors and spaces on the basis of their relative composition in respect to cities’ averages and their actual composition which is usually much less distinctive. Segregation and Contextual Difference The simple definition of segregation as the spatial separation of population groups has permitted this concept to appear decontextualized, i.e. sufficiently abstract and therefore of general validity. However, behind this simple definition lies a concept that is rather halfway decontextualized,14 since it continues to carry numerous contextual attachments that remain implicit to a large extent, and is imbued with the connotations it has acquired through decades of segregation research and urban policy-making. Thus, even though segregation appears, by definition, to be a simple notion, its comparative study presents several problems illustrated by the practical difficulties of measurement across contexts. In fact, a fundamental problem stems directly from its very definition, which leads to considering segregation much more as an outcome than as a process. This means that attention is focused on degrees of segregation rather than on its formative processes and on its effective impact. The simple definition implies, in a sense, that the content of segregation is self-evident; but the reality of segregation across contexts proves to be much more complex and less prone to immediately meaningful comparison. It is much more meaningful to compare segregation as a process; to examine, that is, whether it is generated by similar causes and whether similar mechanisms are mediating its development, whether it has similar social consequences, whether segregation is increasing or decreasing across contexts, whether it is framed in similar ways as a political problem and whether similar policies are devised to regulate it. As an outcome, segregation is inevitably much more embedded in the particularities of local contexts—i.e. much more dependent on contextual causality—while as a process, it may be more relevantly related to global, intercontextual, forces and tendencies.

13 Moran’s I is an index based on the Pearson correlation coefficient, which relates the variance of a variable with its spatial autocovariance (i.e. with the contiguous or dispersed form of its distribution). Values higher than zero indicate similarity, regionalization, smoothness and clustering, while values below zero indicate independence, randomness and dissimilarity (Robinson 1998: 276–7). 14  See Maloutas (2012) for a more elaborate argumentation along this line regarding gentrification.

10

Residential Segregation in Comparative Perspective

Segregation as a Process Invoking contextual diversity may seem sometimes to be an excuse for avoiding comparison, abstraction, generalization and theory. Every city is unique in its detail and, at some level, noncomparable with any other. In this sense, our approach to contextuality, even if limited to state (redistribution)-market (exchange)-civil society (reciprocity) combinations and their interrelation with durable urban structures, histories and ideologies, creates a very large variety of potential contextual situations that cannot be productive if context is to be used as a differentiating parameter in a systematic way. Ideally, we would be aiming to construct a typology of contexts that could be related, more or less, to different forms and degrees of segregation in a theoretically meaningful way. The data we have from the 11 cities of this volume and from previous works can bring us closer to such a target, although they are far from sufficient. Fujita (see conclusions to this volume) elaborates on varieties of capitalism that could be used as basic referents for contextual difference and then tries to fit bottom-up groupings of the 11 cities included in this volume following the logic of these varieties. Here, I try to map residential segregation as a process—i.e. not only as an outcome—and to pinpoint the parts of the process where contextual variation seems particularly important. Residential segregation is an outcome and, at the same time, a part of the process that reproduces inequality and discrimination in capitalist societies. Urban segregation is older than capitalism in the sense that cities were socially partitioned since their first appearance (Marcuse 2002a) and they remained so in the short lived state socialist societies (Pickvance 2002). The distinctive feature of segregation under capitalism is that—like inequality—it is mainly a product of economic mechanisms rather than the outcome of other forms of social violence. As an outcome, segregation is fed by economic inequality and discrimination and shaped by their filtering through space-related mechanisms and structures and, especially, by the shifting and sorting of housing allocation processes. This shifting and sorting is mainly operated by the housing market on the basis of households’ unequal ability to pay; it is usually complemented by administrative allocation that may counteract or reinforce the effect of market mechanisms, as well as introduce different dimensions to the shifting and sorting process (i.e. enforcement of discriminating—or anti-discriminating—rules against groups defined otherwise than by economic condition). Individual choices of residential location—more or less constrained or enabled within different contexts—are systematically aggregated into unequal social and ethno-racial distributions within cities (van Kempen 2002: 46–7). The established reproduction of residential segregation and the neighborhood or area effects it generates—i.e. the effects not attributable to personal and household characteristics, but the additional spatial effects related to the social composition of residential areas, to their intrinsic qualities (e.g. pollution level, quality of services) and their comparative status (Atkinson and Kintrea 2001, Buck 2001, Lupton 2003)—make it part of the structures and mechanisms that reproduce urban social inequalities. Class segregation is fundamentally a market driven process theoretically starting as economic inequality produced in the labor market and transformed to segregation through the housing market. Ethno-racial segregation is mediated by economic inequality—with ethno-racial difference being translated to ethno-racial hierarchy in both the labor and housing markets—and also derives directly from discriminatory rules and practices15 in housing allocation.

15  Such rules and practices do not need to be official and explicit as in the case of “redlining” or “blockbusting” (Knox and Pinch 2006: 140–43).

Introduction

11

The state may be intervening at one or more parts of segregation processes and in one way or another. State intervention may alter considerably the outcome of the segregation process, opposing or reinforcing the expected outcomes of market mechanisms and of discriminatory predispositions and arrangements. This alteration may affect the form and degree of segregation and/or its impact on people’s lives. Expected outcomes may also be altered by the impact of reciprocity networks that civil society builds independently of the state and the market, due to the capacity of these networks to alleviate conditions of poverty and deprivation or to reinforce situations of privilege and advantage. Furthermore, the spatial fixation of these networks—especially those of the lower social strata—may impede the otherwise free shifting and sorting by the market. Finally, the expected outcomes of market mechanisms in terms of socio-spatial separation are influenced by the characteristics of recent and inherited durable urban structures (Vaughan and Arbaci 2011), including the social relations they carry (e.g. property rights or tenure patterns). It is important to stress, therefore, that segregation may be more or less severe—as several segregation indices can indicate—but this severity is not an unequivocal index of its social role. On the one hand, spatial distance cannot be a reliable measure for social distance because proximity does not necessarily reduce social distance (Chamboredon and Lemaire 1970) and, on the other, public policies can reduce the impact of spatial distance on the reproduction of inequality and discrimination, and mitigate its socially negative impact. Schematically the processes that lead to residential segregation and to its reproduction may be depicted as a sequence of systemic causal relations mediated and modified by institutional interference and local historical inertia under the form of durable urban structures, social relations and ideologies (Figure 1.1). Institutional interference has been abstracted from the schematic depiction of the reproduction process of residential segregation since it may occur at all its levels. The first part of the causal mechanism identifies the labor market and the unequally accumulated wealth that jointly generate and reproduce economic inequality as an expected—and often intended—outcome in capitalist societies. The degree of inequality is often assumed to be related to the exposure or insertion of local labor markets to globalized economic processes, even though the exposure to such processes is not necessarily positively correlated with the degree of inequality produced by the labor market. Theoretical schemes, like the global city thesis (Friedman and Wolff 1982, Sassen 1991) give some insight regarding the unequal pressure that urban labor markets experience toward more inequality, and especially polarization, depending on their position in the global urban hierarchy. It is widely contested, however, that a polarized social structure, with a distribution shaped like an hourglass rather than an onion (Pahl 1988), is causally related to segregation, since simple inequality can also fuel segregation under the same conditions (i.e. when housing allocation mechanisms do not operate against free shifting and sorting) (Hamnett and Cross 1998, van Kempen 2002: 49). Ethno-racial discrimination is less directly related to systemic causal relations even though capitalism tends to transform any form of difference into hierarchy and inequality. Consolidated ethno-racial hierarchies embody the translation of accumulated discriminations into inequality, while the dominant ideological frames pertaining to alterity affect discrimination norms and arrangements. The extent of ethno-racial diversity and the degree of ethno-racial discrimination are both important contextual characteristics, as well as institutional intervention to impose and solidify discrimination, or to reduce it through positive discrimination measures. Finally, different modes of social regulation, ranging from extreme laissez-faire that favors the growth of inequalities in the name of economic efficiency, to situations where restraining inequality is an explicit social and political objective, can frame the functioning of the labor market and the reproduction of accumulated wealth in significantly different ways and produce very

12

Figure 1.1

Residential Segregation in Comparative Perspective

A simplified mapping of the causal mechanism in the reproduction process of residential segregation

diverse outcomes in terms of inequality. Esping-Andersen’s (1990) types of welfare capitalism have provided a theoretically informed insight on how to group cases following the degree of decommodification that different welfare regimes attain in their social regulation. However, this does not do justice to the redistributive functions of the Asian developmental state, which affects the labor market much more directly (e.g. by imposing a compressed wage system) than the European welfare state that mainly tries to redress inequalities produced by a less constrained labor market. Exploring further the distinction between stronger and weaker government—suggested by van Kempen and Murie (2005: 384)—or among varieties of capitalism (Hill and Fujita 2003, Kazepov 2005: 11 and Fujita, conclusion to this volume) may be productive options. In this first part of the causal mechanism, therefore, the important contextual characteristic is the degree of income inequality, generated by the labor market, and ethno-racial discrimination both potentially mitigated by state regulation, leading to the distinction between rather equal and rather unequal cities. In the second part of the mechanism, inequality and discrimination are translated into social and ethno-racial segregation through a series of space related structures and mechanisms that act

Introduction

13

as filters in the process of housing allocation, i.e. as specific enabling and constraining conditions for the housing choice of households. In a capitalist society housing allocation may justifiably be expected to function by market criteria. Therefore, it is also expected that the housing market will shift and sort the unequal individuals and households into unequal places. Institutional intervention at this level has been important and multifaceted in different contexts, ranging from minimum intervention to full-blown policies of social housing, greatly affecting the shape of segregation patterns although not always in the desired way. North European welfare states with a universalistic distribution of services show much lower levels of segregation compared to the residual welfare regimes of the US and increasingly of the UK (Domburg-De Rooij and Musterd 2002). The production of a large volume of social housing projects in the developed welfare societies of Western and Northern Europe has contributed to tackling the housing problem on the short to medium term in the early post-war decades, and to keeping segregation at rather low levels (van Kempen (2002a) but has also, unintendedly, led to producing the main physical support of residential segregation in the long term (Marcuse and van Kempen 2002a, Andersen 2004). History endows cities with different building stocks. The diversity in the quality of the housing stock—and therefore the socially diversified access to it—as well as the spatial distribution of this diversity are important parameters for shaping residential segregation. Equally important is the diversity in the spatial distribution of other attributes of the housing stock, such as unit size, that may exclude certain types of households from certain areas or, on the contrary, compel them to choose among very few areas offering the required size. The legal and social relations that tie people to housing and neighborhoods (in the form of property rights, tenure and social networks) are also of great importance. It may be expected that cities with a relatively uniform distribution of housing stock in terms of quality and sizes, with strong property rights, high rates of homeownership and extensive solidarity networks based on family or common origin ties usually entail low levels of residential mobility and inhibit shifting and sorting, and segregation. In the third part of the mechanism, residential segregation feeds back inequality and discrimination through the positive or negative area effects produced by the diversified social composition of neighborhoods. Institutional intervention may mitigate or exacerbate these effects by confronting segregation in different ways (e.g. policies for social mixing or policies guaranteeing similar quality of schools across neighborhoods). A typology of contexts where residential segregation is reproduced could be constructed using a number of empirical measures in order to control whether certain types of context are systematically related to different degrees of segregation.16 16  A measure for income inequality—a gini coefficient of income distribution for example—would be useful to appraise the degree of inequality as a fundamental indicator of unevenness characterizing the demand for housing. The percentage of the dominant ethno-racial group in the city’s population and the difference in average salaries for the same jobs between the dominant and the other main groups may account for the importance and magnitude of discrimination in the labor market. The impact of the exposure of the local labor market to globalized economic processes on the housing market and segregation could be measured by the percentage, on the one hand, of foreign-born managers and higher grade professionals and, on the other, of foreign born routine job holders. The effectiveness of social regulation can be measured by the percentage point reduction of poverty rate after social transfers, and the general level of redistributive justice by an indicator like child poverty rate. The percentage of people in decommodified tenure and a gini coefficient for the distribution of house prices and rents could add some insight into the role of housing allocation mechanisms. Networks and attachment to place are much harder to approach by easily accessible data; the ties to housing and neighborhoods can be roughly assessed through the rate of residential mobility, while the average age of leaving the parental house can serve as an indicator of the impact of social networks—especially

14

Residential Segregation in Comparative Perspective

Such measures are certainly not sufficient to address segregation issues in a particular city; they are certainly not the most pertinent, but they are relatively accessible approximations of the relevant information, and combined they could help to broadly map contextual diversity in a rather organized and tangible manner. In fact, such measures can lead to a city typology that takes into account the breadth of inequality created by the labor market, the importance of ethnoracial discrimination, the extent of the mitigating effect of the welfare state on inequality, the degree of commodification of housing provision and the degree of unevenness in the distribution of housing prices, the level of residential mobility and the possible existence of solidarity networks. Eventually, this typology can be related to the intensity of social and ethno-racial segregation and their tendency to increase or decrease. This means that the typology could distinguish between more or less equal cities, with a more or less pronounced element of ethno-racial discrimination, with a more or less effective welfare—or otherwise interventionist—state, with more or less commodified and unequal housing provision, with more or less residential mobility, with indications for the existence of solidarity networks and, finally, relate the different types with more or less intense and increasing or decreasing social and ethno-racial segregation. Multiple measures would permit to include the time dimension and assess the dynamics of the typology. The assumption is that more unequal and discriminating cities, within less developed welfare states, with a more commodified housing provision, with high residential mobility and with less solidarity networks are expected to be highly and increasingly segregated. And since capitalist globalization drives toward more inequality, less welfare state and more commodification of housing, it may reasonably be expected that segregation should be found to be on an increasing trend. In this book we do not have information from a sufficiently large number of cities to create such a typology; the different contextual dimensions we have identified will serve, however, as broad guidelines to assess the profiles of segregation in the 11 city chapters as a first step to such a typology and a complement to other works that have investigated the causality of segregation levels and trends. Most of the works challenging mono-causal interpretations of segregation focus on differences and difficulties of comparison between the US and Western/ Northern Europe, both because of their significant differences and of the fact that they are the most researched areas in this respect. In the following I briefly discuss the contextually embedded nature of segregation, first, as a path dependent process whose specificity is illustrated by the comparison between segregation in the US and the (West) European metropolis; second, as an ideological and political issue whose contextual variegation entails significantly different approaches to its nature as a problem and to the ways it is addressed; and, third, as a contextually diversified social impact in terms of neighborhood effects. Distinct Segregation Paths: The US and Western Europe The comparative understanding of urban segregation on the two sides of the Atlantic is that on the one side segregation is high and mainly ethno-racial while on the other it is substantially lower and mainly social (i.e. based on occupation and income). These differences are probably responsible for the small number and the rather unsatisfactory character of comparisons between segregation in US and European cities. Sako Musterd (2005) family networks—on residential mobility and, potentially, segregation. Finally, the dependent variables—i.e. the level and trend of both social and ethno-racial segregation—can be assessed through segregation indices while the existence (or not) of explicit anti-segregation policies can function as a dummy variable.

Introduction

15

attempted to gather comparative evidence in order to illustrate the expected differences between the two, but found the European evidence wanting and fragmented due to different standards and categorizations among European countries or simply to missing information. Lack of adequate data is always a problem, but, in this case, it is probably not the most important one. Simple segregation measures are not sufficient to convey the different nature of segregation processes in Europe and the US. In this sense, comparative readings of dissimilarity indices in Chicago and Paris, for example, may not be very meaningful if essential contextual information is missing; and this is not simply a question of difference regarding the choice of segregated groups or the size of spatial units. Segregation in the modern metropolis developed initially on the patterns of socio-spatial division inherited by pre-modern cities. The great leap forward in the development of segregation occurred when rapid urban growth stimulated spatial expansion under the form of suburbanization. This pattern—clearly depicted in Burgess’s model and in its subsequent modifications by Hoyt and others (Timms 1971: 211–29, Badcock 1984: 8–10, Knox and Pinch 2006: 161–3)—characterized initially, and mainly, the Anglo-American world, where the elites massively opted for suburban residence, especially after the First World War, in response to the rapid growth of manufacturing activities and the concentration of working class groups around the core of industrial cities (Fishman 1987). The mechanisms through which suburban residential space was produced (large developments corresponding to specific segments of the housing market) resulted in making suburbanization the par excellence process generating socially homogeneous spaces and, therefore, an important mechanism of segregation. Marcuse (2002: 27) characterized the evolution of the suburbs as “the sharpest representation of the increasing division of urban space.” In the US, at least during the first half of the 20th century, this mechanism worked in tandem with the racist regulation of residential space producing a social hierarchy of spaces within a clearer and more severe separation founded on ethno-racial discrimination and ghetto formation, especially for the African-American population in downgraded inner-city areas with the sharpest dividing lines created where the lines of race and class overlapped (Marcuse 2002a). At the same time, the melting-pot side of the American dream was working for most immigrant groups that were less racially distinct from the White majority than African-Americans. The relatively rapid socio-spatial assimilation of immigrant groups rendered the segregation of the Black population even more severe, and indicated that social mobility was immediately leading to the decrease of segregation. Decrease in social distance meant decrease in spatial distance as well. It was in those conditions that segregation was coined as a term for urban sociology. Its simple definition reflected these clear-cut spatial outcomes and gave the impression of a full correspondence between social and spatial distance. Within that context, Robert Park and the Chicago School were impelled to consider segregation and spatial distance a direct and adequate measure for social distance and inequality. It is because geography, occupation, and all the other factors which determine the distribution of population determine so irresistibly and fatally the place, the group, and the associates with whom each one of us is bound to live that spatial relations come to have, for the study of society and human nature, the importance which they do. It is because social relations are so frequently and so inevitably correlated with spatial relations; because physical distances so frequently are, or seem to be, the indexes of social distances, that statistics have any significance whatever for sociology. And this is true, finally, because it is only as social and psychical facts can be reduced to, or correlated with, spatial facts that they can be measured at all. (Park 1916 [1957]: 177)

16

Residential Segregation in Comparative Perspective

European cities, on the other hand, have not developed a similarly clear ethno-racial and social segregation pattern. In terms of ethno-racial composition, most of them were dominated—until recently at least—by one ethnic group, while the specific weight of outcast groups—like travelers— remained relatively marginal wherever they were present. The phenomena of ghettoization and white flight have not shaped, therefore, their socio-spatial structure, even though ghettoes were first instituted in the European continent (Wacquant 1997). In terms of social segregation, welfare policies have effectively opposed the development of high levels of spatial division. The second important parameter differentiating the pattern of segregation in continental European cities is that suburbanization has not been an equally strong generator of socially homogenous space. The relatively belated suburban expansion and the much greater reliance on public transport systems and spatial planning has produced less suburban sprawl and less social homogeneity in suburban spaces, while the attachment of ruling classes to central areas in major cities like Paris, Vienna or Madrid gave substantially less propensity to suburban expansion. As a result, cities in Europe remain more compact than their American counterparts; and since suburbanization was much less the choice of social elites who did not abandon city centers, they— as most cities around the world—were led, according to Timms (1971) and Leontidou (1990), to an ‘inverse Burgess model’. Moreover, cities in Europe, compared to those in the New World, rarely developed ex nihilio; their development patterns were usually grafted onto urban tissues inherited from the pre-industrial city whose spatial division was characterized, to a large extent, by the occupational maze corresponding to the spatial organization of the guild system (Sjoberg 1960, Vance 1971) rather than by clear segregation lines. In this sense, the definition of segregation corresponding to the context of the booming and racially divided American metropolis of the mid-west in the early twentieth century, has not been directly relevant for European cities. The same applies to measurement tools (especially segregation indices) and to research agendas of the Chicago School and the early post-war segregation research, which ignored politics and the state (Pahl 1975: 236–40, van Kempen and Murie 2009: 378) developed on the basis of this contextual model. Indices of segregation and factorial ecologies of European cities have never revealed clear images of shifting and sorting directly attributable to discrimination practices and unleashed market forces. They rather revealed complex and mitigated situations involving the interplay of inherited, socially mixed urban structures, of social structures and networks with increased spatial embeddedness and low residential mobility and, often, of policies directly checking the development of segregation. Thus, a staggering segregation index of 0.80 for African Americans in Chicago creates a clear, powerful and to some extent self-explanatory image, compared to a relatively modest 0.35 for high status professionals in Paris, Madrid or Athens, which reveals that the degree of isolation of the highest and most segregated occupational categories in these cities is nowhere near that of racially discriminated groups in the US. The significance of segregation as a major issue and a measure of urban social inequality has diminished for the Anglo-American metropolis in recent decades since urban socio-spatial change progressively depended less on expansion (suburbanization) and more on the ‘new urban frontier’ (Smith 1996) of gentrification. Gentrification is the dominant way of remodeling central urban areas affected by deindustrialization. It is leading to the attraction of more profitable land uses, higher status residents and the displacement of working class and marginal groups following neoliberal urban policies focused on commodification and competitiveness. Urban segregation and gentrification are particular spatial manifestations of ethno-racial and social inequality since they are both related to specific processes within particular contexts. They have respectively emerged out of the extremely clear ethno-racial division based on racial discrimination in the modern

Introduction

17

American metropolis of the early twentieth century, and of the massive socio-spatial remodeling of city centers in the post-industrial metropolis of the Anglophone world. It is no wonder then that they are most clearly expressed in those contexts rather than anywhere else in the world (Maloutas 2012). Gentrification has not left segregation unaffected. As a process of (re)appropriation of central city spaces by middle and upper middle-classes at the expense of working class groups it may be assumed to lead eventually to more segregation, but, for some time at least, it increases the social mix in gentrified areas and may reduce segregation indices.17 Moreover, gentrification changes the scale of segregation by diversifying spaces at the micro-level. Further complexity stems from the uneven use of local services (e.g. schools)18 by different social groups, which develops as segregation is reduced, and blurs the correspondence between residential spaces and social profiles. Therefore, the simple and seemingly universally applicable definition of segregation—as well as of gentrification19—and the tools to measure it come bundled with implicit contextual assumptions (Butler 2007: 162) that may distort the analysis of socio-spatial inequality within different contexts. Perceptions of Segregation and Policy Responses The perception of segregation as a social issue has been constructed in strikingly different ways following the diverse paths and content it acquired in different contexts, its contextually differentiated effects and mainly the ideological substratum on which it stands. Again, the difference observed between the two sides of the Atlantic is informative. On the American side the perception of segregation is founded on the dominance of economic liberalism, personal merit and on a very high rate of residential mobility.20 From the era of Chicago School’s natural areas the high rates of social and residential mobility led to intense shifting and sorting on the housing market, and the relation of people to places became increasingly fluid and temporary. People and places formed two distinct, even though interrelated, hierarchies: places according to quality, accessible by people according to merit. As the market became dominant in the allocation of housing, there was a widespread belief—illustrating its ideological dominance— 17  In most cities of the advanced capitalist world gentrification runs parallel to the declining numbers of the working class, and to some extent at least, it leads to the replacement rather than displacement of the working class (see Hamnett 2003 about London, and Slater 2006 for a different approach). In both cases, however, the effect of first or second stage gentrification (Smith 1987)—i.e. of the phase in which pioneer groups, like artists, start living in an area and setting a trend to be followed by a second phase when affluent groups corroborate the trend—is the mitigation of working class segregation in gentrified areas, which serves to legitimate gentrification policies. Otherwise, gentrification increases overall working class spatial retrenchment even in the case of replacement: the spatial distribution of the working class becomes more uneven as its vanishing part is not replaced by middle class groups in working class strongholds and in gentrified or socially mixed areas at the same rate. 18  See van Zanten (2001) and Butler et al. (2006) on strategies of school choice, especially concerning middle-class groups. 19  Clark (2005: 258) has proposed a simplified definition of gentrification which should serve its universal use: “a change in the population of land-users such that the new users are of a higher socio-economic status than the previous users, together with an associated reinvestment of fixed capital.” 20  Comparative data show that cities of the New World were the champions of residential mobility in the 1980s with annual rates between 15 percent and 20 percent. European cities were much lower with rates around 5 and 10 percent (Knox and Pinch 2006: 252). More recent figures for Southern Europe show rates clearly below the European average (Allen et al. 2004).

18

Residential Segregation in Comparative Perspective

that where people live reflects where they deserve to live and hence that whatever residential segregation exists should not be considered a social problem. Racial discrimination, however, has been distorting the image of the meritocratic system obstructing potentially deserving African Americans (and others) from accessing better places, while the cracks of the market produced barriers to deserving poor (Whites as well). Following the same ideological doctrine, segregation becomes an equal opportunity problem limited to the lower social strata. Policies devised to confront it aim at providing opportunities to escape from bad areas rather than to improve them, and people may be moved to less segregated residential areas or to non-segregated schools. Policies like Moving to Opportunity,21 the HOPE program [www. thehopeprogram.org/] or school bussing are within such a conceptual and contextual frame. The tendency to dissociate, in policy terms, the fates of people and places in the US should certainly be related to the contradictory coexistence of a long history of racial discrimination—that flagrantly obstructed access to the land and housing markets for a substantial part of the population— with the high rates of social and spatial mobility for the numerous others that participated in the American dream. Thus, an important difference between the US and European constructions of segregation as a social and political problem consists of the still paramount presence of racialized segregation in the former. This begged for the liberalization of residential mobility for the racial groups victimized by discrimination, and made their unrestricted participation in the housing market an obvious improvement over normative, or otherwise imposed, discriminatory residential space allocation on the basis of racial hierarchy (Massey and Denton 1993). At some point, the free movement of individuals for residential location anywhere they could afford became at the same time a recommendation of economic liberalism and a progressive claim for the civil rights movement. However, this liberalization of residential mobility, combined with urban structures inherited from a long period of racial discrimination and to the impact of economic restructuring, has led according to Wilson (1987: 49–56) to further segregation of the African-American poor in increasingly jobless and socially disorganized inner-city ghettos following the flight of Black middle class and working class households. In Europe, on the contrary, concern for segregation developed regarding the negative impact of the freely relocating individuals and households through the mechanisms of land and housing markets that produce an uneven spatial distribution of social groups and, at the same time, uneven living conditions and life prospects in different localities. The major policy response in Western and Northern Europe has been the extensive investment in the social housing sector that, for some decades at least, has opposed segregation, especially wherever it was addressed to a wide social range of beneficiaries in a rather oecumenical welfare state spirit. The perception of segregation in the European city is substantially different and the life itineraries of people are much more tied to places, regardless of whether they become attached to them or feel entrapped. This is practically expressed by much lower residential mobility and is mainly founded on the comparatively reduced ideological influence of economic liberalism during long periods. Questions of residential area quality are constitutive parts of social equality in the French republican ethos or of the socialist tradition22 and social rights in Scandinavian welfare 21  MTO is a pilot project in the US whose rationale is to move people from downgraded social housing projects and control how they fare in less disadvantaged surroundings (Goering and Feins 2003, Orr et al. 2003). As a pilot program it had rather limited size and impact, while its basic procedures in terms of choice of households to be supported, and of the fate of those left behind, are questionable. According to Lupton (2003) such a policy rationale would be out of context in the UK. 22  See, for example, Halbwachs’s attention to the improvement of working class areas in the municipal socialism spirit (Topalov 2001).

Introduction

19

societies and, therefore, segregated areas are a problem to which organized society must provide answers. Socially mixed residential areas have resulted from policies founded on strong welfare states in Western and Northern Europe (Musterd and Ostendorf 1998, van Kempen 2002a, Musterd et al. 2006). Häussermann (2005) claims that the interventionist welfare state originates from the autonomy of European bourgeoisies in the nineteenth century and represents the main legacy of the European city. Scandinavian cities have been accustomed to the regulation of both labor and housing markets in ways that used to avoid segregation.23 Such framings of segregation tend to devise policies targeted to places apart from people and were, partly at least, the grounds on which area based policies were developed as a way to combat segregation in several countries around Europe (Burgers and Vranken 2003, cited in Musterd and Murie 2006). In certain countries and namely in France, the UK and Netherlands, there is considerable emphasis on anti-segregation policies, related to social issues and strong mobilizations considered to emanate from negative neighborhood effects. However, the emphasis on area based policies and, particularly, on social mixing in a receding welfare state can be associated with policies that are displacing the focus from social to spatial issues, are legitimating different objectives—like gentrification (Lees 2008)—and may eventually lead to increased segregation. Contextual diversity in terms of the ways segregation is perceived is also an issue within Europe. In Southern European cities, for example, segregation has not been until lately on the political agenda. Relatively low segregation indices, infrequent social unrest related to segregation, family centered social organization and very low residential mobility are probably part of the explanation. In the family centered welfare regimes of this region, people’s fates are even more tied to their place of residence than in Western or Northern Europe, but not due to increased public responsibility and policies. On the contrary, public intervention is much less developed and legitimated, and it is expressed in less direct ways, with less public housing among other things (Allen et al. 2004). Families cater for their weakest members’ needs and, since family networks have to coalesce in space in order to be effective, the resulting reduced residential mobility tends also to reduce the visibility of segregation as a social problem. However, even though ideological frames in Europe have been different from the US, the increasing dominance of neoliberalism in the last three decades has produced approaches and remedies to segregation that follow the American way, i.e. they promote the spatial redistribution of poverty and consider gentrification an effective way to improve social mixing (Ostendorf 2002, Lees 2008). The receding welfare states of Western and Northern Europe, the collapse of state socialism and the progressive decline of clientelism in Southern Europe have weakened—in different ways—the defenses against increasing inequality and its spatial expressions. Segregation Impact: Contextually Different Neighborhood Effects Contextual parameters that affect the ways segregation is constructed as a social and political problem also affect its impact and the solutions devised to combat it in terms of anti-segregation policies. Segregation is generally considered as an important issue due to its assumed impact on living conditions and on chances of social mobility. There has been a substantial growth in the literature addressing the impact of segregation, i.e. the neighborhood or area effect. This literature 23  There are claims, however, that even though policies of the Danish welfare state prevent social polarization, recent housing policies have increased segregation (Andersen 2004 and Chapter 8 in this volume); segregation appears also to be an issue for Sweden (see Holgersson et al. (2010) on Gothenburg) and Finland (Vaattovaara and Kortteinen 2003).

20

Residential Segregation in Comparative Perspective

has mainly been developed in the US (Ellen and Turner 1997) focusing on social isolation and ghetto culture; on the lack of role-models, related to the absence of successful middle class groups; on forms of social capital that constrain rather than enable social mobility; on poor quality of services and reduced access to new jobs (Atkinson and Kintrea 2001: 2278). The central issue is whether there are specific spatial effects on peoples’ lives and life prospects “over and above non-spatial categories such as gender and class (…)” (Atkinson and Kintrea 2001: 2277). These additional effects may originate from the different socio-demographic composition of neighborhoods, from their intrinsic quality—e.g. the quality of their environment or of the locally provided services—and from neighborhoods’ comparative status, ranging from privileged to stigmatized (Buck 2001). Does living in an area of concentrated poverty raise the chances of not finding a job or of not doing well at school compared to someone equally poor living in a mixed or a middle class neighborhood, and to what extent? Is there a linear relation between area effects and the intrinsic characteristics of neighborhoods or are there thresholds after which things change dramatically? Even though such questions about neighborhood effects can be formulated rather clearly, the research design for their empirical investigation is quite complicated because it is difficult to disentangle the complex ways in which individuals interact with neighborhoods (Buck 2001, Lupton 2003). The question of neighborhood effects is further complicated by the fact that they may refer to different spatial scales, they may be negative or positive and they are not necessarily the same for different class categories. According to Gordon and Monastiriotis (2006) neighborhood effects in education performance in the UK appear more important as a middle class advantage than as a disadvantage of working class groups. Research from the UK (Atkinson and Kintrea 2001, Buck 2001, Buck and Gordon 2004) and Netherlands (Ostendorf et al. 2001, Ostendorf 2002) reveals a relatively low, but significant level of neighborhood effects compared to individual/household characteristics. Musterd and Murie (2006) found effects of varying magnitude from a number of European city neighborhoods that were not always what was expected according to the local welfare regime; these effects were considered important—even though not fundamentally important for people’s lives—and not necessarily either positive or negative. Proving the existence and importance of neighborhood effects does not seem necessary in order to convince policy makers in the UK, France or Netherlands, as they seem already convinced of their existence and for the need to develop area based policies or initiatives (Lupton 2003). This conviction stems from the fear—substantially echoed by the media—that poverty in Europe is getting Americanized, with increased ghettoization and racialization. Musterd and Ostendorf (1998a: 6–7) criticized the paradoxical media suggestion to fight Americanization in Europe the American way, i.e. by converging to a workfare system and cutting welfare benefits. To some extent, however, these paradoxical media suggestions have been implemented by several European countries in the meantime, and the current sovereign debt crisis in Southern Europe imposes discipline and punishment along this line. The neighborhood effects literature is unevenly developed geographically, and this partly reflects the unevenness of these effects in different contexts. Enforced and strict spatial isolation, as in the excluded black ghetto, obviously reduces opportunities for social mobility to a much higher degree (Wilson 1987, Massey and Denton 1993) than spatial separation in comparatively low segregation environments and relatively evenly serviced residential areas, as in Dutch cities. In the latter, neighborhood effects may be found to be of considerably less importance for social mobility than the personal/household characteristics of the relatively isolated and deprived groups (Ostendorf et al. 2001, Musterd et al. 2003). Neighborhood effects in Southern European cities can be expected to be somewhere in the middle due to the contradictory influence of, on the one

Introduction

21

hand, the absence of highly segregated areas and groups and, on the other, the relatively poor and unevenly distributed social services. Overview of the 11 Cities: Segregation in Context The profile of segregation patterns, levels and trends for the 11 cities discussed in this book is very diverse. The same applies to these cities’ contextual features. Each one combines a different degree of inequality and discrimination, with strong or weak state intervention, with more or less socio-spatially dividing house allocation mechanisms, with strong or weak social networks, and forms a contextual profile which corresponds to segregation patterns, levels and trends that cannot be directly derived from any of these contextual parameters in isolation. Tokyo, in spite of its magnitude and importance for global financial and other circuits and Sassen’s contentions about its socio-spatial polarization, is a rather equal city with highly mixed residential areas. Fujita and Hill describe the unequal distribution of income among the city’s spatial units, which, however, is not accompanied by a similarly unequal spatial distribution of occupational categories. This paradox is explained by the outstanding functional primacy of its four central wards, in respect to the rest of the units in the central city as well as the suburbs, that were the object of intense investment leading to two real estate bubbles in the last 35 years. The owner-residents of these wards have gained much during the rising price periods, and lost much during the falls. What comes out of Fujita and Hill’s analysis is that the unequal spatial distribution of income in Tokyo is not a product of the labor market, but of the rocketing prices of landed assets in the four central wards that boosted the income of the local owner-residents and pushed others to different areas. Otherwise, the city has inherited a socially mixed urban structure since the end of feudalism, and has been reproducing its social structure in rather egalitarian patterns following the developmental state model. The latter involves features like the compressed wage system, which keeps low wage differentials even between managerial and production jobs; the corporate community ethos, which leads corporations to behave in a very protective way toward their employees, mitigating the imperatives of profit making and stakeholders’ interests; the long-term investment against the short-term profit making approach, which is coupled with a collectivist spirit in terms of accomplishments and rewards that blurs the limits between public and private; the highly redistributive tax policies that prevent wealthy communities from using their richer tax bases selfishly; and finally, the egalitarian education system and the non-residual character of public housing that promotes social mix and is present in all parts of the city—even in the four most exclusive central wards. The developmental state model checks in fact the inegalitarian impact and the spatial shifting and sorting of liberally regulated capitalism, leading to similar outcomes with the European welfare state. The difference is that the latter intervenes mainly through redistribution to redress the impact of the labor market, which is left more or less free to operate on market principles, while the former imposes stricter norms on the labor market and its inequality generating mechanisms. Both the welfare state and the developmental state are receding, but their regulating effects are still considerable and bear witness to the continuing power of politics at different levels. Segregation in Beijing, following Logan and Li, seems still mainly influenced by the previous decades of socialist regulation. Social inequality—measured by broad education and occupation categories—is growing, but the spatial separation of these categories remains quite limited. This is because although different forms of housing tenure correspond to different forms of building

22

Residential Segregation in Comparative Perspective

stock with quite distinct spatial patterns, and to different processes of production and allocation, they correspond to only moderately different social groups of potential occupiers. The growing commodification of housing, illustrated by the impressive increase of owner-occupation to 89 percent in 2010 from 20 percent within 30 years, has been dominated by the subsidized process of selling public housing units to sitting tenants, which means long established social segregation patterns have remained unchanged, especially since the new homeowners do not have any particular reason to move. Ethno-racial discrimination is traditionally important in China, but it applies to a relatively small part of the urban population and affects negatively only those with non-urban residence status (see also Chaolin and Kesteloot 2002). Hong Kong presents, following Yip, the unexpected combination of extremely high levels of income inequality and rather low levels of segregation, especially between middle and lower occupational and income groups for reasons connected mainly with planning and housing policies. With a recent colonial past that largely predefines the enclaves of the rich, and rather limited space that has to be densely built up, Hong Kong has to spatially accommodate a highly unequal social structure deriving from its position and great dynamism within global and regional financial and other key economic processes. In the last two or three decades the city has shifted very rapidly from manufacturing to services and its occupational structure was affected accordingly losing skilled and semi-skilled workers and gaining higher-end jobs (managers and professionals). Even though occupational polarization has not developed—since middle status jobs in the services have also increased substantially (Borel-Saladin and Crankshaw 2009)—income inequality, as well as polarization, have grown, especially since the mid 1990s. However, the high level of inequality and polarization does not translate into segregation, except for the richest decile, whose dissimilarity index is high and increasing. For the rest, segregation indices are relatively low and decreasing. This seems to be the outcome of planning and housing policies, with the very high specific weight of public housing affecting the private sector in a decisive way. Public housing is not only quantitatively important, of relatively good quality and attractively priced; it has also been targeted on a much broader clientele than the neediest groups, leading to reduced segregation that becomes more sustainable following the disincentives to homeowners—who purchased their house from public authorities—to resell it in the free market. On the other hand, the scale of housing operations is extremely large and the conditions usually imposed on developers are such that the latter have to diversify their housing supply in order to guarantee their investment. As in Beijing, the population is much more segregated by tenure than by socioeconomic characteristics; however, tenure categories are internally very diversified and this mitigates the importance of tenure segregation in socioeconomic terms. In contrast with Copenhagen and other Scandinavian cities, where effective policies for income equality are no longer accompanied by effectively egalitarian spatial policies (Andersen 2004 and Chapter 8 in this volume), Hong Kong applies antisegregation policies—although they are not termed and intended as such. Finally, as in Athens, Madrid and other Southern European cities (Arbaci 2007, 2008, Arbaci and Malheiros 2010) the low degree of segregation may not be a cause for celebration, as it does not preclude deprivation for the vulnerable groups even if they are not spatially distanced from the rest. Taipei shows a low level of residential segregation in terms of class, while ethno-racial segregation is rather insignificant due to the homogeneous constitution of its population. This virtual absence of working class segregation in particular should be attributed, according to Wang and Li, to the workings of the developmental state that produced this effect through its industrial, and mainly through its planning and housing policies during the last 30 years. Local policies and conditions have led to weak segregation in a path dependent way, and the influence of global forces

Introduction

23

seems rather reduced as well as the explanatory power of theoretical schemes stressing the effect of globalization on socio-spatial outcomes. There is a broad differentiation between Taipei City and its periphery (Xinbei City) with a higher concentration of upper social positions in the former. At the same time, failures and weaknesses of the housing and planning systems have permitted land speculation and housing price increases that have affected certain quarters of Taipei City and its immediate periphery, putting them out of reach for lower socioeconomic groups. However, the defects of planning and housing policies have unintendedly induced lower levels of segregation by not producing massive social housing projects concentrated in particular areas and leaving large numbers of the city’s population to be accommodated with alternative solutions or moving to the periphery and to neighboring cities. São Paulo is a very unequal and segregated city with a long tradition of inequality and segregation. According to Marques et al. its Centro Expandido has always been, and continues to be, the area of the higher social categories, while the peripheries—near as well as distant—are where the lower status groups are located. This dichotomy is not only spatial, but denotes also the strong historical unequal division in access to jobs, goods and services. The center/periphery dichotomy has been fed by the intense internal migration of the 1960s and 1970s, when industry was developing and those seeking working class jobs found residence in irregular settlements produced by private developers without state regulation and infrastructure, and with poor access to public services (see also Schiffer 2002). The poor were hardly becoming part of the city in a period otherwise characterized by rapid economic growth. The city’s occupational structure is characterized by a rather small top (low percentage of higher categories) and a much bigger bottom; the trend in the 1990s was relative growth at the top and stability at the bottom, therefore without any particular polarization tendency. In terms of income, however, inequality has been growing and polarization cannot be ruled out. At the same time, democratic change brought a better distribution of services, and schooling has increased and benefited almost all categories, especially the lower ones. Increased economic inequality, however, has not been translated to increased segregation. The strong center/periphery segregation division has been preserved with a decrease of homogeneity within each part, especially in the peripheries where the lower middle-classes have been increasing their presence. Another trend that reinforces the mosaic pattern is the location of gated communities near poor neighborhoods leading to a more complex combination of spatial proximity and social distance. Overall, the high level of inequality and segregation in São Paulo cannot be attributed to global forces, whose influence affected the city mainly after the 1980s. Inequality and segregation are persistent features of Brazilian cities on which globalization has had a certain effect together, however, with opposite effects related to the country’s democratization. Paris, according to Préteceille, is a comparatively unequal city in a not so unequal country due to its appeal for the very rich—natives and foreigners—and its position in global financial and other business networks. Residential segregation is, however, less marked than portrayed in the media and expected by certain researchers, even though the city is clearly divided into socially diversified spaces following decades and even centuries of upper categories’ residential choices, and urban planning. Social segregation is mainly defined by the opposition between the location of the higher categories of business managers and professionals and that of the working class; an opposition which is mitigated, however, by the existence of very substantial socially intermediate and mixed spaces that comprise almost half of the city’s population. The importance of these mixed spaces is concomitant with the growth of several middle social categories in the service sector that also contradicts the dual city model.

24

Residential Segregation in Comparative Perspective

On the other hand, the exclusive focus on the problematic areas of the banlieues glosses over the fact that the most segregated categories are those of the rich, especially the professionals and managers in the private sector; who are not only the most segregated, but whose spatial isolation increased in the 1990s, contrary to the decrease for most other social categories. At the same time, other high occupational categories—like teachers, professors, artists and journalists—are getting less segregated following different location patterns. It may be true that spaces at the extremities of the social hierarchy get respectively richer and poorer, but the exclusive focus on them (and usually just on the poorer segment) leaves most of the city fabric out of the picture. Ethno-racial segregation has become stronger than social segregation in Paris. However, it remains relatively moderate compared to Chicago or New York, with 70 percent of immigrants living in areas where the natives dominate. The unequal spatial distribution of ethno-racial groups cannot be entirely explained by their members’ class affiliation; the rest of the explanation should be attributed to various forms of discrimination and to practical necessities that bring immigrants of the same groups closer, especially when they are first established in the city. Préteceille concludes that we need to avoid fascination with the social and spatial extremes as well as with mono-causal explanations based on globalization trends. The global forces that affect the labor and housing markets do not necessarily produce the outcomes theoretically expected in terms of segregation since the outcomes are filtered by the inertia of the urban structure and the enabling or constraining effect of related policies. Urban policies in Paris have changed several times over the past decades until they became openly anti-segregation in the 1990s and, even though they have remained far from fully implemented, they have definitely had an impact on outcomes. This impact should not be taken at face value: social housing estates, for example, depending on their location and quality, may either promote or oppose segregation. Andersen argues that Copenhagen is certainly outstanding for its low level of social inequality, which is the outcome of the very high Danish employment rate and high wages that prevent poverty among the employed, and of the Scandinavian Social Democratic welfare state that protects effectively all individuals and households with no income from work. The current social structure derives from a growth period led by the service sector after the deindustrialization and economic restructuring of the 1980s. An important component of this structure are the immigrants who represent 22 percent of the city’s population, even though half of them originate from developed economy countries. Segregation, however, can be developed even when inequality is relatively reduced and Copenhagen—as well as other cities in the same welfare regime, e.g. Gothenburg (Holgersson et al. 2010)—is socio-spatially divided along lines inherited from the nineteenth century and steadily reproduced since. In comparative terms, segregation in Copenhagen remains limited and, as in most cases, the highest occupational or income categories are the ones that are most distant from the rest. Two mechanisms that induce socio-spatial change are depicted by Andersen: the first is the breaking of the traditional life cycle in many western cities, that involved suburban living for young households, leading to smoother age segregation patterns—i.e. less clear concentration of young couples with children in the suburbs returning at a later stage in the center—but also to gentrification pressure.24 The second is the “paradoxical impact of welfarism”: on the one hand, vulnerable groups are relieved of the burden of housing deprivation, but on the other, they become spatially isolated in residualized public housing estates due to the flight of other groups and their own increased presence. Housing policies seem to have lost their egalitarian effect

24  See also Marcuse and van Kempen (2000: 11–12) who argue that these changes increase the complexity of spatial divisions.

Introduction

25

as well as education services which—even though distributed equitably—have much less effect than they used to in bridging inequalities in social mobility. Segregation in Budapest appears extremely path dependent according to Kovács. Following the city’s initial rapid growth in the late nineteenth century, which endowed the city center with good quality buildings and rather low segregation, there developed the dominant segregation pattern between a privileged center and a deprived periphery as a result of the massive settlement of rural migrants in the latter. Segregation was again remodeled after the Second World War under the socialist regime that attacked both income inequalities and socio-spatial separation by policies that downgraded the center and created modern housing projects in the periphery. Inequalities and segregation started to grow again after the relaxation of socialist regulation in the late 1960s and more markedly after the collapse of state socialism (see also Ladányi 2002). The current East/ West division of the city, with particular affluent enclaves within both the center and the periphery, developed as a result of reforms both in the labor and housing markets. The liberalization of the former, accompanied by unrestricted openness to foreign investment, led to the rapid growth of income inequality. The liberalization of the housing market was decisive in boosting segregation as it was implemented in conditions of housing shortage and in the presence of considerable demand for upscale housing from foreign citizens. Thus, the subsidized privatization of public housing turned out to be an unequally profitable operation; those occupying higher value units and having the required resources were given the opportunity to gain much more than the rest. The re-valorization of properties in the center, the degradation of outdated and poorly maintained public housing projects in the periphery and the growing demand for quality housing accelerated the shifting and sorting, while the loose regulation of the market permitted the development of new processes—like “organised gentrification” by municipalities, “white flight” from social housing, filtering-down in certain areas of the center and mushrooming of “gated communities”—all of which have also boosted segregation tendencies. Madrid is a large metropolis of the European South, very well inserted in the globalized corporate networks and having experienced substantial economic growth from the late 1990s to the eve of the current crisis in 2007. According to Domingez et al. economic growth was accompanied by the rapid growth of higher occupational categories as well as of certain intermediate ones and by the decrease of those at the bottom of the occupational hierarchy. The range of income inequality increased with those with higher incomes increasing their distance from the rest. However, the group of those with minimum earnings has decreased. Thus, neither the occupational nor the income hierarchies have become polarized. In fact, income has increased for top categories and lower ones, while for the broad upper and upper-middle category of professionals it remained rather stable, partly due to their own over-fast growth—leading to considerable unemployment and precariousness—as a consequence of rapid intergenerational social mobility. This mobility has depleted lower occupational positions that were filled to a large extent by immigrants whose services became consequently more in demand. During the same period the importance of accumulated wealth has increased in respect to income from work: income from real estate and business grew much faster than income from salaries, pensions or unemployment benefits from 1994 to 2006. Under these circumstances, with housing prices increasing very fast and mortgages being extended ever longer, it makes a big difference—in terms of income—whether a homeowner has finished paying for his/her house or not. These social and economic trends were accompanied by demographic diversification, especially in terms of household structure that complicated even further the spatial impact of socioeconomic changes. Madrid seems to offer plenty of evidence that segregation is not a very pertinent indicator of urban social inequality. On the one hand, increased inequality has not been followed by

26

Residential Segregation in Comparative Perspective

increased, but rather by decreased segregation and, on the other, the new immigrant population is characterized mainly by housing deprivation rather than segregation. The clear division between a bourgeois center and a working class periphery in the 1950s and 1960s has been complemented by a broad North-South division and, although they both remain clear, there are internal diversification processes of different sorts that reduce segregation without leading necessarily to reduced social distance. Thus, the higher occupational categories, who have the higher segregation indices, have also presented the sharpest decrease in segregation levels. This was the result of their numeric growth and their expansion to residential areas where they were not previously present, through processes of embourgeoisement or gentrification. Another process that, temporarily at least, reduces segregation also takes place in lower status peripheral areas where the upwardly mobile offspring of working class parents remain in new improved housing projects in order to continue benefiting from family solidarity networks. Finally, there is a similar desegregation impact from immigrant housing strategies of overcrowding rented, or of subletting part of owner-occupied accommodation in intermediate or higher status areas due to the scarcity of adequate housing supply in traditional working class areas. Reduced segregation, therefore, should not be taken at face value since it glosses over different forms of inequality including, mainly, a substantial degree of housing deprivation for the immigrant population (see also Arapoglou 2006, Arbaci 2007, 2008, Maloutas 2007a, Arbaci and Malheiros 2010). In Istanbul, according to Taşan-Kok, the importance of ethnic diversity for segregation is not related to the city’s present ethno-racial profile—apart from the Kurdish minority perhaps—but to a rather distant past when non-Muslim groups lived quite separately from the city’s Turkish population. Ethnic cleansing in the southern Balkans after the First World War deprived the cities of the region of their multi-ethnic character. With the departure of non-Muslim groups, the privileged areas they were occupying became progressively derelict and inhabited by marginalized groups of poor rural migrants. However, the architecturally interesting building stock that survived many decades of dereliction and their privileged location has helped turn them to gentrifiable areas when social, political and economic conditions changed. During the same period of national consolidation, and in parallel with the dereliction of these areas, the city experienced large waves of rural migration that led to peripheral squatter settlements following family and common origin networks that enabled their settlement, survival and integration. Although poverty was a main feature of these areas, they have also been areas of social mobility, partly based on profit from the land and housing market dynamic and politically monitored by the clientelist legalization of illegal settlements in exchange for votes. The dynamic of these areas changed when Turkey became economically and politically outward-turned, and Istanbul became increasingly socially diversified attracting both unskilled and highly skilled labor. Some of these areas were progressively redeveloped targeting middle and upper-middle social groups, while lower-income social groups have often been displaced in the process. More recently, neoliberal policies have lowered the protectionist barriers of the previous period and brought foreign investors into land and housing development. Combined with the dual labor market and the residual welfare state, these policies deepened further the shifting and sorting of unequal social groups and the formation of socially very diverse spaces—like gated communities for the super-rich and deprived settlements for excluded groups—often close to one another. These processes have led to a broad segregation pattern in which higher income groups are mainly situated in central areas along the coast, lower income groups in the periphery, and middle income groups in between (see also Güvenç and Iştk 2002). Athens, following Maloutas et al., is a rather large regional metropolis that, comparatively, is neither particularly unequal nor intensely segregated in class or ethno-racial terms. This seems to be the unintended consequence of the combined influence of a relatively non-polarized occupational

Introduction

27

structure with a long history of high social mobility in the post-war decades as well as of the rather limited integration of the city into the high-end of the global labor market and the ensuing limited presence of an international corporate elite exercising pressure on the higher end of the housing market. Reduced segregation is also related to the dominance of two housing provision systems that had an ambivalent impact on both class and ethno-racial segregation. These systems developed as parts of the family centered welfare model that, as in the rest of Southern Europe, has grown to depend on family solidarity networks, reducing both residential mobility and segregation. They were also constitutive parts of the clientelist/populist political system that relied on defending both high social mobility rates and massive access to homeownership for its reproduction, often at the expense of the free function of market mechanisms that could put them under threat. However, residential segregation is a reality in Athens with the socially most extreme spaces becoming even more homogeneous and, as in most other cities, with the higher social categories more separated spatially than the lower ones. This is due to the gradual reversal of the dominant segregation divide between a bourgeois center and a working class periphery with the formation of extensive middle and upper-middle class suburbs in the eastern periphery that have since the late 1970s progressively become the city’s most homogeneous residential spaces. At the same time, traditional working class suburbs have become more socially mixed—as in Madrid (Leal 2004)— following the spatial ‘entrapment’ of endogenous social mobility, i.e. the fact that the upwardly mobile next generation avoids moving to a better residential address in order to preserve the advantages from participating in kinship networks (Maloutas 2004). On the whole, desegregation was the trend for all major occupational categories in the 1990s, except the shrinking and aging skilled industrial workers who seem to be increasingly confined to their traditional strongholds. Immigration, on the other hand, has not boosted segregation in Athens. Due to the location of the available and affordable housing stock in the densely populated areas around the city center, the presence of immigrants has reduced occupational segregation as large numbers of migrants holding routine jobs were mixed with native middle and upper-middle occupational groups. Reduced neighborhood segregation in Athens coexists, however, with other forms of sociospatial separation, like ‘vertical segregation’ (the systematic class and ethno-racial division by floor of residence in densely built areas around the center) or school segregation, which seems to be a strategy of middle-class households to overcome what they perceive as negative effects of reduced residential segregation. Immigrants, on the other hand, may not be highly segregated, but nonetheless suffer significant housing deprivation. Overall, residential segregation may be of relatively low intensity in Athens, but it is accompanied by real barriers to effective social mixing and contact that create social distance in spite of spatial proximity and, lately, even conflict. Summing Up Most of the chapters included in this volume depict segregation as a complex process that usually contradicts the assumptions of the polarization thesis and the dual/divided city imagery. With our focus on contextual causality—and having defined context as specific articulations of market, state, civil society and durable urban structures including the social relations, practices and ideologies they carry—the contributions of this volume bring evidence from around the world about the importance of such articulations for the shape, intensity and social impact of residential segregation. We can summarize this evidence as follows:

28

Residential Segregation in Comparative Perspective

1. We cannot assume that global forces induce unequivocally a high and increasing level of residential segregation in metropolitan centers around the world. Even though global forces contribute to increasing inequality, they do not immediately affect residential segregation; their impact is mediated by a host of contextual parameters and, usually, this leads to variable outcomes. Capitalist globalization theoretically leads to increased inequality, and sometimes to social polarization and segregation. However, in some cases the growth of inequality is mitigated or even reversed—as in Copenhagen—and in many more—like Paris, Tokyo or Madrid—inequality does not lead to polarization; but, even where there is growth of inequality and polarization, increased segregation does not automatically follow. On the contrary, in almost all of the 11 cities segregation has been decreasing for most occupational and ethno-racial categories. 2. Important socio-spatial dichotomies appear inherited from the past rather than the product of forces related to global economic restructuring. Such sharp dichotomies are observed in some of the cities—mainly in São Paulo and Istanbul—but they were formed well before the recent period of economic restructuring. Moreover, all the cities discussed in the book are more or less clearly divided between areas of different status, with those for the rich having been established as such for many decades, if not for centuries. These divisions, produced under quite different circumstances from current conditions, testify much more to cities’ path-dependent formation than of convergence to a global urban model. Global forces may be currently assisting the deepening of divisions, but the diversity of outcomes indicates that they do not determine the outcome on their own. An argument for dualization was identified in Paris and Athens, where socially extreme spaces have become more extreme during the 1990s. At the same time, however, the mode of socially mixed living was dominant (socially mixed areas cover half of these cities) and not regressing, while extreme spaces remained of rather reduced importance. 3. Ethno-racial identity does not appear to be the primary axis of segregation outside deeply discriminating contexts, even though in several cities included in this book ethno-racial diversity is gaining importance with the increasing size of immigrant communities. In most of the cities where ethno-racial diversity is not negligible, discrimination is not absent but socioeconomic position remains the main criterion for segregation. However, wherever anti-segregation policies are developed, they are almost exclusively related to ethnoracial spatial concentrations and are usually inspired by an unwarranted fear of US style ghettoization. The reality of segregation is, therefore, not necessarily related to the way it is socially and politically perceived. 4. The rich are more segregated than the poor. In almost all of the 11 cities, the higher occupational or income categories are the more segregated and the ones that often continue to increase their spatial distance from the rest; this happens at the same time that segregation for most of the other categories is decreasing. However, the seclusion of the rich is not regarded as constituting a political problem, and the focus remains exclusively on the segregation of the poor. Again, the social and political perception of segregation is dependent more on ideological rapports de force than on documented accounts of its condition and of its effects. 5. There is increasing socioeconomic diversity in urban space following the subdivision of former broader social and functional divisions between Centre/Periphery, East/West or North/South, which does not necessarily lead to increased segregation. There are multiple processes leading to the increase of socio-spatial subdivision in metropolitan space. Most of them involve some form of invasion of groups or functions into territories occupied

Introduction

29

by other groups or functions. Urban renewal and its potentially gentrifying outcome; the expansion of upper-middle classes into mixed and middle social areas leading to their embourgeoisement; “gated communities” implantation near lower status areas; the spatial ‘entrapment’ of socially mobile households in working class areas and overcrowding of large apartments by immigrants in high status areas etc. are all processes that socially subdivide urban space. Some see in these processes a further partitioning of urban space and the erection of new walls between social or ethno-racial groups (Marcuse and van Kempen 2002). In fact all these processes bring social and functional differences closer in space; they change the scale of segregation and decrease segregation levels—at least temporarily. What will come out of these processes represents an open social and political stake vested with contradictory interests and forces rather than a done deal; outcomes can reasonably be expected to be varied and to complicate further the relation between social and spatial distance. 6. The level of segregation depends to a large extent on state policies and this has not fundamentally changed under conditions of capitalist globalization. Energetic state policies usually oppose the increase of segregation. This is true of welfare state policies developed in Northern and Western Europe, even though their range and anti-segregation impact may be decreasing in Copenhagen or their implementation being segmented and problematic in Paris. In the case of the East Asian developmental state, the anti-segregation effect may be less intentional, but derives from the largely egalitarian regulation of the labor market. In Southern Europe the residual-clientelist welfare model impedes segregation through family centered practices and networks that inhibit to some extent the shifting and sorting by the housing market. There is an obvious decline in the impetus and effectiveness of policies and practices that have opposed segregation. 7. Policies affecting segregation are not always planned with such an objective and, when they are, their objective is not necessarily attained. Moreover, similar policies may lead to different outcomes in different cities. In different contextual conditions, similar policies appear to have dissimilar effects: the selling of social housing to sitting tenants, for example, has increased segregation in Budapest much more than in Beijing following the higher marketability of housing, the greater diversity of housing types that favors residential mobility, and the openness to foreign demand in the former. Another example is the impact of public housing on segregation, which has been changing over time in Europe especially as it lost its appeal to social groups other than the neediest, following the cutting of welfare expenditure (see also Marcuse and van Kempen 2002a, Andersen 2004). On the contrary, in Hong Kong and Tokyo public housing is targeted on a wide range of social groups and counteracts segregation, even though in the former it is spatially very unevenly distributed. In Hong Kong, the model of social housing and planning regulations affects the business strategy of the dominant private and very large housing projects that seek internal diversity to ensure their profitability, and produce social mixing as an unintended outcome. Social housing may, thus, affect segregation in different ways depending on its quality, spatial distribution and social targeting. 8. Even the extreme commodification of housing provision does not necessarily lead to increased segregation. Housing production and land development in Hong Kong are very large scale operations within a sector of highly concentrated capital that make housing production in Athens—where building operations rarely comprise more than one relatively small building—look like petty commodity production. However, in both cases the segregating potential of the produced stock is relatively reduced due to the social profile of

Residential Segregation in Comparative Perspective

30

the agents involved (Athens) and to business considerations affected by public housing and planning policies (Hong Kong). 9. The decreasing segregation trend in many cities does not necessarily mean less inequality and more intense and effective social contact between different groups. Less segregation is often combined with increased deprivation among lower social groups and minorities. It may be the outcome of gentrification processes or of the implantation of gated communities in lower status residential areas—as in Istanbul and São Paulo—or of the dominant mode of immigrant integration in the housing market in Southern Europe, where public housing is scarce and reserved for natives. Immigrants are compelled to find solutions in the private rented sector, in overcrowded apartments in middle and upper-middle class areas in Madrid or in lower floors and basements in Athens’s apartment buildings. Moreover, in some cases the relatively low level of residential segregation is accompanied by a higher segregation of services (especially schools). Outside “American exceptionalism” (Marcuse and van Kempen 2000a) spatial distance remains a poor indicator of social distance and the relationship is becoming weaker. Segregation, in its simple traditional definition as the spatial separation of population groups, was a concept devised to address the clear-cut separation of ethno-racial groups in the residential areas of the booming industrial metropolis of the US in the 1920s. For various reasons, it has progressively assumed general validity and has been applied in different contexts around the globe, in most cases taking with it the contextual assumptions it was bundled with, and in particular the confusion between spatial and social distance, the exclusive focus on lower status groups and the assumption of important negative effects. As a simple and stand-alone index it becomes poor and often misleading when used to make inferences about urban social inequality in settings where socio-spatial separation is more intricate, and social distances are far more complex than their spatial reflexion. Research on segregation continues, however, to provide important insight into cities’ socio-spatial structures and into the ways urban social inequality is reproduced, provided that it is adequately informed theoretically and properly contextualized. References Abu-Lughod, J. 1969. “Testing the theory of social area analysis: The case of cairo, Egypt.” American Sociological Review 34, 198–212. Allen, J., Barlow, J., Leal, J., et al. 2004. Housing and Welfare in Southern Europe. Oxford: Blackwell. Andersen, H.T. 2004. “Spatial – not social polarization: Social change and segregation in Copenhagen.” The Greek Review of Social Research 113(A), 145–65. Arapoglou, V.P. 2006. “Immigration, segregation and urban development in Athens: The relevance of the LA debate for southern European metropolises.” The Greek Review of Social Research 121(C), 11–38. Arbaci, S. 2007. “Ethnic segregation, housing systems and welfare regimes in Europe.” European Journal of Housing Policy 7(4), 401–33. Arbaci, S. 2008. “(Re)Viewing ethnic residential segregation in Southern European cities: Housing and urban regimes as mechanisms of marginalisation.” Housing Studies 23(4), 589–613.

Introduction

31

Arbaci, S. and Malheiros, J. 2010. “De-segregation, peripheralisation and the social exclusion of immigrants: Southern European cities in the 1990s.” Journal of Ethnic and Migration Studies 36(2), 227–55. Atkinson, R. and Kintrea, K. 2001. “Disentangling area effects: Evidence from deprived and nondeprived neighbourhoods.” Urban Studies 38(12), 2277–98. Badcock, B. 1984. Unfairly Structured Cities. Oxford: Blackwell. Baum, S. 1997. “Sydney, Australia: A global city? Testing the social polarisation thesis.” Urban Studies, 34(11), 1881–901. Baum, S. 1999. “Social transformations in the global city.” Urban Studies 36(7), 1095–117. Berry, B.J.L. and Rees, P.H. 1969. “The factorial ecology of Calcutta.” American Journal of Sociology 74, 445–91. Boal, F.W. 1987. “Segregation.” In Social Geography: Progress and Prospect, ed. M. Pacione. Beckenham: Croom Helm, 90–128. Borel-Saladin, J. and Crankshaw, O. 2009. “Social polarisation or professionalisation? Another look at theory and evidence on deindustrialisation and the rise of the service sector.” Urban Studies 46(3), 645–64. Brenner, N. and Theodore, N. 2002. “Cities and the geographies of ‘actually existing neoliberalism’.” Antipode 34(3), 349–79. Brun, J. 1994. “Essai critique sur la notion de la ségrégation et sur son usage en géographie urbaine.” In La Ségrégation dans la Ville, eds J. Brun and C. Rhein. Paris: L’Harmattan, 21–57. Buck, N. 2001. “Identifying neighbourhood effects on social exclusion.” Urban Studies 38(1), 2251–75. Buck, N. and Gordon, I. 2004. “Does spatial concentration of disadvantage contribute to social exclusion?” In City Matters. Competitiveness, Cohesion and Urban Governance, eds M. Boddy and M. Parkinson. Bristol: The Polity Press, 237–54. Burgers, J. and Vranken, J. eds 2003. How to Make a Successful Urban Development Program; Experiences from Nine European Countries. Antwerp: UGIS. Butler, T. 2007. “For gentrification?” Environment & Planning A 39(1), 162–81. Butler, T., Hamnett, C., Ramsden, M., et al. 2006. “The best, the worst and the average: Secondary school choice and education performance in East London.” Journal of Education Policy 21(6), 7–29. Chamboredon, J.C. and Lemaire, M. 1970. “Proximité sociale et distance spatiale. Les grands ensembles et leur peuplement.” Revue Francaise de Sociologie XI(1), 3–33. Chaolin, G. and Kesteloot, C. 2002. “Beijing’s socio-spatial structure in transition.” In Studies in Segregation and Desegregation, eds I. Schnell and W. Ostendorf. Aldershot: Ashgate, 285–311. Charre, J. 1995. Statistique et Territoire. Montpellier: RECLUS. Clark, E. 2005. “The order and simplicity of gentrification – a political challenge.” In Gentrification in a Global Context: The New Urban Colonialism, eds R. Atkinson and G. Bridge. London: Routledge, 256–64. Coulter, P.B. 1989. Measuring Inequality. A Methodological Handbook. Boulder: Westview Press. Desrosières, A. and Thévenot, L. 1988. Les catégories socioprofessionnelles. Paris: La Découverte. Domburg-De Rooij, T. and Musterd, S. 2002. “Segregation and the welfare state.” In Studies in Segregation and Desegregation, eds I. Schnell and W. Ostendorf. Aldershot: Ashgate, 107–31. Duncan, O.D. and Duncan, B. 1955. “Residential distribution and occupational stratification.” American Journal of Sociology 60, 493–503. Ellen, I. and Turner, M. 1997. “Does neighbourhood matter? Assessing recent evidence.” Housing Policy Debate 8(4), 833–66.

32

Residential Segregation in Comparative Perspective

Esping Andersen, C. 1990. The Three Worlds of Welfare Capitalism. Cambridge: Polity Press. Fainstein, S.S., Gordon, I. and Harloe, M. eds 1992. Divided Cities. New York and London in the Contemporary World. Oxford: Blackwell. Fishman, R. 1987. Bourgeois Utopias: The Rise and Fall of Suburbia. New York: Basic Books. Friedmann, J. and Wolff, G. 1982. “World city formation: An agenda for research and action.” International Journal of Urban and Regional Research 6(3), 309–43. Fuzita, K. 2003. “Neo-industrial Tokyo: Urban development and globalization in Japan’s Statecentered developmental capitalism.” Urban Studies 40(2), 249–81. Goldthorpe, J.H. 2000. On Sociology. Oxford: Oxford University Press. Gordon, I. and Monastiriotis, V. 2006. “Urban size, spatial segregation and inequality in educational outcomes.” Urban Studies 43(1), 1–24. Goering, J. and Feins, J. eds 2003. Choosing a Better Life? Evaluating the Moving to Opportunity Experiment. Washington, DC: Urban Institute Press. Hamnett, C. 1994. “Social polarisation in global cities. Theory and evidence.” Urban Studies 31(3), 401–24. Hamnett, C. 1996. “Social polarisation, economic restructuring and welfare state regimes.” Urban Studies 33(8), 1407–30. Hamnett, C. 2001. “Segregation and social polarization.” In Handbook of Urban Studies, ed. R. Paddison. London: Sage, 162–76. Hamnett, C. 2003. Unequal City. London in the Global Arena. London: Routledge. Hamnett, C. and Cross, D. 1998. “Social polarisation and inequality in London: The earnings evidence, 1979–95.” Environment and Planning C 16(6), 659–80. Häussermann, H. 2005. “The end of the European city?” European Review 13(2), 237–49. Hill, R.C. and Fujita, K. 2003. “The nested city: Introduction.” Urban Studies 40(2), 207–17. Hill, R.C. and Kim, J.W. 2000. “Global cities and developmental states.” Urban Studies 37(12), 2167–98. Holgersson, H., Thörn, C., Thörn, H. et al., eds 2010. (Re)Searching Gothenburg. Gothenburg: Glänta Produktion. Johnston, R., Forrest, J. and Poulsen, M. 2002. “Are there ethnic enclaves/ghettos in English cities?” Urban Studies 39(4), 591–618. Johnston, R.J., Gregory, D. and Smith, D.M. eds 1986. The Dictionary of Human Geography. Oxford: Blackwell. Kazepov, Y. 2005. “Introduction.” In Cities of Europe. Changing Contexts, Local Arrangements, and the Challenge to Urban Cohesion, ed. Y. Kazepov. Oxford: Blackwell, 3–42. Kesteloot, C. 1998. “The geography of deprivation in Brussels and local development strategies.” In Urban Segregation and the Welfare State. Inequality and Exclusion in Western Cities, eds S. Musterd and W. Ostendorf. London: Routledge, 126–47. Kesteloot, C., Murie, A. and Musterd, S. 2006. “European cities: Neighbourhood matters.” In Neighbourhoods of Poverty. Urban Social Exclusion and Integration in Europe, eds S. Musterd, A. Murie and C. Kesteloot. Basingstoke: Palgrave, 219–38. Knox, P. and Pinch, S. 2006. Urban Social Geography. Harlow: Pearson. Knox, P. and Taylor, P.J. eds 1995. World Cities in a World System. Cambridge: Cambridge University Press, 48–62. Ladányi, J. 2002. “Residential segregation among social and ethnic groups in Budapest during the post-communist transition.” In Of States and Cities. The Partitioning of Urban Space, eds P. Marcuse and R. van Kempen. Oxford: Oxford University Press, 170–82.

Introduction

33

Leal, J. 2004. “Segregation and social change in Madrid metropolitan region.” The Greek Review of Social Research 113, 81–104. Lees, A. 1985. Cities Perceived. Urban Society in European and American Thought, 1820–1940. Manchester: Manchester University Press. Lees, L. 2008. “Gentrification and social mixing: Towards an inclusive urban renaissance?” Urban Studies 45(12), 2449–70. Leontidou, L. 1990. The Mediterranean City in Transition. Cambridge: Cambridge University Press. Logan, J.R., Stults, B. and Farley, R. 2004. “Segregation of minorities in the metropolis: Two decades of change.” Demography 41, 1–22. Lupton, R. 2003. ‘Neighbourhood Effects’: Can We Measure Them and Does it Matter? CASE Paper 73. London: LSE. Maloutas, T. 2004. “Segregation and residential mobility: Socially entrapped social mobility and its impact on segregation in Athens.” European Urban and Regional Studies 11(2), 171–87. Maloutas, T. 2007a. “Segregation, social polarisation and inequality in Athens during the 1990s: Theoretical expectations and contextual difference.” International Journal of Urban and Regional Research 31(4), 733–58. Maloutas, T. 2007b. “Socio-economic classification models and contextual difference. A look at the European Socio-economic Classes (ESeC) from a South European angle.” South European Society & Politics 12(4), 443–60. Maloutas, T. 2012. “Contextual diversity in gentrification research.” Critical Sociology 38(1), 33–48. Marcuse, P. 1989. “Dual city: A muddy metaphor for a quartered city.” International Journal of Urban and Regional Research 13(4), 697–708. Marcuse, P. 2002. “The layered city.” In The Urban Lifeworld, eds P. Mandsen and R. Plunz. London: Routledge, 94–114. Marcuse, P. 2002a. “The partitioned city in history.” In Of States and Cities. The Partitioning of Urban Space, eds P. Marcuse and R. van Kempen. Oxford: Oxford University Press, 11–34. Marcuse, P. 2002b. “The shifting meaning of the Black ghetto in the United States.” In Of States and Cities. The Partitioning of Urban Space, eds P. Marcuse and R. van Kempen. Oxford: Oxford University Press, 109–42. Marcuse, P. and van Kempen, R. 2000. “Introduction.” In Globalizing Cities. A New Spatial order? eds P. Marcuse and R. van Kempen. Oxford: Blackwell, 1–21. Marcuse, P. and van Kempen, R. 2000a. “Conclusion: A changed spatial order.” In Globalizing Cities. A New Spatial Order? eds P. Marcuse and R. van Kempen. Oxford: Blackwell, 249–75. Marcuse, P. and van Kempen, R. eds 2002. Of States and Cities. The Partitioning of Urban Space. Oxford: Oxford University Press. Marcuse, P. and van Kempen, R. 2002a. “States, cities and the partitioning of urban space: Conclusions.” In Of States and Cities. The Partitioning of Urban Space, eds P. Marcuse and R. van Kempen. Oxford: Oxford University Press, 258–68. Marcuse, P. and van Kempen, R. 2002b. “States, cities and the partitioning of urban space.” In Of States and Cities. The Partitioning of Urban Space, eds P. Marcuse and R. van Kempen. Oxford: Oxford University Press, 3–10. Massey, D.S. and Denton, N.A. 1989. “Hypersegregation in US Metropolitan areas: Black and Hispanic segregation along five dimensions.” Demography 26(3), 373–91. Massey, D.S. and Denton, N.A. 1993. American Apartheid: Segregation and the Making of the Underclass. Cambridge, MA: Harvard University Press.

34

Residential Segregation in Comparative Perspective

McElrath, D.C. 1962. “The social areas of Rome. A comparative analysis.” American Sociological Review 27, 376–91. Meert, H., Mistiaen, P. and Kesteloot, C. 1997. “The geography of survival: Household strategies in urban settings.” Tijdschrift voor Sociale en Economische Geografie 88(2), 169–81. Mollenkopf, J.H. and Castells, M. eds 1991. Dual City: Restructuring New York. New York: Russel Sage Foundation. Musterd, S. 2005. “Social and ethnic segregation in Europe: Levels, causes and effects.” Journal of Urban Affairs 27(3), 331–48. Musterd, S., Murie, A. and Kesteloot, C. eds 2006. Neighbourhoods of Poverty. Urban Social Exclusion and Integration in Europe. Basingstoke: Palgrave. Musterd, S. and Ostendorf, W. eds 1998. Urban Segregation and the Welfare State. Inequality and Exclusion in Western Cities. London: Routledge. Musterd, S., Ostendorf, V. and De Vos, S. 2003. “Neighbourhood effects and social mobility: A longitudinal analysis.” Housing Studies 18(6), 877–92. Musterd, S. and van Kempen, R. 2009. “Segregation and housing of minority ethnic groups in Western European cities.” Tijdschrift voor Economische en Sociale Geografie 100(4), 559–66. Orr, L., Feins, J.D., Jacob, R., et al. 2003. Moving To Opportunity for Fair Housing Demonstration Interim Impacts Evaluation. Washington DC: US Department of Housing and Urban Development, www.huduser.org/publications/fairhsg/mtofinal.html [accessed: 18/07/2011]. Ostendorf, V., Musterd, S. and De Vos, S. 2001. “Social mix and the neighbourhood effect. Policy ambitions and empirical evidence.” Housing Studies 16(3), 371–80. Pahl, R.E. 1975. Whose City? And Further Essays on Urban Society. Harmondsworth: Penguin Books. Pahl, R.E. 1988. “Some remarks on informal work, social polarization and the social structure.” International Journal of Urban and Regional Research 12(2), 247–67. Pahl, R.E. 2001. “Market success and social cohesion.” International Journal of Urban and Regional Research 25(4), 879–83. Park, R.E. 1916 [1957]. “The city: Suggestions for the investigation of human behaviour in the urban environment.” American Journal of Sociology XX, 577–612. Reprinted in Park, R.E. 1957. Human Communities. The City and Human Ecology. Glencoe, IL: The Free Press, 13–51. Park, R.E. 1936 [1957]. “Human ecology.” American Journal of Sociology XLII, 1–15. Reprinted in Park, R.E. 1957. Human Communities. The City and Human Ecology. Glencoe, IL: The Free Press, 145–58. Pickvance, C. 2002. “State socialism, post-socialism, and their urban patterns: Theorizing the Central and Eastern European experience.” In Understanding the City, eds J. Eade and C. Mele. Oxford: Blackwell, 183–203. Polanyi, K. 1944. The Great Transformation. New York: Rinehart. Préteceille, E. 1993. Mutations urbaines et politiques locales. Seégrégation sociale et budgets locaux en Ile-de-France. Paris: CSU. Préteceille, E. 1995. “Division sociale de l’espace et globalisation.” Sociétés Contemporaines 22/23, 33–68. Préteceille, E. 2004. “Décrire et analyser la ségrégation sociale.” In Les mécanismes fonciers de la ségrégation. Paris: ADEF-Études Foncières, 9–29. Robinson, G.M. 1998. Methods & Techniques in Human Geography. Chichester: Wiley. Robson, B.T. 1969. Urban Analysis. A Study of City Structure with Special Reference to Sunderland. Cambridge: Cambridge University Press.

Introduction

35

Rose, D. and Harrison, E. 2007. “The European socio-economic classification: A new social class schema for comparative European research.” European Societies 9(3), 459–90. Sassen, S. 1991. The Global City: New York, London, Tokyo. Princeton: Princeton University Press. Schiffer, S.R. 2002. “Economic restructuring and urban segregation in São Paulo.” In Of States and Cities. The Partitioning of Urban Space, eds P. Marcuse and R. van Kempen. Oxford: Oxford University Press, 143–69. Schnell, I. 2002. “Segregation in everyday life spaces: A conceptual model.” In Studies in Segregation and Desegregation, eds I. Schnell and W. Ostendorf. Aldershot: Ashgate, 39–65. Shevsky, E. and Bell, W. 1955. Social Area Analysis. Stanford: Stanford University Press. Sjoberg, G. 1960. The Pre-Industrial City. Chicago: Free Press. Slater, T. 2006. “The eviction of critical perspectives from gentrification research.” International Journal of Urban and Regional Research 30(4), 737–57. Smith, N. 1987. “Gentrification and the rent gap.” Annals of the Association of American Geographers 77(3), 462–65. Smith, N. 1996. The New Urban Frontier: Gentrification and the Revanchist City. London: Routledge. Swyngedouw, E., Moulaert, F. and Rodriguez, A. 2003. “‘The world in a grain of sand’: Large scale urban development projects and the dynamics of ‘glocal’.” In The Globalized City. Economic Restructuring and Social Polarisation in European Cities, eds F. Moulaert, A. Rodriguez and E. Swyngedouw. Oxford: Oxford University Press, 9–28. Timms, D. 1971. The Urban Mosaic. Towards a Theory of Residential Differentiation. Cambridge: Cambridge University Press. Topalov, C. 2001. “Maurice Halbwachs et les villes. Les expropriations et les prix des terrains à Paris (1909).” In La Ville des Sciences Sociales, eds B. Lepetit and C. Topalov. Paris: Belin, 11–45. Vaattovaara, M. and Kortteinen, M. 2003. “Beyond polarization versus professionalisation? A case study of the Helsinki region, Finland.” Urban Studies 40(11), 2127–45. van Kempen, R. 2002. “The academic formulations: Explanations for the partitioned city.” In Of States and Cities. The Partitioning of Urban Space, eds P. Marcuse and R. van Kempen. Oxford: Oxford University Press, 35–56. van Kempen, R. 2002a. “Towards partitioned cities in the Netherlands? Changing patterns of segregations in a highly developed welfare state.” In Of States and Cities. The Partitioning of Urban Space, eds P. Marcuse and R. van Kempen. Oxford: Oxford University Press, 88–108. van Kempen, R. 2005. “Segregation and housing conditions of immigrants in Western European cities.” In Cities of Europe. Changing Contexts, Local Arrangements, and the Challenge to Urban Cohesion, ed. Y. Kazepov. Oxford: Blackwell, 190–209. van Kempen, R. and Murie, A. 2009. “The new divided city: Changing patterns in European cities.” Tijdschrift voor Economische en Sociale Geografie 100(4), 377–98. van Zanten, A. 2001. L’école de la périphérie. Scolarité et ségrégation en banlieue. Paris: Presses Universitaires de France. Vance, J.E. Jr. 1971. “Land assignment in pre-capitalist and post-capitalist cities.” Economic Geography 47, 101–20. Vaughan, L. and Arbaci, S. 2011. “The challenges of understanding urban segregation.” Built Environment 37(2), 128–38. Wacquant, L. 1997. “Three pernicious premises in the study of the American ghetto.” International Journal of Urban and Regional Research 21(2), 341–53.

36

Residential Segregation in Comparative Perspective

Wacquant, L. 2008. Urban Outcasts. A Comparative Sociology of Advanced Marginality. Cambridge: Polity. Walks, R.A. 2001. “The social ecology of the post-fordist/global city? Economic restructuring and socio-spatial polarisation in the Toronto Urban Region.” Urban Studies 38(3), 407–47. Wessel, T. 2000. “Social polarisation and socio-economic segregation in a welfare state: The case of Oslo.”. Urban Studies 37(11), 1947–67. White, P. 1984. The West-European City. A Social Geography. London: Longman. Wilson, W.J. 1987. The Truly Disadvantaged: The Inner City, the Underclass and Public Policy. Chicago: Chicago University Press. Wong, D.W.S. 1993. “Spatial indices of segregation.” Urban Studies 30(7), 559–72.

Chapter 2

Residential Income Inequality in Tokyo and Why it Does Not Translate into Class-based Segregation Kuniko Fujita and Richard Child Hill

Introduction The degree of class-based residential segregation in a society can carry enormous implications for the life chances of urban residents, including their educational and occupational opportunities, the vitality of their neighborhoods, and the attractiveness of their central cities. International comparisons suggest that the extent of spatial class segregation and its social implications are heavily influenced by national and regional contexts. Social and economic institutions, politics and policies, family structures and social networks, norms and culture shape residential patterns in urban societies and may intensify or dampen the impact of income inequality on residential segregation. Peter Marcuse and Ronald van Kempen (2000) question whether globalization has led to greater spatial inequality in cities of the world in their edited collection, Globalizing Cities: A New Spatial Order? Their answer is ambiguous. The authors contend that globalization, a multifaceted process embracing deindustrialization, international migration flows, economic structural change, and heightened capital mobility, may have generated a new urban spatial order polarized between a higher income class linked to the global economy and a lower income class engaged in local services but acknowledge that cities, nonetheless, vary in the degree and impact of class residential segregation depending upon local factors. In the United States, control over territory is a basic means of reproducing class privilege since residential location provides differential access to valued resources that are critical to social mobility and class reproduction. Municipal jurisdictions in metro areas are starkly stratified by income and education. Wealthy communities pay lower taxes for higher quality public services than poor communities because privileged groups can wield place-based jurisdictional powers to exclude others from their tax-revenue and service-rich enclaves (Dreier et al. 2004, Swanstrom et al. 2004). Place matters for American urban residents and it matters most for inner city black residents whose life chances are severely limited by poverty and racial discrimination (Wilson 1987, 2009). European cities display a different pattern. For example, Musterd, Murie and Kesteloot (2006) examine whether class-based spatial segregation exists in European cities by focusing on poor neighborhoods where international immigrants also tend to concentrate. They find that the welfare state, NGOs and clientele relations have helped keep poor areas from developing into classsegregated neighborhoods in Northern, Central, and Southern European cities and that low-income neighborhoods are fluid enough to alleviate class segregation. With the exception of the UK, they conclude that spatial location is not an important factor in reproducing urban poverty in Europe. The few studies of spatial segregation in Japanese cities, including Osaka (Fujita and Hill 1997), Tokyo (Waley 2000, Jacobs 2005), Nagoya (Jacobs 2003), and Kyoto (Fielding 2004)

38

Residential Segregation in Comparative Perspective

indicate that Japanese cities are closer to the European than to the American and British pattern. Waley (2000) and Jacobs (2005) find that spatial inequality in Tokyo is growing but is not wide enough to indicate social polarization.1 Fujita and Hill (1997) conclude that geography plays a minor role in social stratification in Osaka while Fielding finds that Kyoto is less spatially stratified by occupation than Edinburgh, Scotland.2 This chapter describes and seeks to explain the distribution of income and occupational groups among places in Tokyo, including levels and patterns of spatial income inequality and trends over time. The paper further explores how various political, economic and cultural factors prevent spatial income inequality from growing into class based segregation in Tokyo. We draw upon data from the census, other governmental publications and the Geographical Information System (GIS) statistics to measure income and occupational differences between Tokyo’s central city (23 wards) and surrounding suburbs (26 cities); and among the 23 wards, the 877 smaller sized neighborhoods (chos) and the 2,970 smallest sized spatial units (chomes) making up the central city. In this study Tokyo’s “central city” refers to the 23 ward area, and Tokyo’s “suburbs” refer to the remaining 26 cities encompassed by the Tokyo metropolitan government.3 Tokyo has a special status in Japan’s intergovernmental structure. Tokyo functions as a prefecture, of which Japan has 47, rather than as a typical city. Tokyo has a metropolitan government headed by an elected Governor and Assembly (see Map 2.1). The area encompassed by the 23 special wards constituted the city of Tokyo until 1943 and can be thought of as the central city. The special wards are now separate, self-governing municipalities, each with an elected Mayor and Assembly, but they differ from the ordinary Japanese city in that some government functions are handled by the Tokyo Metropolitan Government (TMG). The suburban cities, towns and villages also have local governments with elected Mayors and Councils. Tokyo had about 13 million inhabitants in 2010: over 8.9 million lived in the central 23 ward area and 4 million in the suburban area. TMG combines with the adjoining three prefectures of Chiba, Kanagawa and Saitama to form the greater Tokyo area. With over 35 million inhabitants, the Tokyo capital region which includes four more prefectures surrounding the greater Tokyo area ranks among the world’s most populous metropolitan regions. Per capita income data for Tokyo’s wards and suburban cities come from fiscal tax records compiled by Japan’s Ministry of Internal Affairs and Communications and contained in publications by the Japan Real Estate Institute, the Japan Marketing Education Center, and JPS. Government census and per capita income data exclude registered foreigners who made up 3.2 percent of Tokyo’s population in 2010.4 1  Jacobs finds that spatial inequality has been growing in Tokyo because public sector investment has been skewed to Tokyo’s downtown areas causing uneven development between central Tokyo and outlying neighborhoods. Nonetheless, argues Waley, spatial stratification is relatively weak in Tokyo because the gentrification prevailing in Western cities is largely absent in Japan. 2  Fujita and Hill’s (1997) study of income distribution among wards in Osaka drew criticism from Fielding (2004) who argued that ward sized areal units were too large to generate meaningful conclusions about spatial inequality. However, his own study of occupational distribution among smaller area units in Kyoto comes to much the same conclusion. 3  The five towns and eight villages are excluded in this analysis. 4  Eighty-three percent of registered foreigners lived in Tokyo’s central city as 2010. The registered foreign residents scattered throughout 23 wards but they were relatively more concentrated in the four wards of Minato (6 percent), Shinjuku (10 percent), Edogawa (7 percent), and Adachi (6.7 percent) than the rest of the central city (TMG Bureau of General Affairs 2010). Koreans and Chinese composed over 50 percent of the registered foreigners in Tokyo and Filipinos, Americans, Thais and others made up the rest.

Tokyo Metropolitan Government: Central city (23 wards) and suburbs (26 cities, 5 towns and 8 villages)

Source: Tokyo Metropolitan Government, Administrative Areas of Tokyo, Tokyo—City profile and government: http://www.metro.tokyo.jp/ENGLISH/ PROFILE/area.htm [accessed: 27 March 2010].

Map 2.1

Residential Segregation in Comparative Perspective

40

Residential Income Inequality in Tokyo The first part of this chapter documents and analyzes the distribution of income and occupational groups between Tokyo’s central city wards and suburban cities and among the wards, chos and chomes making up the central city. Income and occupational inequalities are measured by the coefficient of variation.5 Our summary findings are as follows. Spatial income inequality in Tokyo is largely between the four central core wards which constitute Tokyo’s CBD and the rest of the metropolitan area. Tokyo’s four core wards (Chiyoda, Chuo, Minato and Shibuya) have by far the highest per capita income in the city; there is little variation in per capita income among the rest of Tokyo’s 23 central city wards and 26 suburban cities. Tokyo’s spatial income inequality rises and falls over time, in a wave-like pattern, in line with fluctuations in the fortunes of the four core wards. The income of the four core wards in turn moves in lockstep with fluctuations in their land value. Land value and per capita income correlate highly in Tokyo; but occupation and per capita income correlate less. Tokyo’s spatial income inequality is not a function of social class segregation. Functional Primacy Property assets peak in Tokyo’s central core area, decline precipitously in adjoining areas, and flatten out over the rest of Tokyo (see Figure 2.1). The price of housing land in central Chiyoda

Figure 2.1

Housing and commercial land value per square meter in Tokyo, 2010

Source: TMG Bureau of Finance (2010).

5  The coefficient of variation is the standard deviation of a distribution divided by the mean.

Residential Income Inequality in Tokyo

41

ward, for example, is seven times higher than outlying Adachi ward, and 17 times higher than in Ome, a suburban city the farthest distance from Tokyo’s central core.6 The sky high price of land in the central core wards and the L-shaped, tapering-with-distance land price gradient reflect the functional primacy of Tokyo’s CBD.7 Functional primacy refers to the central management activities of Japan’s major economic, political and socio-cultural institutions concentrate overwhelmingly in Tokyo’s four central wards (Hill and Fujita 1995).8 The functional primacy of Tokyo’s central core wards can be traced to the Feudal Era. Chiyoda, Chuo and Minato became premium places after Tokyo became Japan’s national political capital in

Map 2.2

Tokyo’s 23 central city wards and four core wards

Source: Author’s own.

6  In 2009, commercial land value averaged over 3 million yen per square meter in Tokyo’s central core, over 1 million yen in immediately surrounding wards, 0.6 million in the next ring of wards, and 0.4 million yen on the periphery of the central city. Housing value per square meter follows the same L-shaped pattern (TMG Bureau of Finance 2010). (Exchange rates were 86 yen for US$1 and 112 yen for 1 euro as of August 2010.) 7  The centrifugal power or functional primacy exerted by the CBD is also indicated by the ratio of daytime to resident population in the central wards: The daytime population is 20 times larger than the resident population in Chiyoda ward; 6.6 times in Chuo ward; and 5 times in Minato wards (TMG Bureau of Industrial and Labor Affairs 2008). 8  Approximately half of Japan’s major corporations are headquartered in Tokyo and half of those are located in the central core wards. These include 83 percent of Japan’s largest banks, 82 percent of the country’s major insurance companies, 59 percent of big security firms and the headquarters of all foreign banks and security companies (TMG Bureau of Industrial and Labor Affairs 2008: 34).

42

Residential Segregation in Comparative Perspective

1603 because they surrounded Edo castle, the domicile of the Tokugawa Shogunate government. Edo castle later became the Emperor’s residence and so it remains today. Residential segregation in feudal Tokyo mirrored the rigid class pecking order among ruling samurai, farmers, craftsmen, and merchants. The samurai class lived in Chiyoda and Minato. Craftsmen and merchants resided in Chuo. The Meiji government abolished the feudal class system, including its residential counterpart, in the middle of the nineteenth century. As the Meiji state promoted industrialization, the samurai residential areas became the locale for central government offices, corporate headquarters and public institutions. Chiyoda evolved into Japan’s political nerve center, housing the Japanese Diet (Parliament) and Kasumigaseki (office compound for the national government bureaucracy) adjacent to Edo castle. Chuo, home to Ginza, is Japan’s best known commercial, cultural and entertainment center. Minato hosts most of the foreign embassies in Japan. Shibuya joined Chiyoda, Chuo and Minato as functionally prime in 1980 (see Map 2.2). Sub-centers have emerged elsewhere in metro Tokyo but they have not altered the functional primacy of the four core wards. Spatial Income Inequality Japan evolved during the half century after World War II into one of the world’s most egalitarian societies. But the perception of Japan as a middle class nation has come into question in recent years. The pumping up of a huge asset bubble in the late 1980s, and its abrupt rupture in 1990, propelled Japan into a trajectory of near-zero growth that has persisted to the present day. Public concern with stagnant employment and wages led to a debate in the government, academy and media over the extent to which Japan was turning into an unequal society. According to data collected by the Ministry of Internal Affairs and Communications, income inequality, measured by the Gini coefficient, was quite low in Japan in the mid-1980s and grew only slightly by the end of the 1990s—from 0.25 in 1984 to 0.27 in 1999 (MIC 2002). However, a different body of data, gathered by the Ministry of Health, Labor and Welfare, showed a much higher level of income inequality, 0.50, in 2002 (MHLW 2004). A 2006 OECD study, drawing upon the MHLW data, concluded that institutional mechanisms for income redistribution were weak in Japan and the nation had indeed become one of the most unequal societies among OECD member states (OECD 2006). The OECD conclusions buttressed arguments made by some Japanese researchers (Tachibanaki 1998, Hara and Seiyama 1999, Sato 2000) that Japan was turning into a highly stratified class system. However, a number of analysts have contested the “unequal society” thesis. Skeptics argue that the MHLW data, which underpin the OECD conclusions, greatly overestimate the degree of income inequality in Japan because they exclude government employees and workers employed in very small businesses (MHLW 2007). A recent United Nations study reported that income inequality in Japan averaged 0.25 in 2007, still ranking Japan’s income distribution among the most egalitarian in the world (UN 2009: 195). Others contend that income redistributive mechanisms specific to Japan must be taken into account in the measurement of inequality (Ota 2006). Of particular salience, some argue, is the seniority wage system, which, by tying salaries to age, concentrates higher incomes among older workers yet also enables younger workers to raise their incomes the longer they stay in their workplace (Ohtake and Saito 1998, Ohtake 2005). Social mobility remains high in Japan, according to research sponsored by the Ministry of Finance, dampening the impact of income inequality (Ministry of Finance Policy Research Institute 2004).

Residential Income Inequality in Tokyo

43

These issues are far from settled.9 In the study that follows we address the neglected yet potentially significant spatial dimension of the inequality debate. We document the level of income inequality among residential areas in Tokyo, we explore the extent to which income and occupational disparities among local districts have been increasing over time, and we analyze the institutional forces that promote or prevent class segregation in the city. Table 2.1 shows trends in the distribution of per capita and per household income and spatial income inequality among Tokyo’s 49 local districts (23 central city wards plus 26 suburban cities) between 1971 and 2005. Spatial income inequality, measured by the coefficient of variation, increased over the 35 year period, in an undulating, wave like pattern. Inequality in per capita income among Tokyo’s localities fell 13 percent in the 1970s, rose 42 percent in the 1980s, fell 23 percent in the 1990s, and rose 46 percent in the 2000s, with an overall, net rise of 63 percent during the three and a half decades. Inequality in income per household followed the same spatial trajectory.10 Table 2.1

Per capita and per household income inequality among Tokyo’s 49 local districts (23 central city wards plus 26 suburban cities), 1971–2005 (unit: 10,000 yen) 1971

1980

1990

2000

2005

Mean

37

107

202

204

207

Standard deviation

8

20

63

49

73

0.216

0.187

0.312

0.241

0.353

Mean

103

286

492

449

431

Standard deviation

14

36

116

78

113

0.136

0.126

0.236

0.174

0.262

Per capita

Std/M Per household

Std/M

Note: Per capita income and per household income refer to total taxable individual and household income divided by the total number of individuals and households respectively in each ward and city. Taxable individual and household income include the sum of salaries, bonuses, income from individual businesses, pensions, interest on savings, dividends, rents and real estate sales, after some work related expenses are deducted. Source: Japan Real Estate Institute (Nihon Fudosan Kenkyujo); Statistics on Real Estate (Fudosan Toukei) for 1980–1990; Japan Marketing Education Center, Individual Income Indicators (Kojin Shotoku Shihyo) for 1971 and 2000; and JPS, Individual Income Indicators for 2005.

9  With continued economic stagnation, the focus of the inequality debate has shifted to the national government’s neoliberal reform policies. Inequality research now centers on especially disadvantaged groups, including women, youth, and the homeless (Ishida and Slater 2009; Tachibanaki 2010). 10  Inequality in income per household among Tokyo’s wards fell 7 percent in the 1970s, rose 87 percent in the 1980s, declined 26 percent in the 1990s, and rose 5 percent in the 2000s, with a net rise of 93 percent between 1971 and 2005.

Residential Segregation in Comparative Perspective

44

Figure 2.2

Per capita and per household income inequality among Tokyo’s 49 local districts (23 central city wards plus 26 suburban cities), 1971–2005

Source: Based on Table 2.1.

The inequality trend line crests twice, as shown in Figure 2.2. The first and highest peak occurs in 1990.11 1990 marks both the pinnacle and the termination point for Japan’s financial bubble. The 1980s financial bubble created a steep rise in land values in Tokyo’s four central wards, relative to the rest of the city. Soaring land values spawned a huge increase in the average incomes of individuals and families living in the central core, relative to outlying areas, thus issuing in the steep rise in income inequality among Tokyo’s 49 districts between 1980 and 1990. Spatial income inequality falls between 1990 and 2000, and then rises to a second peak in 2005, again reflecting parallel movements in land values and incomes in the central wards. The resurgence of land values in the central core wards between 2000 and 2005 mirror a governmentinitiated urban redevelopment boom designed to boost Japan’s stagnant economy (Fujita 2011).12 How does the spatial relationship between land value and income account for the large income gap between Tokyo’s four core wards and the rest of the city and the wave like fluctuations in that income disparity overtime? We do not have data to document our hypotheses, but we believe the following four hypotheses are important. First, it is likely that a substantial segment of Tokyo’s rentier class lives in the four central wards. By rentier class we mean the land owners, brokers, builders, mortgage bankers and service providers who make their living from the real estate game. Rising land prices boost rentier profits, through increased land sales, rents, and fees for service, thus raising per capita ward income. Second, when land values soar, local homeowners have an incentive to lease or sell land they own in order to realize capital gains and/or pay for rising property taxes, thus raising per capita ward income. Third, the white collar employment and cultural amenities generated by new commercial and office development in the central wards entice higher income professional and managerial 11  At 0.312 per capita income and 0.236 household income. 12  Specifically, the 2001 Urban Renaissance Plan. See Urban Renaissance Projects at Prime Minister’s Office http://www.toshiseisei.go.jp.

Residential Income Inequality in Tokyo

45

groups to move into the area, thus raising per capita ward income. Finally, as higher land prices make it more expensive to live in the four central core wards, lower income residents are forced to move out, also raising per capita ward income. The process operates in reverse as well: as land values decline, rentier income declines, as does homeowner income from leases and property sales, and as does the residential price barrier facing low income Tokyoites.13 The 23 central city wards and 26 suburban cities  Table 2.2 and Figure 2.3 display separate income trends for Tokyo’s central city wards and suburban cities. The distinctive wave-like movement in income inequality among Tokyo’s 49 districts, revealed when city wards and suburbs are combined at the metropolitan level, actually applies only to central city wards; the inequality trend among suburban cities is flat, displaying almost no variation over the years between 1971 and 2005. Table 2.2

Per capita income inequality among Tokyo’s 23 central city wards and 26 suburban cities, 1971–2005 (unit: 10,000 yen) 1971

1980

1990

2000

2005

9 40 0.225

25 115 0.217

82 231 0.355

63 227 0.278

95 244 0.389

5 34 0.147

11 101 0.109

23 178 0.129

21 186 0.113

24 178 0.135

23 central city wards Standard deviation Mean Std/M 26 suburban cities Standard deviation Mean Std/M Source: See Table 2.1.

Tokyo’s central city is wealthier than the suburbs, and there are no signs that the central city’s dominant economic position in the metropolitan area is weakening, just the reverse. As can be calculated from Table 2.2, Tokyo’s central city to suburban mean income ratio increased between 1971 and 2005, from 118 percent to 137 percent. Table 2.2 also shows that income inequality is much lower among Tokyo’s suburbs than among wards in the central city—roughly two to three times lower depending upon the year selected. In sharp contrast to the central city, income inequality among suburbs, negligible to begin with, actually fell slightly (8 percent) over the 35 year period. The 1980s financial bubble, which had such a profound impact on income inequality among central city wards, barely affected the distribution of income among Tokyo’s suburban communities.

13  The inequality pattern is exactly the same for per capita and for per household income inequality throughout the period. To avoid repetition we will just focus on per capita income inequality in the analysis that follows.

Residential Segregation in Comparative Perspective

46

Figure 2.3

Per capita income inequality among Tokyo’s 23 central city wards, 26 suburban cities and Tokyo (23 wards plus 26 cities), 1971–2005

Source: Based on Tables 2.1 and 2.2.

Four central core wards and the 23 central city wards  Table 2.3 shows per capita income for the four core wards (Chiyoda, Chuo, Minato and Shibuya) and the rest of the central city wards between 1971 and 2005. Per capita income in the four core wards peaked in 1990, at the high point of the bubble, and again in 2005 with the urban redevelopment boom in the CBD. Per capita income declined in the four core wards between 1990 and 1995, indicating the loss of resident income caused by the financial collapse. By contrast, with the exception of Bunkyo, income in the wards outside the four core wards continued to grow, reaching a high point in 1995, five years after the financial collapse.14 Prior to the asset bubble, per capita income growth in the four core wards paralleled that of the rest of the city. The bubble economy broke that pattern, sharply pushing up per capita income in the four core wards. Income growth in the core wards averaged 72 percent between 1985 and 1990, compared to 30 percent among wards in the rest of the central city. Table 2.3

Per capita income among Tokyo’s four central core wards and the rest of 23 central city wards, 1971–2005 (unit: 10,000 yen)

Four core wards Rest of central city

1971

1975

1980

1985

1990

1995

2000

2005

54 37

107 69

153 106

217 137

371 198

330 211

331 203

409 205

Source: See Table 2.1.

14  Adjacent to the central core, Bunkyo ward experienced a spillover of bubble inflated land prices from the CBD.

Residential Income Inequality in Tokyo

Figure 2.4

47

Per capita income inequality among 23 central city wards, with and without Tokyo’s four core wards (Chiyoda, Chuo, Minato and Shibuya), 1971–2005

Source: See Table 2.1.

Figure 2.4 shows that the fluctuating fortunes of four core wards largely determine the rise and fall of income inequality among all wards in the central city. Two trend lines display income inequality among central city wards between 1971 and 2005; one includes, the other excludes the four core wards. Removing the four core wards nearly erases the increase in spatial inequality over the 35-year period. When the four core wards are included, income inequality grows by 73 percent. When the four core wards are excluded, income inequality among the remaining wards grows by only 14 percent. Income inequality among central city wards outside of the four central core wards is negligible, as in the suburbs. And like the suburbs, per capita income in noncore wards did not peak until 1995, five years after the financial collapse. Incomes in the noncore wards grew little after 1995, as in the suburbs, due to the onset of the deflationary economy. Clearly, Tokyo’s cyclical ups and downs in real estate prices are largely confined to the central core area; and it is the fluctuating fortunes in the four core wards that largely determine the rise and fall of spatial income inequality in the central city and in the metropolitan area as a whole. Land prices and per capita income  The income of a Tokyo ward and city correlates very highly with the value of land in the district. As shown in Figure 2.5, the Pearson product moment correlation between a ward and city’s per capita income and the ward and city’s residential land value per square meter is 0.898. As expected, the four core wards stand out in the upper right quadrant (high income, high land value) of the figure among 49 local districts. Unsurprisingly, per capita ward income correlates highly with the ward and city’s housing prices and rent.15

15  Housing prices, like land prices, are highest in Tokyo’s central core wards and decline abruptly with distance from the CBD. According to a recent government study, the average price for condominium housing in central Minato ward was 11.2 million yen per square meter; 10.2 million yen in Chiyoda ward and 10

48

Figure 2.5

Residential Segregation in Comparative Perspective

Correlation between per capita income and housing land value per square meter among Tokyo’s 49 local districts (23 central city wards and 26 suburban cities), 2005

Source: Authors’ calculations based on the following data: JPS, Individual Income Indicators for 2005; Ministry of Land, Information, Transport and Tourism (MLIT); 2005 Chika koji [published land value in 2005], http://www.tochi.mlit.go.jp [accessed: January 2010].

Cyclical real estate booms and busts  Functional primacy continues to inflate Tokyo’s land values relative to other Japanese cities, especially in the four core wards where the headquarters of the nation’s central institutions are concentrated, while at the same time rendering Tokyo’s traditional CBD particularly sensitive to ups and downs in the real estate and financial markets. As noted above, land values in central Tokyo ballooned in the latter half of the 1980s, collapsed in the early 1990s, remained low throughout the 1990s, and turned upward again in the mid-2000s. Government policies that counteracted the rise of the yen against the dollar in the mid-1980s helped create speculative gambling on real estate and pushed land prices in central Tokyo sky high in the latter half of the 1980s (Fujita 2011).16 Individuals who lived in the four central wards and who owned land and did business in city real estate gained immense fortunes from speculationmillion yen in Shibuya ward (TMG Bureau of Housing 2006: 22). The least expensive housing, 4.6 million yen per square meter, exists in the outlying, lower income Adachi ward. A correlation between per capita income and rent among Tokyo’s 49 districts was high at 0.84 in 2005. 16  Three decades of high growth, deregulation of foreign currencies, and successful American pressure on Japan to raise the value of the yen against the dollar all contributed to the bubble. The higher yen contracted Japan’s export growth and brought recession. To boost the economy, the state lowered interest rates. Lower interest rates flooded the financial market with cheap money and easy credit, enabling speculation on land and stocks, thus driving up asset prices.

Residential Income Inequality in Tokyo

49

inflated prices during in the 1980s bubble. When the bubble economy collapsed in the early 1990s, land prices plummeted 60 percent. Residents in the four core wards lost fortunes as fast as they had gained them. As their income plummeted, income inequality among Tokyo’s wards and suburban cities also dropped. Today, 20 years later, the price of land in Tokyo’s central area is only one-fifth the peak 1990 level. The upward trend in land values in the 2000s was triggered by a public policy led urban redevelopment boom. The central government launched a nationwide Urban Renaissance Policy to stimulate economic growth and designated central Tokyo as one of the principal redevelopment areas. Officials from the central state and TMG encouraged corporations to renovate their headquarter offices and participate in adjacent urban redevelopment projects. Tokyo’s top corporations responded with huge new investments in the CBD (Fujita and Hill 2008). As Tokyo’s functional primacy among Japanese cities continues to stand out in new office development, the supply of new offices in Tokyo’s central core has kept growing. The new office supply in the CBD was 39 times higher than in Osaka City in 2006 (TMG Bureau of Industrial and Labor Affairs 2008: 102). Within the central city, the CBD accounted for 80 percent of new office supply between 2000 and 2005 (MLIT 2010a). Residents in the central wards once again began reaping fortunes from an investment boom, thus accounting for the second peak in Tokyo’s spatial income inequality in 2005. The bubble economy was short-lived. Yet the ensuing financial crisis deflated Japan’s economy over the next 10 years, the so-called “lost decade” of the 1990s. Japan’s economic stagflation in the 1990s went hand in hand with a decline in Tokyo’s spatial income inequality. The urban redevelopment boom in the early 2000s renewed income growth in the four central wards while incomes continued to slump in the rest of Tokyo. Income inequality among the city’s districts turned upward again, reaching its second and highest peak in 2005. Tokyo’s CBD land values began falling again in 2010 in response to the economic downturn caused by the 2008 global financial crisis (MLIT 2010b).17 Predictably adverse impact of plummeting land prices on four central core ward incomes suggests that spatial income inequality in the city may be once again on the decline by 2020. Place Stratification by Occupation in Tokyo We have documented the distribution of income among metro Tokyo’s 23 central city wards and 26 suburban cities and analyzed trends in income inequality among those districts between 1971 and 2005. Place stratification in metro Tokyo, as indicated by the coefficient of variation in per capita income among these localities, remains quite moderate but has tended upward over the 35-yearperiod, in a wave-like pattern. Income inequality among Tokyo’s 49 districts essentially boils down to a large income gap between residents in Tokyo’s four core wards and residents in the rest of the city. That income gap mirrors huge differences in land prices between the four central core wards and outlying districts. Land values reach stratospheric levels in Tokyo’s core wards, and there is very high, indeed a nearly perfect, positive correlation between housing land value and per capita income among 17  Between 2009 and 2010, housing land value per square meter declined 13.5 percent from 2 million yen to 1.86 million yen in Chiyoda, 12.2 percent from 0.87 million yen to 0.76 million yen in Chuo, 8.7 percent from 1.4 million yen to 1.3 million yen in Minato, and 9.4 percent from 1.04 million yen to 0.95 million yen in Shibuya. Tokyo as a whole declined 6.8 percent (MLIT 2010b).

50

Residential Segregation in Comparative Perspective

Tokyo’s wards and suburban cities. On the other hand, per capita income and professional and managerial occupations are less strikingly correlated among Tokyo’s 46 local districts.18 Four core wards’ incomes rise and fall with booms and busts in land prices, as does income inequality among 49 districts in Tokyo as a whole. Spatial units at the scale of central city wards and suburban cities aggregate sizeable populations and sometimes fairly wide territories. Averages at this scale may mask considerable variation among smaller scale sub-communities and therefore underestimate the actual degree of place stratification in Tokyo. We would like to do a more fine-grained analysis. Tokyo’s census does subdivide the central city into spatial units smaller than wards: these include “cho,” of which there are 877, and at the smallest scale, “chome,” of which there are 2970. Unfortunately, income data is not available for these smaller spatial units. However, data on the occupation of residents is available for these smaller districts. Occupation cannot serve as a surrogate for income but it can provide insight into another dimension of spatial inequality in Tokyo: place stratification based upon social status. Do birds of an occupational feather tend to flock together in Tokyo? Are Tokyo’s neighborhoods segregated among occupational groups with some residential districts made up predominantly of production workers, for example, and others predominantly occupied by higher status professional and managers? Or, alternatively, do occupational groups tend to mix together in Tokyo’s residential areas? Occupational trends in Tokyo  Unlike New York and London, Tokyo’s economy as a whole and central city economy in particular are not skewed toward the finance, insurance and real estate service industry (FIRE). The effects of industrial, technological, and urban changes on the occupational structure have been gradual as Japan’s state-centered capitalism emphasized full employment (Fujita 1991, 2003). So have the effects of globalization. Opening up the domestic market for foreign trade and investment and increasing Japan’s foreign investment abroad have been incremental as the state adopted a managed trade policy (Fujita and Hill 1993, 2012). Tokyo residents are today employed in a wide range of industries: construction (6 percent), manufacturing (11 percent), information and telecommunication (7 percent), transport (5 percent), wholesale and retail trades (19 percent), FIRE (7 percent), restaurants and hotels (7 percent), medical and welfare (7 percent), education (4 percent), services (19 percent), and government work (3 percent). Tokyo’s resident workforce also spreads rather evenly across the major occupational categories: professional and managerial (20 percent), clerical (25 percent), sales (17 percent), services (11 percent), and production and laborers (19 percent) in 2005. The percent of metro Tokyo’s labor force employed in professional, technical, clerical, sales and service jobs has been increasing over time, while the percent in managerial and production jobs has been declining.19 Occupational distribution among central city wards, chos and chomes  The four core wards contain a slightly higher percentage of residents employed in professional, technical and 18  The correlation is 0.747. 19  For the entire Tokyo metro area, between 1980 and 2005, professional and technical workers increased from 11.3 percent to 17.1 percent and clerical, sales and services workers grew from 54.0 percent to 56.2 percent, while managerial workers decreased from 7.0 percent to 3.1 percent and production and laborers from 26.7 percent to 19.9 percent in Tokyo. Japan’s industry classification also changed in 2002. Newspaper and other media related publications moved from manufacturing to information (services). This reflected the decline of the workforce in manufacturing. Furthermore, corporate restructuring for flatter organization reflected on the declining number of managers.

Residential Income Inequality in Tokyo

Table 2.4

51

Occupational inequality among Tokyo Central City’s 23 wards, 877 chos, and 2,970 chomes, 2005 Professional, technical and managerial

Production and laborers

5 21 0.231

6 17 0.351

8 21 0.381

8 18 0.444

8 21 0.381

8 18 0.444

23 wards Standard deviation Mean Std/M 877 chos (smaller neighborhoods) Standard deviation Mean Std/M 2,970 chomes (smallest neighborhoods) Standard deviation Mean Std/M

Note: Chos and chomes with less than 80 jobs are merged with adjacent units. Source: Authors’ calculations based on Statistical GIS from Population Census in 2005; Ministry of Internal Affairs and Communication (MIC) Bureau of Statistics, Statistical GIS, 2005.

managerial (P&M) positions than wards in the rest of the city, but by and large, Tokyo’s P&M jobs are spread rather uniformly throughout the central city and suburbs.20 The highest residential concentration of this high status occupational group is not among the four core wards but in Bunkyo at 29.8 percent. Table 2.4 displays occupational inequality among Tokyo’s 23 wards, 877 smaller sized neighborhoods (chos) and 2,970 smallest spatial units (chomes). These occupational categories broadly represent the highest and lowest status groups in Tokyo’s labor force. P&M workers have the highest average incomes and the most years of education, while production workers and laborers have the least. We excluded the middle level job categories (e.g. clerical, sales, and service) because residents employed in these positions are evenly distributed throughout the city. We measured occupational inequality for 2000 and 2005. As there was little change between 2000 and 2005, we omitted the 2000 data. Occupational inequality measured by coefficient of variation among the 23 wards was 0.23 for the P&M group and 0.35 for the production group in 2005. The inequality figures increase somewhat for the smaller spatial units: 0.38 for the P&M group and 0.44 for the production group among both 877 chos and 2,970 chomes. Among the 877 smaller “cho” neighborhoods, only three had over 50 percent of their residents employed in P&M occupations in 2005: Kanda-Surugadai (71 percent) and Rokubancho (52 percent) in Chiyoda ward and Yushima (53 percent) in Bunkyo ward. Among the smallest 2,970 chomes, only 11 units had a majority of P&M residents (see Figure 2.6). Seven of these 20  Measured by coefficient of variation, the P&M distribution inequality among Tokyo’s 49 districts is quite low at 0.213 in 2005.

Residential Segregation in Comparative Perspective

52

Figure 2.6

Percent distribution of professional and managerial jobs, and production and laborer work among 2970 chomes of Tokyo Central City, 2005

Source: Authors’ drawing based on data from Ministry of Internal Affairs and Communication (MIC) Bureau of Statistics, Statistical GIS, 2005.

chomes are in the four central core wards and the rest are located in Bunkyo, Ota, Setagaya and Shinjuku wards (see Table 2.5). Tsukiji 5-chome in Chuo ward has the highest concentration of P&M residents at 97 percent and followed by Hongo 8-chome in Bunkyo. The professionals, technicians and managers living in these 11 majority P&M chomes tend to work in the medical industry. As it turns out, these chomes contain residential facilities for employees in hospitals and medical research institutes. Tsukiji 5-chome in Chuo ward, with the highest P&M concentration, is home to the National Cancer Center’s hospital and research facilities. Hongo 7-chome in Bunkyo ward has the second highest P&M concentration and houses Tokyo University’s General Hospital and medical research centers. Kanda-Surugadai 2-chome in Chiyoda ward is home to Tokyo Medical and Dental University’s hospitals and research institutes. Table 2.6 displays the 10 chomes with the highest residential land value in 2005.21 While all but one has a higher percentage of P&M residents than their respective wards and central city as a whole, none fall among the majority P&M chomes seen in Table 2.5. Four of the chomes have a higher percentage of residents working in the finance, insurance and real estate (FIRE) industries than exists in their respective wards and in the city as a whole, which supports the rentier hypothesis discussed earlier, but six do not. Places with the highest land value are not predominantly occupied by individuals who work in FIRE, or in P&M positions or managerial positions; they are populated by residents who work in all kinds of industries and occupations. As these 10 places exemplify only a tiny portion of the four core wards where land prices are the highest in Tokyo, we have looked at the correlation between percent P&M residents and percent FIRE residents among the 75 chomes that make up Chiyoda ward and the 355 chomes that make up the four core wards. The correlation is negligible at 0.059 for Chiyoda and 0.082 for the entire four central core wards. The fact that land prices do not correlate highly with the occupational and industry characteristics of residents suggests that the predominant factor behind the high spatial correlation between land value and per capita income may lie not with a wealthy rentier class or a privileged P&M status group, but with average individual property 21  The ranking of the highest 10 places changes every year but falls within the four core wards.

Residential Income Inequality in Tokyo

Table 2.5

Ward Chuo Chiyoda

53

Tokyo Central City’s chomes with over 50 percent residents employed in professional, technical and managerial (P&M) jobs, and industry with highest percent P&M, 2005 Name of chomes

Tsukiji 5-chome Kanda-Surugadai 2-chome* Kanda-Sakumacho 4-chome Rokubancho** Minato Nishi-Shinbashi 3-chome Toranomon 4-chome Bunkyo Hongo 7-chome Ota Denenchofu 3-chome Setagaya Okura 2-chome Shibuya Hiroo 4-chome Shinjuuku Koyama 1-chome

Total jobs Percent P&M Highest percent industry in P&M 375 318 117 366 314 221 409 792 595 872 1,525

97.1 71.4 51.5 52.2 62.7 77.4 92.7 53.9 57.3 55.5 52.2

93 in medical 77 in medical 88 in medical 80 in medical 20 in wholesale 66 in medical 88 in medical 23 in services 47 in medical 23 in services 41 in medical

Note: * Kanda-Surugadai 2-chome combines 1-chome which consists of less than 80 jobs; ** Rokuban-cho is a cho but is included because it does not contain any smaller sized chome. Source: Authors’ calculations taken from Statistical GIS based on 2005 Population Census Data; Bureau of Statistics; Ministry of Internal Affairs and Communication.

Table 2.6

Ten chomes with highest housing land prices by percent professional, technical and managerial (P&M) and percent employed in finance, insurance and real estate (FIRE) industries, 2005 Location of highest land prices

Chiyoda ward 5 bancho* 1 bancho* 3 bancho* Hirakawa 2-chome Kudan-kita 2-chome Fujimi 2-chome Minato ward Motoazabu 2-chome Akasaka 1-chome Akasaka 6-chome Minami-Aoyama 4-chome

Total jobs

P&M

FIRE

21,053

28.6

10.7

340 1,098 1,044 235 181 938

40.3 41.4 39.4 30.2 45.3 40.6

10.1 12.0 8.0 13.2 10.3 5.0

74,581

28.4

10.3

547 110 874 1,096

33.8 40.9 26.4 34.8

13.1 20.1 13.1 17.6

Note: * 5 bancho, 1 bancho and 3 bancho are chos but are included because they don’t contain any smaller sized chome. Source: Ministry of Internal Affairs and Communication (MIC) Bureau of Statistics; Statistical GIS 2005; Ministry of Land, Information, Transport and Tourism (MLIT); 2005 Chika koji [published land value in 2005].

Residential Segregation in Comparative Perspective

54

owners, who finding themselves unexpectedly caught up in a speculative real estate game boom they had no part in creating, have an incentive to lease or sell their land to increase their income and/or pay for rising taxes. Social mixing  Table 2.7 shows the mix of occupational groups at the ward, cho and chome levels in Tokyo’s central city and compares occupational mixing among higher income, middle income and lower income areas. Minato ward, in the CBD, represents a higher income area (H); Suginami ward, a residential neighborhood in the Western part of Tokyo, represents a middle income area (M); and Adachi ward, an industrial and commercial district on Tokyo’s East side, represents a Table 2.7

Occupational mix of residents at Tokyo Central City’s ward, cho and chome levels, 2005

Tokyo Central City

Professional & technical

Managerial

Clerical

Sales

Service

Safety & Transport & Production protection communication

17.2

4.1

24.7

17.9

11.4

1.3

2.6

17.5

A. Higher income ward Minanto ward (29 chos) (H)

20.4

8.0

25.4

16.9

12.4

1.6

1.4

9.7

Roppongi-cho (consists of 7 chomes) (Ha)

19.8

11.9

23.3

18.6

14.4

1.7

5.5

5.9

5-chome (1)

23.8

13.4

21.8

19.9

12.1

0.3

0.3

4.2

7-chome (2)

20.5

7.1

22.9

17.6

19.3

0.3

1.1

8.0

Shiba-cho (consists of 7 chomes) (Hb)

15.1

5.7

25.5

16.8

16.7

0.6

1.9

14.2

2-chome (3)

15.6

6.3

27.4

19.1

15.5

0.6

1.2

11.5

5-chome (4)

11.1

3.1

25.2

14.5

20.1

0.8

2.3

19.1

Suginami ward (36 chos) (M)

21.3

3.9

27.6

17.6

10.4

1.2

1.7

15.1

Shimotakaido-cho (consists of 5 chomes) (Ma)

21.6

3.9

26.3

17.6

11.2

0.9

1.5

13.3

1-chome (5)

21.7

1.9

28.1

18.2

11.7

1.1

1.4

12.7

5-chome (6)

21.1

4.2

24.9

16.1

9.4

0.7

1.6

14.9

Zenpukuji-cho (consists of 4 chomes) (Mb)

23.6

6.2

28.7

15.5

8.7

0.9

3.7

10.9

3-chome (7)

21.1

6.6

30.2

16.4

9.1

0.5

1.6

12.1

4-chome (8)

25.5

6.5

26.8

16.2

7.7

0.8

1.7

9.7

Adachi-ward (83 chos) (L)

10.7

2.0

21.2

16.4

11.2

1.4

5.5

28.3

Aoi-cho (consists of 6 chomes) (La)

9.5

1.6

21.6

16.6

12.2

1.1

5.5

29.1

1-chome (9)

11.6

2.2

18.3

21.3

11.8

0.7

5.3

26.1

3-chome (10)

9.6

1.2

16.3

13.1

10.9

0.8

4.7

23.1

Adachi-cho (consists of 4 chomes) (Lb)

12.8

2.6

24.7

18.1

11.6

1.5

3.4

22.1

1-chome (11)

10.9

2.1

23.8

19.7

11.3

1.3

3.5

23.1

3-chome (12)

15.3

3.2

25.1

18.3

10.5

1.7

3.0

19.8

B. Middle income ward

C. Lower income ward

Source: Ministry of Internal Affairs and Communication (MIC) Bureau of Statistics, Statistical GIS, 2005.

Residential Income Inequality in Tokyo

Figure 2.7

55

Percent occupational mix of Tokyo Central City residents among higher, middle and lower income wards

Source: Based on Table 2.7.

Figure 2.8

Percent occupational mix of Tokyo Central City residents among higher, middle and lower income chos (based on average figures of 2 chos from each ward)

Note: L indicates averages of La and Lb; M and H indicate averages of Ma and Mb, and Ha and Hb respectively. Source: Based on Table 2.7.

Residential Segregation in Comparative Perspective

56

Figure 2.9

Percent occupational mix of Tokyo Central City residents among higher, middle and lower income chomes (based on average figures of 4 chomes from each ward)

Note: L indicates averages of 1, 2, 3 and 4; M indicates averages of 5, 6, 7 and 8; and H indicates averages of 9, 10, 11 and 12. Source: Based on Table 2.7.

lower income area (L).22 Two smaller chos are randomly selected from each ward (indicated by Ha and Hb , Ma and Mb , and La and Lb . And two smallest scale chomes are selected from each cho (indicated as 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, and 12 ). Figure 2.7 graphically displays the occupational mix among wards at different income levels, Figure 2.8 among chos, and Figure 2.9 among chomes. Three findings stand out. First, a complete mix of occupational groups resides at all three spatial scales—ward, cho and chome. Second, the mix of occupational groups at each spatial scale and income levels is nearly identical. And third, even the highest income ward, cho and chome has a substantial percentage of residents who work in production jobs, and even the lowest income ward, cho and chome has a substantial percentage of residents who work in professional and technical fields. In sum, there is little evidence of social class segregation in Tokyo, as indicated by the distribution of occupational groups among residential areas. Tokyo’s neighborhoods tend to be socially mixed at every spatial scale and at every income level. Why Residential Income Inequality Does Not Translate into Class Segregation in Tokyo The institutional power centers that underpin Tokyo’s functional primacy in Japan’s urban system are concentrated in the city’s four core wards. The gap between the fortunes of the four core wards 22  Mix refers to statistical mix but not to interaction between people of different occupational groups.

Residential Income Inequality in Tokyo

57

and the rest of Tokyo’s districts essentially determines spatial income inequality in metro Tokyo; there is almost no variation in per capita income among communities outside the central core. Tokyo’s spatial income inequality is not a function of social class segregation. Land value and per capita income correlate highly in Tokyo; but occupation and residential segregation do not. Apart from a few sub-communities in the core wards with high concentrations of residents in the medical professions, Tokyoites employed across the whole occupational spectrum live in proximity to one another in neighborhoods throughout the metro area. The class characteristics of their neighbors have little bearing on where Tokyoites choose to live; rather they choose their place of residence based upon commuting time to the workplace, access to commercial and cultural amenities, and stage in the life cycle. A 2008 survey asked Tokyoites to list their preferred residential areas. The top choice went to Kichijoji neighborhood located in the city of Musashino. Kichijoji was selected for its easy access to public transport and the abundance of commercial facilities near its main train station.23 The next most favored location was Jiyugaoka neighborhood in Setagaya-ward, followed by Ebisu neighborhood in Shibuya-ward. The latter two areas have streets lined with fashionable shops and restaurants and are especially attractive to young adults. Residential location has little bearing on the life chances of Tokyoites. Several attributes of Tokyo’s political economy help explain why. A Different Form of Capitalist Society The compressed wage system  Japan’s compressed wage system is perhaps the primary reason why residential income inequality does not translate into social class segregation in Tokyo.24 Managerial salaries in Japan are on average only three times larger than the wages of rank and file workers. The average wages of managerial and technical employees in the manufacturing sector out-distance production employees by only 41 percent (MHLW 2009). A corporate president’s salary seldom exceeds 10 times that of the lowest paid rank and file worker. By contrast, executive pay at Standard & Poor’s 500 corporations in the USA averages 200 times the pay of the typical American worker (New York Times 2010). Salaries and wages in Tokyo’s finance sector rank highest among the nation’s industries25 but they are only 7 percent above the next highest paying utility industry; 17 percent above information and telecommunication, education, medical and welfare; 32 percent above construction, manufacturing, wholesale, retail trade, and services; and 67 percent above the lowest paying industry, restaurants and hotels (TMG Bureau of Industrial and Labor Affairs 2009). The corporate community  The compressed wage system is linked to a widespread sense of the “corporation as a community” in Japan. The corporation as community idea evolved out of postwar workers’ control movements and is buttressed by enterprise based social compacts between labor unions and corporate management (Gordon 1985, 2001). Corporate practices such as the seniority-based wage system, long-term investment management, and the weak power of 23  The real estate information provider, Next Co., conducted an online survey in November 2008 querying 4,607 Tokyo residents, aged 20 and older, about their residential location decisions (Nikkei Net 2009). 24  Wilkinson and Pickett argue that there are different routes to greater income equality; a large welfare state is one, Japan’s compressed wage system is another (Wilkinson and Pickett 2009: 236–7). 25  The average monthly wage in the finance industry was only 32 percent higher than Tokyo’s average monthly wage in 2008 (TMG Bureau of Industrial and Labor Affairs 2009).

58

Residential Segregation in Comparative Perspective

shareholders add further substance to the corporate community concept (Matsumoto 1991). The corporation is viewed more as a public institution owned by employees than as a profit-making tool for shareholders. Management and labor cooperate to reach corporate goals and share the fruits of better performance through bonuses distributed among the entire workforce. Profits are invested in R&D and in worker training for future growth. During economic downturns, maintaining corporate employment is the first priority; managerial salaries are the first to be reduced; and shareholder interests come last.26 The corporate community has weakened in the past decade, as Japanese companies have become more profit-oriented, reformed long-term relations with their main bank and supplier stakeholders, altered their employment practices, and reduced their workforces. Yet, while the diversity of business practices has increased, the corporate community principle still remains in place (Inagami and Whittaker 2005, Vogel 2006). For example, lifetime employment may not be as prevalent today, but large companies continue their efforts to retain worker loyalty through practices like employee shareholding systems, a practice engaged in by 63 percent of Japanese companies listed on the Tokyo Stock Exchange (Tokyo Stock Exchange 2007, Ono 2010). Compressed wages are particularly identified with the seniority wage system, a practice prevailing in both the private and public sectors.27 Wages are lowest for young workers who are at the beginning of their careers, reach the highest level when employees are between the ages of 50 and 60, and gradually decline thereafter. In principle, the longer employees serve in the same company, the greater the skill levels they attain and the higher the wages and bonuses they earn. Years of schooling have less of an impact on promotion and earnings in the corporate community. Wage differences among workers do not become significant until their peak earning years, between 50 and 60, when the highest wages are reached for company employees at every education level: college, junior college, technical school, and high school. At that point the average difference in wages between college and high school graduates is 47 percent (MHLW 2009). Top executives often come out of the rank and file work force, often were union members at one time, and often possess only a high school degree. Long-term investment vs. short-term profit making  The corporate practice of long term investment also makes the compressed wage system possible. As seen earlier, bond and equity markets and the finance industry do not play a major role in the Tokyo economy (or in the Japanese economy as whole). More Japanese corporations raise capital from the stock market these days but they do not depend upon shareholders alone for capital. Large corporations still look to the main banks with which they are interlocked to purchase their bonds. Playing the shareholders’ role, main banks do not place a priority on profit maximization. Freed from the pressure of quarterly dividend reports, companies are free to make long-term investment plans that in turn make intra corporate training possible.28 The compressed wage system is also linked to a collective approach to innovation. The creation of new products and technologies is attributed to a team, a division, and the corporation as a 26  The corporate community concept spread to the public sector. Firms in the small business sector (which 97 percent of Tokyo firms belonged to) could not afford to fully practice the concept but in principle adhered to it. 27  Many corporations have partly replaced their seniority-based salary systems with systems based on performance. This change is not indicative of a move toward profit maximization, however, and can be interpreted as reforms that preserve the corporate community. 28  The interlocking shareholding and bank-credit financial systems also underpinned the emergence of the corporation as community in postwar Japan. These systems have changed over the past two decades.

Residential Income Inequality in Tokyo

59

whole. Individual rewards exist but are less valued. The team approach sharply contrasts with the American reward system, where an individual researcher who invents a new product while working for a public research institute or a university could go into business for herself and then be rewarded in the form of a high IPO in the stock market. Compressed wages also shape households incomes. A 2005 TMG survey found that the income of employed households headed by managers exceeded Tokyo’s average household income by just 41 percent. Managerial household income was 16 percent higher than professional, 35 percent higher than clerical, and 94 percent higher than production household income (TMG 2006). Government policy also helps the corporate community retain the compressed wage system. In economic downturns, like the recession triggered by the 2008 global financial crisis, the state provides subsidies to corporations to assist them in maintaining employment. The corporate community, with its unique system of labor and industrial relations, and with the state providing support indirectly to workers through subsidies to their employers, has been a source of collective cohesion in postwar Japan. However, the cohesion engendered by corporation as community practices is eroding. The corporate community concept only ever fully applied to large companies. Non-regular workers have grown to a third of Japan’s labor force; they lack employment security, and are not covered by the corporate welfare system offered by big companies. Needed are fundamental reforms that will extend corporation as community benefits to all strata of the workforce.29 Intergovernmental Relations, Public Policies and Institutions Tokyo in Japan’s unitary state structure  While the decentralization law of 2000 has provided Tokyo with more autonomy in policy making, Japan’s unitary state structure still reflects an integrated approach to intergovernmental relations, emphasizing the benefits to be gained from central-local coordination, shared responsibilities and intertwined competences among levels of government. Because control over public education, the tax system, and government policies concerning the environment, health and welfare, science, technology, industry, regional and urban development is centralized in the national government in Japan, public resources and services are more or less evenly allocated among local districts throughout the Tokyo Metropolitan Area. Public policies: The social investment state  Public policy in Japan is mostly fiscal investment policy achieved through the redistribution of tax revenues. The state’s primary goals are to promote national economic growth and maintain Japan’s competitiveness in the world economy. While the central government’s fiscal investment policies have been embedded in pork barrel, clientele politics, they have also distributed new infrastructure spending and jobs to regions where economic growth has been limited, thus evening out development among Japans regions and cities (Fujita and Hill 2012). Public policies could also generate unintended consequences, creating financial bubbles in the 1980s and 2000s and intensifying income inequality among places as seen in this study. Corporations have shifted emphasis from bank borrowing to stock to raise capital. Yet they continue to maintain the corporate community as their basic philosophy through lifetime employment, a less hierarchical organizational style, internal labor markets, and a corporate welfare system. 29  Large corporations provide their workforce with education and training in their own technical colleges, health insurance and hospital care, credit and pension systems, housing subsidies and mortgages, travel agents and resort hotels, sports and cultural facilities.

60

Residential Segregation in Comparative Perspective

Tokyo’s urban policy is couched within the central state’s fiscal investment policy framework. Tokyo’s planners are also sensitive to the distribution of public investment. The city’s urban and industrial policies add local specificity to central state policy, like environmental protection, an aging society and future population decline. Within this centrally coordinated but locally specified plan, planners choose new areas for investment and job growth and plan the revitalization of older industrial areas and declining downtown commercial districts. They allocate funds for transport networks, for example, not only to alleviate traffic and population congestion, but also in response to public concerns about unemployment rates, population decline, and depressed industries (Fujita and Hill 2008). Centralized tax system  A centralized, redistributive tax system is another institution that prevents Tokyo’s spatial income inequality from developing into class segregation. The decentralization law of 2000 called upon the central state to delegate more policy-making power to cities, but it did not radically alter Tokyo’s approach to policy-making since the process already involved extensive central state consultation with local governments. As the headquarters site for large corporations and the dwelling place for high income residents, Tokyo’s central core is a tax-rich locale. Ninetyeight percent of the central city’s corporate tax revenues are generated in just three core wards: Chiyoda, Chuo and Minato (TMG Bureau of Taxation 2005).30 But Japan’s tax system does not allow tax rich ward governments to simply retain the revenue for their own purposes. The state redistributes the revenue throughout Japan. Tokyo Metropolitan Government also collects taxes from the 23 wards and redistributes the revenues throughout the central city based upon urban development priorities and public service needs.31 Residents in the four core wards pay much higher income taxes per capita and per household than Tokyoites dwelling in districts elsewhere in the metro area. Figure 2.10 shows that tax inequality among Tokyo’s 23 central city wards and 26 suburban cities follows the same cyclical trend pattern as income inequality. Figure 2.10 also shows that tax inequality among Tokyo’s districts is much higher than income inequality, indicating that Tokyo’s income tax system is progressive and redistributive. The correlation between per capita income and per capita tax was indeed very high (0.96) in 2005. Responsibility for public investments, like urban redevelopment projects and the industry cluster plan, and for public services, like kindergartens, day care and elderly centers, is divided between TMG and Tokyo’s wards and cities and coordinated by the central state. Tokyo’s wards and cities take responsibility for running kindergartens, for example, while TMG approves and enforces the educational standards and regulations. These services are provided rather evenly throughout the Tokyo metro area, although a shortage of day care and elderly centers plague some parts of the central city. When redevelopment projects are legally required to provide public services, like day care centers, TMG provides the private developer with tax subsidies. Tokyo’s younger, middle income families continue to want public services but they also want more choice among offerings and now seem to favor a combination of market choice and government oversight. 30  Japan’s corporate tax is the highest among OECD countries. For example, Japan’s statutory rate for 2010 was 39.5 percent but effective rate for 2008 went up to 46.2 percent. This is quite a contrast with the US where statutory rate was 39.2 percent but effective rate came down to 27.1. The OECD average rate for effective rate was 24.9 (Kocieniewski 2011). 31  Japan has decentralized its tax system recently. Local governments, like Tokyo’s 26 cities, have received more tax gathering authority from the state. But Tokyo’s 23 wards have a special status. TMG collects taxes on behalf of the 23 wards and reallocates the revenues back to them according to their public service needs.

Residential Income Inequality in Tokyo

61

Figure 2.10 Tax and income inequality among Tokyo’s 49 local districts (23 central city wards plus 26 suburban cities), 1970–2005

Note: Resident income tax consists of TMG tax and municipal tax (wards and cities). Individuals and families who live in Tokyo have to pay a proportional tax according their income and a flat tax. Source: TMG Bureau of Taxation, Juminzei (Resident Income Tax) in various years and Table 2.1.

Centralized public education  The centralized public education system is another factor responsible for the relative absence of class-based residential segregation in Tokyo. Private primary schools are rare in the city: only 3.8 percent of elementary pupils attend private schools, as compared to 22 percent of junior high and 53 percent of high school students (TMG Board of Education 2008).32 Although student performance varies between and within schools considerably,33 Japan’s uniform national curriculum emphasizes that children’s ability comes from hard work not hereditary intelligence. The uniform curriculum has been heavily criticized of late but it generated an educated and highly productive work force, helping to spur national economic growth for several decades after World War II. The high level of uniform basic education is still valued by Japanese firms who continue to rely on their own training programs to turn high school and college graduates into skilled employees. Unlike American urban areas where rich municipalities can attract the best teachers by paying higher salaries, and thereby advantage children in their district relative to poorer neighborhoods where teachers’ salaries are lower, teachers’ salaries are uniform among public schools in Japan. Responsibility for the total cost of Tokyo’s primary and junior high facilities is divided among the central state (23 percent), TMG (60 percent) and wards and suburban governments (17 percent). Teachers’ salaries are paid by the central state (38 percent) and by TMG (62 percent) (TMG Board of Education 2008). There is little difference between central city and suburbs in education facilities, quality of education received, or student achievement. Tokyo had 1,314 primary schools for 562,886 children in 2007. The number of students per teacher was 19.1 in Tokyo as a whole. Within the central city, the number of students per teacher was 17.2 in Chiyoda-ward, one of the highest income areas, and 19.2 in Adachi-ward, the lowest income area in the city (TMG Board of Education 2009). 32  According to OECD’s international study on student achievements, the quality of education between private and public sectors in Japan was the same (OECD-PISA 2010). 33  For instance, variation in an OECD measurement of reading performance between schools is 58.7 in Japan, higher than the average for all OECD countries of 42 (OECD-PISA 2010b).

62

Residential Segregation in Comparative Perspective

Senior high schools are the responsibility of prefecture governments in Japan. TMG provides wards and cities with a uniform quality of senior high school education. Comparison of two public high schools illustrates this point. Mita High School is located in Minato, one of the four central wards, while Kokubunji High School is located in Kokubunji, a suburban city. The two schools have nearly identical student enrollments, number of teachers, expenditures on facilities and curriculum. TMG funds about 90 percent of costs in both schools.34 Public housing policy  Public housing has never had much bearing on spatial income inequality in Tokyo. To begin with, public housing plays a small role in the housing market.35 Second, Tokyo’s public housing extends to middle as well as to low income families. In the suburbs, where the price of land is considerably lower than in the central city, public housing is concentrated in cities where public housing was a component of new town projects begun in the 1960s and 1970s. Public housing is spread throughout the central city36 (TMG Bureau of Housing 2006). Tokyo’s mix of low and middle income residents in public housing is part of the historical and ideological legacy of the developmental state. Rich and poor have tended to share the same urban space since the Meiji era when the feudal class system was abolished. Even during the pre-World War II years, when Tokyo was regarded as a class-stratified society, rich and poor tended to live side by side. In Seidensticker’s authoritative history of central Tokyo before World War II, the “High City,” which refers to the central city excluding the old downtown (the “Low City”), has always accommodated, side by side, the extremely well-placed and the extremely disadvantaged (Seidensticker 1983: 241). The single exception was the Denenchofu neighborhood in Otaward. The Toyoko Land Company, owner of huge tracts in Tokyo’s southwestern suburbs and in neighboring Yokohama city, built a railway, the Toyoko line, to spur development, and incidentally, to make Shibuya the thriving center that it is today. The Toyoko Land Company also developed Denenchoufu, a “city in the fields” (Seidensticker 1983: 49). Denenchoufu became an exclusive residential enclave for the rich in the 1920s. Progressive income tax hardly existed in the preSecond World War period and capitalists could reinvest almost all of their incomes for further capital accumulation (Moriguchi and Saez 2008: 728). But Denenchoufu today, consisting of nine chomes spread over Ota and Setagaya wards, is a socially mixed neighborhood differing little from the rest of the central city.37 TMG’s public housing provision was a component of the central state’s public housing policy until 2005. Local governments like TMG are now responsible for public housing but they continue to have cooperative relations with central state agencies like the Urban Renaissance Agency and other public corporations. TMG allocates public housing throughout Tokyo, working with the 23 wards and 26 city governments. Even the four core wards, with the highest housing land price and per capita incomes in the metro area, contain public housing. Since 1997, as a means of redressing residential population decline in the downtown, TMG and central core ward governments have 34  The number of students was 878 in Mita and 973 in Kokubunji; the number of teachers was 69 in both schools; the number of classes 26,250 and 25,200; total cost 915,523,000 yen and 1,056,945,000 yen; and per student cost 1,043,000 yen and 1,086,000 yen; percent burden in per student cost 9.9 percent and 10.2 percent. TMG coverage of costs was 89.7 percent and 88.8 percent (TMG 2004). 35  Tokyo’s public housing (owned by local governments and Urban Renaissance Agency and other public corporations) consisted of only 8.4 percent of the total housing stock in Tokyo in 2008 (MIC 2010). 36  TMG’s 320,000 public housing units are spread throughout the central city. Measured by coefficient of variation, the distribution inequality among Tokyo’s 49 districts is 0.444 in 2005. 37  Denenchoufu’s 9 chomes had higher percentages in managerial jobs than the rest of 2,970 chomes in 2005. But only one chome exceeds over 50 percent of P&M concentration as indicated in Table 2.5.

Residential Income Inequality in Tokyo

63

offered subsidized public housing to qualified young couples who could otherwise not afford to live in the high priced central core. The four core wards have reversed the trend in population decline and are facing shortage of kindergartens and daycare centers today. The welfare state  The welfare state in Japan provides basic services such as unemployment insurance, aid to lower income families, universal health care, and long-term elderly care system. Japan’s state social spending is low compared with Western Europe’s, but the Japan’s welfare state is far more developed today than the corporatist welfare state system described by Esping-Andersen (1990) two decades ago. Japan’s welfare state cannot be measured by state social spending only. As in industrial and economic policy areas, the state guides the private sector in the provision of longterm care for the elderly and health care without increasing state social spending by coordinating stakeholders such as universities, local governments, firms, professional and industry associations, and other NGOs (Fujita and Hill 2012). As described earlier, the welfare state in Japan does not play a major role in after-market income distribution. The Japanese political economy rather emphasizes pre-market distributional mechanisms in which the developmental state plays the major role. State policies, for instance, emphasize a full employment strategy through a fiscal investment policy that redistributes tax revenues from tax rich metropolitan areas to poor rural areas, public policy assistance that enables big firms to maintain their workforce and the corporate welfare system during economic downturns, and subsidies that help maintain a myriad of otherwise uncompetitive small businesses. The developmental state’s full employment policy has thus played a surrogate role for the welfare state. But young people have likely suffered most from Japan’s welfare state system. Two decades of slow economic growth have created a generation with substantial numbers who have missed out on full time employment opportunities. Without full time employment and often living in their parent’s home, they have received few benefits from the welfare state. In this sense, families bear the cost of their unemployed young. The un-and-under-employed younger generation poses serious issues for the future of Japanese society. TMG and the state have belatedly begun some social programs that could help reduce the gloomy situations facing the young but their positive effects may come too late. Social Norms and Culture Collectivism  The Japanese preference for collectivism over individualism also contributes to the relative absence of class segregation in Tokyo. Collectivism is well illustrated in the corporate workplace where pursuing one’s self-interest at the expense of others is discouraged. In the public school system, a standardized curriculum is given to all children. Japanese believe that every child can achieve high performance if sufficiently encouraged to make the effort. In R&D centers, employees work in a team environment where individual invention is shared collectively and praise from colleagues is more highly regarded than monetary rewards (Fujita and Hill 2005). Large companies often blur the demarcation between blue and white collar work, call upon everyone to contribute to corporate productivity and performance, and distribute profits across the entire workforce in the form of bonuses.38 The collectivist spirit has its own perils and weaknesses. It does not extend to women. Japan remains a sexist society. Moreover, collectivism has not prevented a lost decade for many young people who continue to suffer high unemployment. The homeless are visible in Tokyo. And the 38  According to a survey by Japan Productivity Center in 2006, 42 percent of newly appointed corporate

64

Residential Segregation in Comparative Perspective

social welfare system is weak. But the collectivist culture has fostered a social consciousness that everybody should be treated the same. Urban residents are particularly sensitive to inequality when they see unemployment rates and income inequality go up in their neighborhoods. Their sensibility finds expression in government research and electoral politics. The popularity of Prime Minister Junichiro Koizumi went down, for example, exactly because people perceived that his allegedly neoliberal policy widened social inequality. LDP administrations that came after Koizumi promised to address the inequality issue and to restore some of the re-distributional public policies that Koizumi abolished during his tenure (Nikkei-Net 2006). The new DPJ administration also came to power campaigning on social inequality issues. In short, collectivist culture reduces Japan’s tolerance for inequality. Conclusion We have argued that spatial income inequality in Tokyo does not reflect class- based residential segregation. Tokyo’s neighborhoods and political sub-communities are not mechanisms for class reproduction. The income gap between the four core wards and districts in the rest of the metro area is induced by Tokyo’s functional primacy in Japan’s urban system, and by the spatial concentration of the functional power centers in these core districts. Organizations of all kinds desire proximity to these national centers of political, economic and cultural power, driving up land values, rents and home prices, and rendering these areas prime locales for real estate speculation. Tokyoites wishing to live in the core areas must pay a steep premium and endure radical fluctuations of their fortunes according to real estate booms and busts. High per capita incomes notwithstanding, residents in the central wards receive the same level of public goods and serves as Tokyoites living elsewhere in the city. And while more high income people live in the core, neighborhoods in the central wards are not exclusive preserves for the rich; they contain much the same mix of occupational groups as communities throughout the rest of Tokyo. One must look to the character of Japan’s political economy to understand why spatial income inequality does not translate into residential class segregation. Tokyo is nested in Japan’s developmental form of capitalism, which, among other things, has rested on the belief that companies should function like communities. Institutions like the compressed wage system and norms valuing collective effort militate against occupational segregation. Although some state policies create real estate booms and busts and intensify spatial income inequality, the state tends to prevent spatial income inequality from developing into class segregation through fiscal investment, education, housing and tax policies. Given the relatively uniform distribution of public goods and services among Tokyo’s neighborhoods, Tokyoites choose where to live according to commuting time to workplace, access to commercial facilities, and stage in the life cycle, not according to the class attributes of their neighbors. In sum, class-based residential segregation does not appear to be growing in Tokyo.

executives chose ‘employees’ when they were asked the question, “whose interest is the most important to you?” followed by customers and shareholders (Japan Productivity Center 2006).

Residential Income Inequality in Tokyo

65

References Dreier, P., Mollenkoph, J. and Swanstrom, J. 2004. Place Matters: Metropolitics for the Twentyfirst Century. Kansas: University Press of Kansas. Second revised edition. Esping-Andersen, G. 1990. The Three Worlds of Welfare Capitalism. Princeton, NJ: Princeton University Press. Fielding, A.J. 2004. “Class and Space: Social Segregation in Japanese Cities.” Transactions of the Institute of British Geographers 29(1), 64–84. Fujita, K. 1991. “A World City and Flexible Specialization: Restructuring the Tokyo Metropolis.” International Journal of Urban and Regional Research 15(2), 269–84. Fujita, K. 2003. “Neo-Industrial Tokyo: Urban Development and Globalization in Japan’s StateCentered Developmental Capitalism.” Urban Studies 40(2), 249–81. Fujita, K. 2011. “Financial Crises, Japan’s State Regime Shift and Tokyo’s Urban Policy.” Environment & Planning A 43(2), 307–27. Fujita, K. and R.C. Hill, eds 1993. Japanese Cities in the World Economy. Philadelphia: Temple University Press. Fujita, K. and Hill, R.C. 1997. “Together and Equal: Place Stratification in Osaka.” In The Japanese City, eds P. Karan and K. Stapleton. Lexington, Kentucky: University of Kentucky Press, 106– 33. Fujita, K. and Hill, R.C. 2005. Innovative Tokyo, World Bank Policy Research Working Paper WPS 3507, February. Social Science Research Network (SSRN). http://papers.ssrn.com/sol3/ papers.cfm?abstract_id=660088. Fujita, K. and Hill, R.C. 2008. Tokyo’s Urban Redevelopment Projects and the Post-Developmental State. Unpublished paper. Presented at the ISA First World Forum of Sociology, Barcelona, Spain, September. https://msu.edu/user/fujitak/Tokyo’sredevelopmentprojects.pdf Fujita, K. and Hill, R.C. 2012. “Industry Clusters and Transnational Networks: Japan’s New Directions in Regional Policy.” In Locating Neoliberalism in East Asia: Neoliberalizing Space in Developmental States, eds B. Park, R.C. Hill and A. Saito. Oxford: Wiley Blackwell, 27–58. Gordon, A. 1985. The Evolution of Labor Relations in Japan. Cambridge, MA: Harvard University Press. Gordon, A. 2001. The Wages of Affluence: Labor and Management in Postwar Japan. Cambridge, MA: Harvard University Press. Hara, J. and Seiyama, K. 1999. Shakai kaiso: Yutakasa no naka no fubyodo [Social Stratification: Inequality in an Affluent Society]. Tokyo: University of Tokyo Press. (English version published in 2006 as Inequality amid Affluence: Social Stratification in Japan. Melbourne: Trans Pacific Press.) Hill, R.C. and Fujita, K. 1995. “Osaka’s Tokyo Problem.” International Journal of Urban and Regional Research 19(2), 181–93. Inagami, T. and D.H. Whittaker, eds 2005. The New Community Firm: Development, Governance and Management Reform in Japan. Cambridge, UK: Cambridge University Press. Ishida, H. and Slater, D.H. eds 2009. Social Class in Contemporary Japan. London: Routledge. Jacobs, A.J. 2003. “Embedded Autonomy and Uneven Metropolitan Development: A Comparison of the Detroit and Nagoya Auto Region, 1969–2000.” Urban Studies 40(2), 335–60. Jacobs, A.J. 2005. “Has Central Tokyo Experienced Uneven Development? An Examination of Tokyo’s 23Ku Relative to American Large Urban Centers.” Journal of Urban Affairs 27(5), 521–55.

66

Residential Segregation in Comparative Perspective

Kocieniewski, D. 2011. U.S. Business Has High Tax Rates But Pays Less. New York Times, May 2, A1. Japan Productivity Center. 2006. Shinnin yakuin no sugao ni kansuru chosa [Survey on Profiles of Newly Appointed Executives], Tokyo, June. Marcuse, P. and van Kempen, R. eds 2000. Globalizing Cities: A New Spatial Order? London: Blackwell. Matsumoto, K. 1991. The Rise of the Japanese Corporate System. London/New York: Kegan Paul International. Ministry of Finance Policy Research Institute. 2004. Japan’s Social Stratification and Consciousness. Tokyo: Ministry of Finance Policy Research Institute. Ministry of Health, Labor and Welfare (MHLW). 2004. Shotoku saihaibun chosa hokoku based on Kokumin Seikatsu Jittai Chosa [Report on the 2002 Income Redistribution based on National Livelihood Surveys], Tokyo. Ministry of Health, Labor and Welfare (MHLW). 2007. 2005 shotoku saihaibun chosa hokokusho [Report on Income Redistribution in 2005]. Tokyo. Ministry of Health, Labor and Welfare (MHLW). 2009. Heisei 20 nen chingin kozo kihon toukei chosa [2008 Basic Statistical Survey on the Wages Structure], Tokyo. Ministry of Internal Affairs and Communication (MIC). 2002. Nihon no shotoku kakusa based on Zenkoku Shohisha Jittai Chosa [Income Inequality in Japan based on National Consumer Surveys]. http://stat.go.jp/data/zensho/topics/1999_1.htm [accessed: May 2006]. Ministry of Land, Information, Transport and Tourism (MLIT). 2010a. The 2010 White Paper on Land. http://tochi.mlit.go.jp/hakusyo/h22tochi.pdf [accessed: August 2010]. Ministry of Land, Information, Transport and Tourism (MLIT). 2010b. Chika koji (Published land value) in 2010. http://tochi.mlit.go.jp/01_04.html [accessed: August 2010]. Ministry of Land, Information, Transport and Tourism (MLIT). 2009. The 2009 White Paper on Land. http://tochi.mlit.go.jp/h20hakusho/chapter3/chapter_03_eng.html [accessed: August 2010]. Ministry of Land, Information, Transport and Tourism (MLIT). 2007. Real Estate Securitization Survey, Tokyo. Moriguchi, C. and Saez, E. 2008. “The Evolution of Income Concentration in Japan, 1886–2005. Evidence from Income Tax Statistics.” Review of Economics and Statistics 90(4), 713–34. Musterd, S., Murie, A. and Kesteloot, C. eds. 2006. Neighborhoods of Poverty: Urban Social Exclusion and Integration in Europe. London: Palgrave Macmillan. New York Times 2010. The Real Say on Pay. Editorial. http://nytimes.com/2010/09/02/ opinion/02thu3.html?_r=1&th&emc=th [accessed: September 2010]. Nikkei-Net. 2006. Koizumi Rebuts Issue Of Economic Disparities Under His Reform, April 25. Nikkei-Net. 2009. Kichijoji Still Hottest Address Among Tokyoites: Poll, March 11. OECD. 2006. Economic Survey of Japan. Paris, July. OECD Programme for International Student Assessment (PISA). 2010a. PISA 2009 Results: What Students Know and Can Do. http://www.oecd.org/edu/pisa/2009 [accessed: 25 March 2011]. OECD Programme for International Student Assessment (PISA). 2010b. Figure II.5.1: Differences between schools and within schools in “Chapter 5 School Systems and the Impact of Socioeconomic Background in Volume II: Overcoming Social Background: Equity in Learning Opportunities and Outcomes.” In PISA 2009 Results. http://www.pisa.oecd.org/document/24 /0,3746,en_32252351_46584327_46609752_1_1_1_1,00.html [accessed: 22 October 2011]. Ohtake, F. 2005. Nihon no fubyodo [Inequality in Japan]. Tokyo: Nihon Keizai shimbunsha.

Residential Income Inequality in Tokyo

67

Ohtake, F. and Saito, M. 1998. “Population Aging and Consumption Inequality in Japan.” Review of Income and Wealth 44(3), 361–81. Ono, H. 2010. “Lifetime Employment in Japan: Concepts and Measurement.” Journal of the Japanese and International Economies 24(1), 1–27. Ota, K. 2006. Is Japan’s Wage Inequality Small? ESRI Discussion Paper Series, No. 172. Economic and Social Research Institute, Japanese Government Cabinet Office, Tokyo. Sato, T. 2000. Fbyodo shakai Nippon [Unequal Society Japan]. Tokyo: Chuo Koronsha. Seidensticker, E. 1983. Low City, High City. Tokyo: Charles Tuttle. Swanstrom, T., Casey, C., Flack, R., et al. 2004. Pulling Apart: Economic Segregation Among Suburbs and Central Cities in Major Metropolitan Areas. Living Cities Census Series. Washington, DC: The Brookings Institution. Tachibanaki, T. 1998. Nihon no Keizaikakusa [Economic Inequality in Japan]. Tokyo: Iwanami. (English version published in 2005 as Confronting Income Inequality in Japan. Cambridge, MA: The MIT Press. Tachibanaki, T. 2010. The New Paradox for Japanese Women: Greater Choice, Greater Inequality. Tokyo: International House of Japan. Tokyo Metropolitan Government (TMG). 2002.Tokyo: TMG Housing Master Plan, 2001–2015. Tokyo Metropolitan Government (TMG). 2006. Tomin no kurashimuki [Tokyoites’ Livelihood in 2005]. Tokyo. TMG Board of Education. 2005. Public Schools in Tokyo. Tokyo. TMG Board of Education. 2008. The 2008 Report on Local Education Cost. Table 5: Total Costs on School Education. http://kyoiku.metro.tokyo.jp/toukei/21chikyohi/2–5toukeisiryou.pdf [accessed: September 2010]. TMG Board of Education. 2009. List of Tokyo’s Public School. http://kyoiku.metro.tokyo.jp/ toukei/21kouritsu/21mokuji.htm [accessed: September 2010]. TMG Bureau of Finance. 2010. Published Land Prices: Average Prices by Land Use and Ward, City, Town and Village. http://www.zaimu.metro.tokyo.jp/kijyunti/22kouji/04heikin.pdf. TMG Bureau of General Affairs. 2010. Registered Foreigners. http:// toukei.metro.tokyo.jp/ gaikoku/ga-index.htm [accessed: January 2010]. TMG Bureau of Housing. 2006. Tokyo: Tokyo Metropolitan Government Housing Master Plan, 2006–2015. TMG Bureau of Industrial and Labor Affairs. 2008. Tokyo no sangyo to koyo. Tokyo: Industry and Employment in Tokyo. TMG Bureau of Industrial and Labor Affairs. 2009. Tokyoto no chingin, rodo jikan oyobi koyo no ugoki [Tokyo’s rends on Wages, Working Hours and Employment]. Tokyo. TMG Bureau of Taxation. 2005. Tokyo to zeimu tokei nempo [2005 Annual Report on Tokyo’s Taxation]. Tokyo. TMG Bureau of Urban Development. 2005. Urban Redevelopment Projects. http://toshiseibi. metro.tokyo.jp/bosai/sai-kai.htm [accessed: January 2010]. TMG Bureau of Urban Development. 2007. Tokyo’s Urban Development. Tokyo. Tokyo Stock Exchange. 2007. The 2006 Survey on Employee Share Holdings. Tokyo. United Nations. 2009. Human Development Report 2009. http://hdr.undp.org/en/media/ HDR_2009_EN_Complete.pdf [accessed January 2010]. Vogel, S.K. 2006. Japan Remodeled: How Government and Industry are Reforming Japanese Capitalism. Ithaca, NY: Cornell University Press. Waley, P. 2000. “Tokyo: Patterns of Familiarity and Partitions of Difference.” In Globalizing Cities: A New Spatial Order? eds P. Marcuse and R. van Kempen. Oxford: Blackwell, 127–57.

68

Residential Segregation in Comparative Perspective

Wilkinson, R. and Pickett, K. 2009. The Spirit Level: Why More Equal Societies Almost Always Do Better? London: Allen Lane. Wilson, W.J. 1987. The Truly Disadvantaged: The Inner City, the Underclass and Public Policy. Chicago: University of Chicago Press. Wilson, W.J. 2009. More than Race: Being Black and Poor in the Inner City. New York: W.W. Norton.

Chapter 3

The Impact of Housing Tenure on Residential Segregation in Beijing, China John R. Logan and Limei Li

Introduction The spatial structure of metropolitan areas of China is being reorganized as a result of economic expansion and market reform. This study uses 2000 census data from Beijing to identify the main dimensions and causal processes of residential segregation. One component is mainly descriptive, using segregation indices as well as GIS maps at the neighborhood (jiedao) level to identify the zones of concentration of residents by personal characteristics such as occupation, education, and migration status. The second and major component will examine the causal processes at the individual level that place certain people in certain types of zones. The chapter will demonstrate a novel method of estimating OLS regression models case despite the lack of geographic identifiers in the publicly available census microdata, showing that how people gained access to housing— that is, their type of housing tenure—is the key predictor of where they live. The Impact of Housing Tenure on Residential Segregation in Beijing, China China is in the midst of transition from a central planned socialist economy to a market economy, with dramatic consequences for the spatial structure of metropolitan areas. There is wide agreement that post-reform Chinese cities are becoming more socially stratified and spatially segregated, but relatively little empirical research has been done on this question. The main effort has been to use census data to depict the evolution of urban social space in the factorial ecology tradition (e.g. Xu et al. 1989, Sit 1999, Feng and Zhou 2003a, Li and Wu 2006, Feng et al. 2007, Li et al. 2010). Less attention has been given to assessing the extent or causes of the social segregation that these patterns imply. We will examine residential segregation in urban China along three axes: migrant status, socioeconomic status and housing tenure. Highlighting these dimensions reflects a major difference between China and many Western countries. Race/ethnicity may be the single most important determinant of residential segregation in the lengthy history of research on segregation in the West. Ethnical/racial enclaves can also be found in some Chinese cities, such as the Hui community in Beijing (Zhou et al. 2002, Liang 2003, Zhou and Ma 2004), Uyghur community in Beijing (Baranovitch 2003, Yang and Wang 2008), foreign gated communities in Beijing (Wu and Webber 2004), and African enclaves in Guangzhou (Li et al. 2009). However, ethnic residential segregation is not a dominant factor in most Chinese cities, given that 96 percent of the urban population is Han according to the 2000 census (Huang 2005, Li and Wu 2008). China’s “minorities” are highly concentrated in the interior provinces. It is in those regions of the country that ethnicity is more likely to play a role in urban space, especially as migrations of Han settlers create more diverse

70

Residential Segregation in Comparative Perspective

populations there. Our research, though, is limited to Beijing, and our results are intended to apply mainly to the large and rapidly developing coastal cities. In some countries ethnic minorities are created through immigration. China has experienced very limited immigration, but migration across regions and between the countryside and city has increased dramatically over the past 30 years. Its consequences are mediated by China’s registration (hukou) system instituted in 1958. The key social divide in Chinese cities is between local people with urban hukou and migrants with agricultural hukou. Migrants are typically treated as a transient, “floating” population in cities. They are disadvantaged by state policies regarding access to housing, job, education, and other social welfare services (Wu 2002, 2004, Fan 2002, Li 2006, Li, Li and Chen 2010), and they are segregated economically and socially from urban residents (Chan 1996, Solinger 1999a, Wang et al. 2002, Zhou and Cai 2008, Zhang and Wang 2010). Indeed, their position is similar to that of undocumented immigrants in the West (Solinger 1999b). Estimates of the floating population show large increases (though precise definitions have varied over time): 6.6 million in 1982, 21.4 million in 1990, 70.7 million in 1995, 105 million in 2000, 147.4 million in 2005, and 221 million in 2010. Migrants now constitute about one third of the total urban population. Partly because of restrictions on where they are allowed to live and partly due to their social and economic disadvantages, they have the fewest housing options and live in the worst housing conditions (Wu 2002, Logan et al. 2009). They are also residentially segregated. Migrants tend to concentrate in certain areas of the cities—the inner suburb or urban periphery—as found in Beijing, Shanghai and Guangzhou (Feng and Zhou 2003b, Zhou and Logan 2008, Wu 2008, Yuan et al. 2007). More specifically, migrants usually live in urban villages or migrant enclaves based on place of origin, factory dormitories, and transient construction sites (Ma and Xiang 1998, Fan and Taubmann 2002, , Wu 2002, Zhang et al. 2003). Class stands out as another important factor causing residential segregation to the extent that higher socio-economic status can be translated into better neighborhood locations. As China transforms to a market-oriented economy, social scientists expect socioeconomic status to become a more important source of urban residential segregation. Only the most affluent, typically private business owners and some professionals and high-level public officials, can afford single-family suburban villas and upscale apartments in well-serviced downtown zones (Huang 2005). The new affluent class in Chinese cities has begun to show similarities with their western counterparts and cluster in “zones of affluence” (Gaubatz 1999, Hu and Kaplan 2001) and western-style gated communities (Wu 2004a, 2010). Since income is not reported in the 2000 census, we use education and occupation in lieu of income to examine its effect on residential segregation. Education attainment is increasingly related to income level as well as social status in post-reform China (Tang and Parish 2000). Migrants who earn a college degree often can gain a local urban hukou, which can enhance their social status in the city (Li and Wu 2008, Li, Li and Chen 2010, Zhang and Wang 2010). Besides migrant status and class, the third determinant of residential segregation in Chinese cities is related to the country’s socialist legacy. A housing census conducted in 1985 showed that state-owned work units controlled 70.4 percent of all floor area in cities; municipal housing bureaus controlled 11.5 percent; urban collectives owned and managed 10.3 percent, and private ownership accounted for just 10.3 percent (Li 1995). In this period social class indirectly affected residential patterns by virtue of the fact that people’s work units more or less determined their residential location, since they were allocated housing by their employers—often within or near the work unit compound (Yeh et al. 1995). Note that not everyone could qualify for public housing. Migrant workers with agricultural registration in their home province did not qualify. In addition, because

The Impact of Housing Tenure on Residential Segregation in Beijing, China

71

there was a housing shortage, not all urban residents (especially young adults and newlyweds and employees of work units with fewer resources to build housing compounds) were able to obtain public housing. Also rooted in the socialist period was segregation by housing tenure. Public rental housing was concentrated in central zones; private rental housing, mainly serving migrant workers, was found on the periphery of the metropolis; self-built housing, the prerogative of local villagers, was found in rural zones. The transition from socialism created two new forms of housing tenure: public purchase (subsidized sale of apartments to former public housing tenants) and new construction for sale at market prices. The incidence of home ownership jumped from 20 percent in 1980 to 24 percent in 1990—at that time still mainly a carryover from older private sector housing—to 72 percent in 2000, to 82 percent in 2005 and further to 89 percent in 2010. This jump mainly reflects the sale of previous public housing that turned former tenants into homeowners and the evacuation of redevelopment zones (whose residents typically became apartment owners in new developments outside the city). Of course only people who had managed to qualify for public housing in the socialist period were in position to purchase their rental unit or to receive compensation for displacement. Increasing homeownership also reflects the early stages of growth of a “commodity housing” market for more affluent residents. Because the location of these several types of housing has a spatial pattern, there is now significant housing tenure-based residential segregation (Li and Wu 2008). In short, throughout urban China people are being sorted into neighborhoods of many types: migrants’ enclaves, dilapidated neighborhoods and redevelopment zones in the central city, work unit compounds, commodity housing developments, and luxury townhouse and villa estates. The sorting process mainly follows the inter-related dimensions of residence status, social class, and housing tenure. We will examine each of these separately, but in the end we will argue that the main line of segregation in Beijing is based on housing tenure. Other factors become important by virtue of their close association with tenure types. The Case of Beijing Beijing was chosen as the national capital after the founding of the People’s Republic of China in 1949. At that time it had a population of 1.6 million, about 1.4 million of whom lived in the old city, i.e. the area within the present-day Second Ring Road encircling Dongcheng, Xicheng, Chongwen and Xuanwu districts (Sit 1999). The inner city core and its adjacent area have been built up to accommodate the administrative staff of the central government, military and party headquarters and their dependents. The northwest suburb (Haidian) was developed into a zone of higher education and research. Industrial development took place in the southwestern suburban areas on the downwind sides of the inner city. Due to the close linkage between work and residence, workers’ quarters were also found in the same area. The urban master plan of Beijing in 1983 deemphasized industry and in the 1980s heavy industry began to shift to the suburbs, leading also to residential suburbanization (Zhou 1996). Since then the central city has been losing population while the inner suburbs gained. However, the monocentric structure remains largely intact despite the emergence of several population subcenters in the 1990s (Feng and Zhou 2003c). Land price and population density decrease with distance to the city core (Feng and Zhou 2003c, Ding 2004). Beijing in 2000 had a total population of over 13 million and was still growing rapidly. We study an area that is formally defined as “urban” by government authorities, includes a population of 9.2

72

Map 3.1

Residential Segregation in Comparative Perspective

Beijing’s central city, inner suburbs, and adjacent outer suburban zones

Source: Drawn by the authors.

million. We treat the Dongcheng, Xicheng, Chongwen and Xuanwu districts as central city and Chaoyang, Fengtai, Shijingshan and Haidian as inner suburbs (Map 3.1). The map also shows six adjacent outer suburban districts. Table 1 provides an overview of the socioeconomic differences among these districts. The percentage of residents (age six and above) with at least some college education is highest in the central city districts (especially Xicheng), somewhat lower in inner suburbs (with the notable exception of Haidian), and quite low by comparison in adjacent outer suburban districts. The share of workers with government or professional occupations follows the same pattern. The subdistrict (more formally: street offices, towns and townships1 or jiedao/zhen/xiang) is the smallest geographic unit for which census data are available and it is the unit that we rely on in this study. Subdistricts are often too large to distinguish between villa developments, gated communities, work compounds, peasants’ enclaves, and dilapidated neighborhoods that are in close proximity to one another despite serving very different populations. But they can still offer valuable information 1  The governance structure within Chinese cities is usually composed of municipal government, urban districts/counties/county-level city, street offices/towns/townships, and residents’ committees/villagers’ committees. Census is also conducted based on this administrative structure.

The Impact of Housing Tenure on Residential Segregation in Beijing, China

Table 3.1

73

Education and high-status occupations by district, 2000 College educated

Government and professionals

Central city

21.5%

26.5%

Dongcheng Xicheng Chongwen Xuanwu

22.3% 27.0% 13.8% 18.3%

26.9% 29.6% 22.0% 25.1%

Inner suburb

23.9%

23.4%

Chaoyang Fengtai Haidian Shijingshan

20.9% 14.7% 34.1% 17.2%

22.2% 15.6% 30.6% 18.9%

5.9%

9.9%

Mentougou Fangshan Tongzhou Shunyi Changping Daxing

5.7% 5.1% 5.6% 4.1% 12.7% 6.7%

12.1% 9.5% 9.0% 8.2% 13.5% 8.9%

Beijing urban area total

18.3%

19.4%

Adjacent outer suburb

Note: College educated as a percentage of all residents age six and above. Government and professional occupations as a percentage of all employed persons age 16 and above, including residents of group quarters for whom occupation is not reported. Source: 2000 population census.

about socio-spatial differentiation. There are 164 subdistricts in the central city and inner suburbs; an additional 18 subdistricts outside this area are also designated as urban areas. We use standard methods to map population characteristics including residence status, education, occupation, and housing tenure and to calculate measures of residential segregation that can be compared to such measures in other cities. Several variables from the census are used here. Counting residents by their migration status is complex because there are at least two kinds of distinctions to be made (Logan et al. 2009). One is the type of registration (hukou) status that the person has, either agricultural or non-agricultural. Another is whether the person is “local” in origin (i.e., born in Beijing) or registration. The available tabulations allow us to define three categories of people: locals (regardless of their registration status), rural migrants, and urban migrants. The segregation of residents by migrant status is modest. The value of the Index of Dissimilarity (D)2 for rural migrants vs. local residents is only 0.33, while the segregation of urban migrants 2  The index ranges from 0 to 1, giving the percentage of one group who would have to move to achieve an even residential pattern—one where every street office replicates the group composition of the city. Values of 0.6 or above are considered very high. Values of 0.3 to 0.6 are considered moderate level of segregation, while values of 0.3 or less are considered low.

74

Map 3.2

Residential Segregation in Comparative Perspective

Distribution of migrants with urban registration in Beijing in 2000

Source: National Bureau of Statistics of China (2005).

from locals is 0.38. To some extent these low measures are due to the scale of measurement; at a higher level of resolution more segregation would be revealed. But in fact migrants with an urban registration should not be expected to be highly segregated from urban locals because these two groups have had relatively similar levels of access to the housing market. Rural migrants (often referred to as the floating population) are the group with the least options, but their typical location in peripheral zones means that they tend to live near (and often rent from) local rural people. A similar study in Shanghai (Li and Wu 2008), using the much smaller residents committee as the unit of analysis, found a value of D between migrants and locals of only 0.26. The geographic pattern for these groups is illustrated in Maps 3.2–3.3. Map 3.2 shows that urban migrants tend to concentrate in the northwestern sector of inner suburbs. Map 3.3 shows that rural migrants are more dispersed to the outer suburbs. We turn now to the indicator of class standing, education. Education can be treated as ordinal categories; we divide into those with no more than an elementary education, those with up to a high school degree, and those with some college or more. Segregation between people with primary and secondary education is low—0.06. College educated people are more sharply separated, 0.29 between secondary and college and 0.32 between the two extremes of elementary and college. The distinctive residential pattern of people with college education is illustrated in

The Impact of Housing Tenure on Residential Segregation in Beijing, China

Map 3.3

75

Distribution of migrants with rural registration in Beijing in 2000

Source: National Bureau of Statistics of China (2005).

Map 3.4. It shows that subdistricts with more than 30 percent college educated residents were highly clustered in the northern and western inner suburbs, which coincide also with the location of Beijing’s major public universities (Haidian) and its high-tech innovation zone. Their location is surprisingly similar to that of urban migrants, which gives a clue about the social standing of that class of migrants. The findings for occupation involve more categories, but the underlying pattern is not complex. On the one hand, people with agricultural occupations (and there are many, despite our effort to restrict the study to Beijing’s urbanized area) are very highly segregated from all other occupations. Values of D are around 0.70 between farm workers and professionals, clerical workers, and government employees, and (at 0.55) also high between farm workers and manufacturing workers. Segregation across other categories is much lower, 0.29 between professionals and manufacturing workers, 0.34 between government employees and manufacturing workers, but less than 0.25 between all other pairs of occupations. The geographic pattern for government employees is representative of the level of spatial variation. As shown in Map 3.5, government workers tend to cluster in the inner city districts or in the near western inner suburbs.

76

Map 3.4

Residential Segregation in Comparative Perspective

Distribution of people with tertiary education attainment in Beijing in 2000

Source: National Bureau of Statistics of China (2005).

Finally we assess segregation by housing tenure. Table 3.2 reports the tenure composition of housing stock in Beijing in 2000, identifying the categories of ownership and renting that are appropriate to this case. Self-built housing occupied the largest share, 31 percent (note that these data include both urban and rural parts of Beijing). Self-built housing is the norm for the suburban peasants around Chinese cities. Local peasants construct their dwelling in the allocated land lot. Self-built housing is typically larger than housing unit built for urban residents, but its quality and neighborhood facilities are inferior to the housing units in the city. The second largest (28 percent) is public purchase housing, referring to the former public housing provided by work units or by local governments. Since the implementation of housing reform at the end of 1980s, the bulk of public housing was sold to sitting tenants. The third component is public rental housing (25 percent), mainly the private housing from pre-1949 era which was confiscated and nationalized by the governments in the 1950s. As a residual of the old system, it is managed by the municipal government. Its rent levels have been raised but it is still less expensive than market rent. It is diminishing in number but seems unlikely to disappear entirely. Private purchase, economic purchase (modestly subsidized to promote new housing for moderate income families) and private rental housing together comprised only about 12 percent of the housing stock.

The Impact of Housing Tenure on Residential Segregation in Beijing, China

Map 3.5

Distribution of government employees in Beijing in 2000

Table 3.2

Housing tenure composition in Beijing, 2000

77

Source: National Bureau of Statistics of China (2005).

Central city Inner suburb Outer suburb Total

Public rental

Public purchase

Private purchase

Economic purchase

Private rental

Self-built

Others

52.7% 30.1% 6.3% 25.1%

30.6% 39.0% 12.8% 27.9%

0.8% 1.9% 8.0% 4.0%

1.0% 2.0% 2.9% 2.1%

3.1% 9.0% 4.7% 6.4%

7.9% 14.4% 62.9% 31.2%

4.0% 3.6% 2.5% 3.3%

Source: National Bureau of Statistics of China (2005).

By the end of 2000, housing distribution in Beijing featured a concentric pattern as shown in Map 3.6. Public rental housing was the primary housing type in the central city because most private housing from the pre-1949 era was in the central city and almost all was nationalized into some forms of public rental housing in the 1950s and managed by housing bureau. Even after the housing reform, the governments have been reluctant to restore the ownership to its original

78

Map 3.6

Residential Segregation in Comparative Perspective

Composition of housing tenure in Beijing in 2000, showing the tenure type with the largest share in each street office

Source: National Bureau of Statistics of China (2005).

owners (Hsing 2010). It was surrounded by a zone of housing built in the 1980s and 1990s as work unit housing which was allocated to their employees with a nominal rent, and these units were subsequently very likely to be purchased by the tenants. In much of the periphery self-built housing is the most common type (a prerogative of local people with agricultural registration). The periphery also includes newer and more heterogeneous housing (including commodity housing estates, suburban villas, and economic housing), all in the larger category of private purchase and also areas of private rental housing (owned by local villagers but occupied by rural migrants). But in very few subdistricts are these categories, even combined, the primary type. The strong association between location and either timing of construction or hukou-related housing types contributes to the very high levels of segregation by housing tenure. Mirroring segregation of farm workers, people in self-built housing are highly segregated from public purchase (0.67), somewhat less segregated from people in and public rental housing (0.58) and private rental housing (0.57), and much less segregated from private purchase housing (0.36) which often is found in lower cost peripheral zones. Public rental and public purchase housing are intermixed (D = 0.31). But these two types have levels of segregation in the 0.50–0.60 range from private purchase and private rentals.

The Impact of Housing Tenure on Residential Segregation in Beijing, China

79

Again we can compare these findings with those of Li and Wu (2008) for Shanghai, where segregation between public rental housing or housing purchased by tenants vs private purchase housing was also in the range of 0.50–0.60. Processes of residential sorting in Beijing How is residential segregation generated? Segregation indices provide some clues about the special importance of types of housing tenure in the Chinese case, which reflect changing public policies that determine how housing is constructed and who has access to it. We have made a case that segregation by migration status, education and occupation, while important in its own right, is highly intertwined with and likely subordinate to segregation by tenure. Another way to examine processes of residential sorting is to estimate at the individual level what kinds of people live where. We do this through locational attainment models, where the spatial outcome is the dependent variable and the traits by which people become segregated are the predictors. Hence the outcome is a locational property while the predictors are at the individual level. To estimate these models would be simple if the census microdata explicitly reported people’s place of residence at a local level. However the Chinese census microdata, like similar publicly available data in the United States, do not include neighborhood identifiers. Alba and Logan (1992) proposed a method to use both microdata and summary tables for neighborhoods to overcome this limitation. Their method has been intensively used to study racial/ethnical residential segregation in the metropolitan areas of the United States (e.g. Alba and Logan 1993, Logan and Alba 1993, Alba et al. 1994, Logan et al. 1996). This is the first attempt to apply this method to China to study the urban residential sorting processes that map individuals into spatial locations. The theoretical reasoning connecting the dependent to the independent variables is that some individual characteristics (X) are presumed to determine access to different kinds of neighborhoods (Y). To estimate parameters of an OLS model linking X to Y requires creation of a correlation matrix to input into the regression program. Correlations among individual-level predictors (X–X) are calculated directly from the individual-level data. Correlations between these predictors and characteristics of neighborhoods (X–Y) are calculated from aggregate data at the neighborhood level using well known procedures for calculating correlations from grouped data. For example, if we have a tabulation of people according to some individual characteristic X for every neighborhood, where we also know the neighborhood’s value for Y, the correlation of X and Y for individuals can be calculated from these grouped data. We use a standard software package (SPSS) to input data to the regression procedure in the form of a correlation matrix, vectors of means and standard deviations, and number of cases. Both individual-level and aggregate data can provide the means and standard deviations. The aggregate data are usually based on a larger sample, often on the entire population, and for this reason we take means and standard deviation from aggregate data. As for the specification of the number of cases (which determines the degrees of freedom for significance tests), we adopt a conservative approach: we use the smaller number of cases present in the microdata. We apply this method to 2000 census data of Beijing. The individual-level data are a 0.1 percent sample of the population enumerated in 2000 Chinese census. At the same time, we also have census data for the same variables aggregated at the street office/town/township level along with the geographic coordinates of the central point in each street office/town/township area. For the first time in 2000, the Chinese census enumerated all the population—both “permanent” and

80

Residential Segregation in Comparative Perspective

“temporary”—in the place they are residing. This is also the first census to count migrants at the street office/town/township level. In this study we examine a single spatial outcome: the distance to the city center (the location of the city hall) from the center of the subdistrict where the individual resident resides. The natural logarithm of the point distance is the dependent variable in the locational attainment models. It is a neighborhood variable and assumed to be constant for residents living in one street office/town/ township. The rationale is that the city center in Chinese city is the most sought-after residential location (Li and Siu 2001, Wu 2004b). Population density, land value and housing price in Beijing gradually decline as the distance from the central city increases (Wang and Zhou 1999, Feng and Zhou 2003b, Ding 2004, Wu et al. 2010). As we will show, there are similar gradients for individuals’ occupational standing and education, clearly suggesting that central location is highly valued in Beijing. Of all dimensions along which metropolitan regions are known to be spatially organized, the center-periphery dimension is the most studied and best understood. Future studies can focus on other residential outcomes. The independent variables include both individual-level characteristics and one set of neighborhood variables, housing tenure. There are three kinds of individual-level variables in the models, and the specific categories used here are constrained by the categories used in the available data sources. The first category has to do with migration and registration status. This includes hukou status, holding agricultural or non-agricultural hukou. It also includes a set of categories related to residential mobility: people with local registration living in their place of registration, local hukou holders who live in a different place (people who have moved within the city), migrants from another coastal province, and migrants from a non-coastal province. Another variable related to residential mobility is the length of time the person has lived in the same subdistrict, and it mainly measures mobility over time within Beijing. Another group of variables are related to socioeconomic status measured by education and occupation. We also include age, gender and marital status as control variables in the models. All individual-level variables are treated in categorical form (as dummy variables) in our models. Finally, to capture the effects of housing structure, we introduce a set of variables on housing tenure (e.g., public rental, etc.). The Chinese census does not report housing tenure for persons in collective households (such as dormitories or construction worksite housing), and so “group quarters” is introduced as another tenure category. What this variable can tell us is crucial— controlling for individual characteristics, how much is spatial segregation on the center-periphery dimension shaped by the spatial pattern of housing tenure? More specifically, how much is the effect of other individual characteristics is mediated through this mechanism? Implementing Issues When we apply this method to investigate the spatially sorting processes in Chinese cities, some issues arise in the use of Chinese census data due to its unique definitions and limitations. First, there are issues related to the aggregate data. The 2000 Chinese census uses a combination of short-form questionnaires (covering 90 percent of the population) and long-form questionnaires (administered to the other 10 percent). The latter asks more detailed information about occupation, migration, and housing conditions. We use the variables from both short-form and long-form tables in our models. The definitions of universe for aggregate data vary across tables. Some tabulations are for individuals, some are for individuals in certain age ranges, whereas others are for households. To estimate correctly a correlation or covariance matrix, it is imperative that the universes be

The Impact of Housing Tenure on Residential Segregation in Beijing, China

81

consistent across variables (Alba and Logan 1992), and this requirement limits the tabulations that can be exploited. The availability of tables can affect decisions about the unit of analysis, placing the analysts in the position of choosing between the most appropriate unit of analysis and the widest set of variables. For analyses of spatial location, the household is probably the most appropriate unit, but theoretically key variables, such as hukou status, place of origin, education, and occupation, are only available in tables for individuals. Actually, tables for household in Chinese census are quite limited (including only housing materials, housing facilities, housing value and rent ranges). Therefore, we base the analyses on individuals. Moreover, some means of representing missing data categories are necessary. Missing data occur because the tables for some variables do not cover the universe of analysis. Census tabulations of education, for instance, do not count persons under six years old, whereas they are included in other tabulations. We extend the education variables by creating a separate category for persons under six, which is not an education category at all but rather a category for missing data. Marital status and occupation are dealt with in the same way. Another important issue is to keep the geographical coverage consistent between aggregate data and individual-level data to ensure that the two levels of data represent the same population. The geographical coverage of Chinese census is based on administrative boundaries, which are subject to criticism (e.g. Zhou and Yu 2004, Chan 2007). Analyses based on the administrative boundary tend to overestimate the size of the city, but analyses entirely excluding the agricultural population undercount the number of inhabitants. The 2000 Chinese census adopted a new definition for “urban” reflecting de facto population density, administrative status and physical features—an average population density of at least 1,500 per km2 and contiguity of the built-up area,3 which affords an opportunity to study Chinese cities in a more meaningful way (Chan 2007). We follow this new definition, so our locational attainment models include only subdistricts (street offices/ towns/townships) defined as city in the aggregate data tables, and we only include the designated “urban” population in the micro-data. Results of Locational Models Individual-level regression analyses that address these questions (locational attainment models) are reported in Table 3.3. Positive coefficients in these models mean that the variable predicts residence farther away from the center of the city. Because distance is logged, coefficients of dummy variables can be interpreted as showing by what percentage a category of the independent variable differs from the reference category. However coefficients that are not statistically significant should be treated as showing no effect at all. Results are presented in two models. Model 1 includes most individual-level predictors. Model 2 introduces housing tenure as an additional variable, in part because tenure itself has been shown to be strongly affected by other characteristics (Logan et al. 2009) and in part because we want to identify the independent impact of this key variable. We focus on three questions: 1) what are the overall effects of other individual predictors in model 1, 2) what is the direct effect of housing tenure, and 3) how does introduction of the tenure change the estimates of the independent effects of individual predictors in model 2? 3  The criteria of city in 2000 census: 1) the whole administrative area of urban district with a population density of 1,500 persons per km2 or above; 2) if the population density of urban district is lower than 1,500 persons per km2, only the district government seat and other street offices under this urban district are treated as city; 3) if the area stated by criteria 2 is encroaching the surrounding town/township, this town/township is also counted as city.

Residential Segregation in Comparative Perspective

82

Table 3.3

Locational attainment model predicting distance (ln) from the city center (N = 9114) Model 1

Model 2

0.343 ***

0.212 ***

Hukou status Non-agricultural hukou a Agricultural hukou Place of origin Local hukou holders living in hukou place a Local hukou holders, living away from hukou place Migrants from coastal provinces Migrants from non-coastal provinces Length of residence Lived here since birth a Moved here 5+ years ago Moved here in last five years

0.032 -0.145 *** -0.093 **

0.037 -0.082 * -0.055

-0.130 *** -0.009

-0.084 *** 0.000

-0.245 -0.100 -0.154 -0.174 0.173 0.051 -0.194 -0.076

-0.264 -0.099 -0.155 -0.162 0.109 0.022 0.199 -0.074

Occupation Manufacturing workers a Government employees Professional, technicians Clerical & related workers Sales and services workers Agricultural workers Student Retired Other and unemployed

*** ** *** *** * *** *

*** ** *** *** * *** *

Education Primary school a Secondary school Some college or more

-0.068 * -0.107 ***

-0.083 ** -0.151 ***

-0.057 0.021 -0.005

-0.023 0.040 -0.002

-0.086 ** -0.096 *

-0.890 ** -0.077 *

Age 60 and above a Under 20 20–39 40–59 Marital status Married a Single Widowed or divorced Housing tenure Public rental a Self-built Public purchase

0.396 *** 0.576 ***

The Impact of Housing Tenure on Residential Segregation in Beijing, China

Model 1

83 Model 2

Housing tenure (continued) Private purchase Private rental Group quarters Constant Adjusted R

2

0.310 *** 0.363 *** 0.246 *** 4.761 ***

4.472 ***

0.056

0.103

Note: a Omitted category; *** significant at 0.001 level; ** significant at 0.01 level; * significant at 0.05 level.

Not surprisingly, hukou status itself is a strong predictor where people live in the city. An agricultural hukou holder lives more than 65 percent farther away from the center than does a non-agricultural hukou holder. Surprisingly, the effect of place of origin shows that—net of hukou status—migrants live more centrally than locals, especially migrants from the more developed coastal provinces. This result makes sense in light of previous findings. Among persons with urban hukou, migrants have been found to be particularly favored in access to housing, possibly because an offer of housing was part of the arrangement that brought them to Beijing (Fan 2002). Among persons with agricultural hukou, locals are mostly villagers in outer suburbs, while migrants seek insofar as possible to find jobs and therefore to live within the urban area. It isn’t possible with these data to distinguish these effects more precisely. The effect of residential mobility within Beijing shows that those who moved five or more years ago live more centrally than those who lived in the same location since birth or those who moved recently. Because such a large share of the population has reached adulthood in the postreform era, the non-movers are of two types. Some are older persons who always lived in the inner core neighborhoods, but their numbers are declining due to relocation forced by redevelopment projects. A larger share are villagers who are unlikely to leave the local lands over which they have use rights, and these people tend to live on the periphery. Recent movers are also more peripheral because (except for people who can afford high-cost new apartments in the city center) they tend to locate in less expensive, more distant neighborhoods (Feng and Zhou 2003b, Zhou and Logan 2008). Intra-urban movement often originates from central city and ends in the inner suburb (Li 2005, Li and Li 2010). Naturally agricultural workers live further from the center than manufacturing workers, but people in all other categories of occupation live more centrally. This relationship has two sources. First, many of the other occupations are higher status than manufacturing workers, and second (especially since public authorities have aggressively sought to relocate factories in the suburbs downwind of Beijing) manufacturing workers are more likely to have jobs on the edge of the city. Higher education is linked with more central location. The size of this coefficient is modest, attenuated because many college educated people (as we saw above) are located in the university/ high technology zone found to the northwest of central Beijing. But it still partly reflects the preference of educated people with more choices (presumably higher incomes) to live closer to the city center. We included age and marital status only as control variables. The effects of age are not significant. Single and widowed or divorced persons are more likely to live closer to the city center. The very strong effect of being unmarried may reflect the willingness of young adults to trade off

84

Residential Segregation in Comparative Perspective

space and privacy for access to jobs and activities in the urban center. The much smaller effect of being widowed or divorced has no clear interpretation. The predictors in model 1 are significant and cumulatively they can have a strong influence on location. Model 2 presents the results of regression analysis including housing tenure. All housing tenure variables are significant and strong predictors of locational outcome; public rental housing (the category omitted from the analysis) stands out as more centrally located than other categories. Adjusted R2 values double from .056 to .103. Equally important, including housing tenure in the model also reduces the effects of some other individual-level variables including hukou status, place of origin, and length of residence. Summary and Conclusions In this chapter, we have described the pattern of residential segregation at the level of subdistricts in Beijing in 2000, providing a baseline against which future development can be evaluated. We then applied locational attainment models to analyze the processes of residential sorting in terms of registration, migrant status, socioeconomic status, and housing structure that shape them. In general, the extent of segregation by migration status and socioeconomic standing is not as high in Beijing as in western cities. But segregation by housing tenure, which directly reflects the policy shift toward marketized housing and built-in advantages inherited from the socialist period, is higher. And housing tenure appears to be the mechanism through which segregation on other social dimensions is maintained. Ethnic/racial segregation in the West is a consequence of three processes: discrimination; economic disadvantage; and individual choice (Johnston, Poulsen and Forrest 2007). The role of discrimination in the West is played in the Chinese case by state policies that dictated who had access to public housing in the socialist era, under what terms public housing was privatized, the special rights of local villagers to build their own housing and the nearmonopoly on rental housing provision to migrants that was exercised by villagers. Marketization has introduced new mechanism of residential sorting through education and occupation, as well as by income (which could not be measured here). Higher educational attainment may place people in the central city and the northwestern sector of inner suburbs, where higher level academic institutions and high-end service job opportunities congregate. Job location also affects location, and we presume that position in the occupation hierarchy strongly influences one’s access to housing and neighborhood resources. Yet in large part our study of residential segregation has echoed the argument that it is necessary to understand the persistence of socialist institutions in order to understand the postreform Chinese cities (e.g. Huang 2005, Li and Wu 2008, Logan et al. 2009). The past 10 years have witnessed the rapid expansion of private market and the further withdrawal of work units from housing provision. The pace of commodity housing construction accelerates even further in the new millennium. Skyrocketing housing prices lure more people to use housing as a means to build wealth. The availability of the 2010 census will provide new clues to how these processes may have affected residential patterns in the last decade. One concluding remark concerns our method, used quite intensively in community research in America but applied here for the first time to Chinese cities. It is well suited for studying locational processes at various geographic scales, such as street office and residents’ committee, with individuals as the unit of analysis and with models specific to different metropolitan areas. The analyses presented here can be replicated in other Chinese cities, and using other locational

The Impact of Housing Tenure on Residential Segregation in Beijing, China

85

characteristics as outcomes, contingent solely on the availability of data on those outcomes at the appropriate geographic level. References Alba, R. and Logan, J.R. 1992. “Analyzing locational attainments: Constructing individual-level regression models using aggregate data.” Sociological Methods & Research 20, 367–97. Alba, R. and Logan, J.R. 1993. “Minority proximity to whites in suburbs: An individual-level analysis of segregation.” American Journal of Sociology 98(6), 1388–427. Alba, R., Logan, J.R. and Bellair, P.E. 1994. “Living with crime: The implications of racial/ethnic differences in suburban location.” Social Forces 73(2), 395–434. Baranovitch, N. 2003. “From the margins to the centre: The Uyghur challenge in Beijing.” China Quarterly 175, 726–50. Chan, K.W. 1996. “Post-Mao China: A two-class urban society in the making.” International Journal of Urban and Regional Research 20(1), 134–50. Chan, K.W. 2007. “Misconceptions and complexities in the study of China’s cities: Definitions, statistics, and implications.” Eurasian Geography and Economics 48(4), 383–412. Ding, C.R. 2004. “Urban Spatial Development in the Land Policy Reform Era: Evidence from Beijing.” Urban Studies 41(10), 1889–907. Fan, C.C. 2002. “The elite, the natives, and the outsiders: Migration and labor market segmentation in urban China.” Annals of the Association of American Geographers 92(1), 103–24. Fan, J. and Taubmann, W. 2002. “Migrant Enclaves in Large Chinese Cities.” In The New Chinese City: Globalization and Market Reforms, ed. J.R. Logan. Malden: Blackwell Publishers, 183– 97. Feng, J. and Zhou, Y.X. 2003a. “The social spatial structure of Beijing metropolitan area and its evolution: 1982–2000” (in Chinese). Geographical Research 22(4), 465–83. Feng, J. and Zhou, Y.X. 2003b. “The latest development in demographic spatial distribution in Beijing in the 1990s” (in Chinese). City Planning Review 27(5), 55–63. Feng, J. and Zhou, Y.X. 2003c. “The growth and distribution of population in Beijing metropolitan area, 1982-2000” (in Chinese). ACTA Geographica Sinica, l58(6), 903-16. Feng, J., Zhou, Y.X., Logan, J.R. and Wu, F.L. 2007. “Restructuring of Beijing’s Social Space.” Eurasian Geography and Economics 48, 509–42. Gaubatz, P. 1999. “China’s Urban Transformation: Patterns and Processes of Morphological Change in Beijing, Shanghai and Guangzhou.” Urban Studies 36, 1495–521. Hsing, Y.T. 2010. The great urban transformation: politics of land and property in China. Oxford: Oxford University Press. Hu, X.H. and Kaplan, D.H. 2001. “The emergence of affluence in Beijing: Residential, social stratification in China’s capital city.” Urban Geography 22(1), 54–77. Huang, Y.Q. 2005. “From work-unit compounds to gated communities: Housing inequality and residential segregation in transitional Beijing.” In Restructuring the Chinese City: Changing Society, Economy and Space, eds F.L. Wu and L.J.C. Ma. London, New York: Routledge, 172–98. Johnston, R., Poulsen, M. and Forrest, J. 2007. “The geography of ethnic residential segregation: A comparative study of five countries.” Annals of the Association of American Geographers. Association of American Geographers 97(4), 713–38.

86

Residential Segregation in Comparative Perspective

Li, L.M. and Li, S.M. 2010. “The impact of variations in urban registration within cities.” in One Country, Two Societies: Rural-Urban Gap in Contemporary China, ed. M. Whyte. Boston: Harvard University Press, 188–215. Li, L.M., Li, S.M. and Chen, Y.F. 2010. “Better city, better life, but for whom? The hukou and resident card system and the consequential citizenship stratification in Shanghai.” City, Culture and Society 1, 145–54. Li, R.J. 2006. “The opportunity and barrier of the status attainment of the migrants: Analysis of the employment structure of migrants settlement” (in Chinese). Chinese Population Science 5, 69–78. Li, S.M. 1995. “Urban housing reform in China”. in Social Development in China, eds. S. M. Li, Y. M. Siu and T.K. Mok, Hong Kong: Hong Kong Education Press, 324–44 (in Chinese). Li, S.M. 2005. “Residential mobility and urban change in China: What have we learned so far?” In Restructuring the Chinese City: Changing Society, Economy and Space, eds F.L. Wu and L.J.C. Ma. London, New York: Routledge, 157–71. Li, S.M. and Siu, Y.M. 2001. “Commodity housing construction and intra-urban migration in Beijing – an analysis of survey data.” Third World Planning Review 23(1), 39–60. Li, S.M., Hou, Q., Chen, S.S., et al. 2010. “Work, home, and market: The social transformation of housing space in Guangzhou, China.” Urban Geography 31, 1–19. Li, Z.G. and Wu, F.L. 2006. “Sociospatial differentiation in transitional Shanghai” (in Chinese). Acta Geographica Sinica 61(2), 199–211. Li, Z.G. and Wu, F.L. 2008. “Tenure-based residential segregation in post-reform Chinese cities: A case study of Shanghai.” Transactions of the Institute of British Geographers 33, 404–19. Li, Z.G., Ma, L.J.C. and Xue, D.S. 2009. “An African Enclave in China: The Making of a New Transnational Urban Space.” Eurasian Geography and Economics 50(6), 699–719. Liang, J.Y. 2003. “From closed to opened: The transition of Hui community in cities” (in Chinese). Journal of the Central University for Nationalities 1, 73–78. Logan, J.R. and Alba, R. 1993. “Locational returns to human capital: Minority access to suburban community resources.” Demography 30, 243-268. Logan, J.R., Alba, R., Mcnulty, T., et al. 1996. “Making a Place in the Metropolis: Locational Attainment in Cities and Suburbs.” Demography 33(4), 443–53. Logan, J.R., Fang, Y.P. and Zhang, Z.X. 2009. “Access to housing in urban China.” International Journal of Urban and Regional Research 33(4), 914–35. Ma, L.J.C. and Xiang, B. 1998. “Native place, migration and the emergence of peasant enclaves in Beijing.” China Quarterly 155, 546–81. National Bureau of Statistics of China. 2005. 2000 China Township Population Census Data with GIS Maps. Beijing: All China Marketing Research Co Ltd. Sit, V.F.S. 1999. “Social areas in Beijing.” Geografiska Annaler 81B(4), 203–21. Solinger, D.J. 1999a. Contesting Citizenship: Peasants, Migrants, the State, and the Logic of the Market in Urban China. Berkeley and Los Angel. California: University of California Press. Solinger, D.J. 1999b. “Citizenship Issues in China’s Internal Migration: Comparisons with Germany and Japan.” Political Science Quarterly 114(3), 455–78. Tang, W.S. and Parish, W. 2000. Chinese urban life under reform: The changing social contract. Cambridge: Cambridge University Press. Wang, F.H. and Zhou, Y.X. 1999. “Modelling urban population densities in Beijing 1982–1990: Suburbanization and its causes.” Urban Studies 36(2), 271–87. Wang, F., Zuo, X.J., and Ruan, D. 2002. “Rural migrants in Shanghai: Living under the shadow of socialism.” International Migration Review 36(2), 520–45.

The Impact of Housing Tenure on Residential Segregation in Beijing, China

87

Wu, F.L. 2004a. “Transplanting cityscapes: The use of imagined globalization in housing commodification in Beijing.” Area 36(3), 227–34. Wu, F.L. 2004b. “Residential relocation under market-oriented redevelopment: The process and outcomes in urban China.” Geoforum 35(4), 453–70. Wu, F.L. 2010. “Gated and packaged suburbia: Packaging and branding Chinese suburban residential development.” Cities 27, 385–96. Wu, F.L. and Webber, K. 2004. “The rise of ‘foreign gated communities’ in Beijing: Between economic globalization and local institutions.” Cities 21, 203–13. Wu, W.J., Zhang, W.Z., Liu, Z.L., et al. 2010. “Tempo-spatial analysis of the residential land’s spatial pattern in Beijing” (in Chinese). Geographical Research 29(4), 683–92. Wu, W.P. 2002. “Migrant housing in urban China: Choices and constraints.” Urban Affairs Review 38(1), 90–119. Wu, W.P. 2004. “Sources of migrant housing disadvantage in urban China.” Environment & Planning A 36, 1285–304. Wu, W.P. 2008. “Migrant settlement and spatial distribution in metropolitan Shanghai.” Professional Geographer 60(1), 101–20. Xu, X.Q., Hu, H.Y. and Yeh G.O. 1989. “A factorial ecological study of social spatial structure in Guangzhou” (in Chinese). Acta Geographica Sinica 44(4), 385–99. Yang, S.M. and Wang, H.S. 2008. “An investigation on the Xinjiang village in Beijing” (in Chinese). N.W. Ethno-National Studies 57(2), 1–9. Yeh, G. O., Xu, X. Q. and Hu, H. Y. 1995. “The social space of Guangzhou City, China.” Urban Geography 16 (7), 595–621. Yuan, Y., Xu, X.Q. and Xue, D.S. 2007. “Spatial distribution, evolution and driving force of nonregistered population of Guangzhou metropolitan area in 1990–2000” (in Chinese). Economic Geography 27(2), 250-255. Zhang, L. and Wang, G.X. 2010. “Urban citizenship of rural migrants in reform-era China.” Citizenship Studies 14(2), 145–66. Zhang, L., Zhao, S.X.B. and Tian, J.P. 2003. “Self-help in housing and chengzhongcun in China’s urbanization.” International Journal of Urban and Regional Research 27(4), 912–37. Zhou, Y.X. 1996. “On the suburbanization of Beijing” (in Chinese). Scientia Geographica Sinica, 16(3), 198–206. Zhou, C.B. and Ma, X.F. 2004. “A discussion of the social structure of Hui in cities: A case study of the Hui community transformation in Beijing” (in Chinese). Journal of Hui Muslin Minority Studies 3, 33–39. Zhou, M. and Cai G.X. 2008. “Trapped in neglected corners of a booming metropolis: Residential patterns and marginalization of migrant workers in Guangzhou.” In Urban China in Transition, ed. J.R. Logan. Oxford: Blackwell Publishing, 226–49. Zhou, S.Y., Zhu, L.A. and Fan, Z.F. 2002. “The impact of urban transportation upon minority community: A case study of Madian community” (in Chinese). Beijing Social Science 4, 33–9. Zhou, Y.X. and Logan, J. 2008. “Growth on the edge: The new Chinese metropolis.” In Urban China in Transition, ed. J.R. Logan. Oxford: Blackwell Publishing,140–60. Zhou, Y.X. and Yu, H.B. 2004. “Reconstructing city population size hierarchy of China based on the Fifth Population Census” (in Chinese). City Planning Review 28(6), 49–55.

This page has been left blank intentionally

Chapter 4

Residential Segregation in an Unequal City: Why are There No Urban Ghettos in Hong Kong? Ngai-ming Yip

Introduction This chapter will explore Hong Kong’s residential segregation patterns which, despite the worsening income inequality trend, do not show sharp socio-spatial divisions and stigmatized enclaves of the poor, or “ghettos.” I argue that local institutions—public housing policies and town planning system—have played the major role in keeping the city from developing such ghettos. I rely upon the Large Street Block (LSBG) analyses based on two Census waves of 1996 and 2006. The use of the LSBG signifies an improvement over previous studies which employed only much larger Census tracts at the Tertiary Planning Unit (TPU) level. Previous studies found that Hong Kong had small enclaves of the rich and middle class as a colonial legacy (Lo 2005) while the rest of the city exhibited no sharp socio-spatial divide or stigmatized enclaves of the poor (Forrest et al. 2004). I share the same findings with previous studies, but I believe the use of the LSBG will provide much finer analyses of the city’s residential segregation than the use of the TPU. There were respectively 1,556 and 1,640 LSBG in 1996 and 2006 with an average of respectively 1,192 households and 3,871 person in 1996 and 1,357 and 3,932 in 20061 (in contrast to 135 TPU in 1996 and 139 TPU in 2001 previous studies were based on). Spatial segregation in this paper refers to the spatial distribution of a specific population group compared with the rest of the population. In this paper, I shall focus on the distribution of income and occupational groups across spatial boundaries in Hong Kong (the LSBG). The dissimilarity index (D), developed by Duncan and Duncan (1955), will be employed as the main index. It is a measure of “evenness” of distribution of the relevant population group in the city and indicates the proportion of a particular population group which would have to be moved across geographic boundaries in order to achieve equality of their distribution among the Census tracts. The dissimilarity index was computed using the multiple-group diversity and segregation indices module in the statistical package Stata 11.0 developed by Reardon (1999). The Gini segregation index (James and Taeuber 1985) will also be used alongside the dissimilarity index. Gini segregation index can be interpreted as “the sum of the weighted average absolute difference in group proportions between all possible pairs of subareas divided by the maximum possible value of this sum (obtained if the system were in a state of complete segregation)” (Reardon 2006: 180). It has more or less the same interpretation as the dissimilarity index with “1” denoting complete segregation and “0” no segregation. The Gini segregation indices were computed using R2.12.0. 1  In the analysis, domestic helpers and people in institutions at the time of Census (respectively 388,561 and 415,327 among which about 118,173 in 1996 and 187,149 in 2006 were domestic helpers) were excluded. The inclusion of domestic helpers (they were included in the Census Main Reports) would create bias in respective to the demographic and income profile (income of domestic helpers were included in the counted toward total household income) of the population in the specific geographic areas.

90

Residential Segregation in Comparative Perspective

With a more refined spatial scale in the population census, this chapter is able to offer a finer analysis of spatial differentiation patterns and fills a gap in our understanding of spatial segregation in Hong Kong. It enables us to indicate whether there is an apparent trend toward greater residential segregation given the continuing transformation of Hong Kong’s employment structure toward a more service sector based economy and the attendant growing income disparity. An initial analysis of the linkage between public housing provision and the planning system for private residential development will be attempted to examine the contribution of these two factors to the mitigation of spatial segregation. Divided Cities and Social Mix Within a heterogeneous society, cities are inevitably not homogeneous. While an undivided city is a myth and a utopia, neither is the conceptualization of divided cities as dual cities, polarized cities, nor fragmented cities useful in describing the complex division of neighborhoods within divided cities (van Kempen 2007). Classic studies from a social ecological perspective view such divisions as a natural process in the evolution of a city. Berry and Rees (1969) conceptualize the emergence of such divisions as the result of the interaction of several distinct spaces. The geographic location of the house a family chooses (the physical space) is the result of social space (economic and family status of the household), housing space (value and quality of the dwelling) and community space. Yet at the core of recent literature on divided cities is the connection between changes in the society of the city and changes in the division of urban space (Hamnett 1994, Marcuse and van Kempen 2002). Particular attention has been paid to the spatial concentration of poor people and the negative impacts resulting from such concentration. For instance, Wilson (1987) examines the negative impact of the spatial concentration of the underclass on people’s opportunities in education, jobs etc. whereas Granovetter (1995) found that the lack of networks for people in poor neighborhoods would hamper their chances of closing the gap with people who have resources. Spatial segregation would also lead to social cleavage and exclusion. A concentration of poor people in certain neighborhoods creates a stigmatization of such neighborhoods which further jeopardizes the chances of poor people to better their lives and reinforces income inequality (Musterd and Ostendorf 1998). Spatial segregation also hampers the full participation in society of residents in poor neighborhoods and thus generates social exclusion (Smets and Salman 2008). To combat the negative impact of differentiated cities, policies on the promotion of social and housing mix are being encouraged. It is hoped that more diversified neighborhoods with a mix of housing and tenure types can be developed so that the housing mix that is created will provide more social mix and subsequently also better conditions for positive socialization; it will also reduce the stigmatization and the risk of individual poor inhabitants becoming excluded (Musterd and Andersson 2005: 764). Within such a context, neighborhoods with homogenous tenure are regarded as causes of spatial segregation. Public housing, with its narrow tenant social base, becomes the target of action and countries like France even set a policy target of minimizing the proportion of public housing in each municipality (Musterd and Andersson 2005). Marcuse and van Kempen (2000) have argued more generally that the processes of social exclusion and segregation have different causes, consequences and trajectories in different urban regions of the world; European cities and North American cities display quite varied social geographies reflecting different welfare regimes and historical land use patterns. In fact, complex factors in states, cities, neighborhoods and among individual citizens, as well as policy responses

Residential Segregation in an Unequal City

91

to neighborhood problems, should be taken into account simultaneously (Musterd 2005). Van Kempen (2007) identifies seven contingencies which are relevant to the evolution of divided cities that need to be considered—the physical setting of a city, its history, economic development, inequality, race and racism, political power and governance. In fact, not all such factors are new but are linked to older theories that explain urban and neighborhood changes within cities. Hence, although the general economic structure of Hong Kong is shaped by global economic factors with the city being increasingly incorporated into the global economy, the dynamics of the local sociopolitical features are far from trivial. Such features work to create unique landscapes of social mix that generate the divided city. The absence of spatial segregation in Hong Kong presents a sharp contrast to cities in many Western countries in which spatial segregation is an important political concern, particularly when such an issue is associated with poverty (Musterd and De Winter 1998) albeit such countries have a much more equal distribution of income. Deliberate policies for enhancing diversification in neighborhoods have been enacted in countries like UK, Finland, France, Germany and Sweden (Musterd and Andersson 2005). Not only is an explicit social mix social policy absent in Hong Kong but there has also been no public concern with spatial segregation until recently when a few incidents of family tragedy have occurred in remote public housing estates. Despite the existence of high, and worsening, income inequality, such a social gap has not led to the creation of urban ghettos or large scale social unrest. While incidents of social disturbance were frequent in the 1950s and 1960s, they have almost disappeared in the last 30 years. Hence, it is worthwhile to explore the underlying factors that have counteracted the impact of worsening income inequality. Hong Kong: An Overview Despite being elected repeatedly as the world’s freest economy in the world by the Heritage Foundation, Hong Kong is far from non-interventionist in its social policy. In fact, involvement of the public sector in social service provision is deep by Asian standards although the level of provision may not be as good as the advanced welfare state in Europe, For instance, nearly half of the population in Hong Kong live in state subsidized housing and universal health care service is offered at minimal charge. While 12 years of free education is provided, nearly all universities are run by the public sector. Hong Kong also has a comprehensive social security system as well as a compulsory provident fund for retirement provision. Yet, with its mean income support system and the short history of the provident fund retirement scheme, protection for the elderly people and poor family is far from adequate. Hence, the family is still pivotal in the provision of welfare. However, with social spending at roughly one fifth of GDP, state provision is beginning to take a more important role in the provision of social welfare and income maintenance. The urban ecology of Hong Kong has been molded by its distinct geographic topology and colonial legacy. Traditional urban areas have developed around the narrow strips of flat land (most of which has been reclaimed from the sea) on both sides of Victoria Harbour. Areas around the Mid-Levels and the Peak close to the CBD have been historically exclusive residential areas for the prestigious class of the expatriates. The rural areas (the New Territories), which are separated from the city center by the rugged landscape, were only sparsely populated. Then, urbanization, which started in the 1970s, could only be accomplished by the construction of “new towns” linked to the CBD by expansive highways and railways. The spatial pattern of Hong Kong in the early twentyfirst century still shows an apparent urban–non-urban contrast (Forrest et al. 2004) with a multinuclei distribution of upmarket residential projects in the non-urban areas and a sectoral pattern

92

Residential Segregation in Comparative Perspective

of distinct land use in the urban area (Lo 2005). While the city’s urban landscape is still described as being differentiated by the Shevky-Bell constructs of socioeconomic, family and ethnic status (Lo 2005), a division is drawn largely between the very small number of affluent and mostly nonChinese households, and the rest of the population consisting of middle and lower income classes living side by side. There is no sign of extensive urban ghettos in the city (Forrest et al. 2004). Such urban landscape continued into 2006. Map 4.1 shows the distribution of areas in which there was a high concentration of the richest and the poorest households. A concentration of the affluence is apparent—16 percent of the LSBG had a concentration of more than 30 percent of households who belong to the richest 10 percent of the population (three times more than the average). Most of such areas were around the Peak and the Mid-levels of Victoria Peak, along the scenic beaches in the South of Hong Kong Island or in Kowloon Tong on Kowloon Peninsula. These areas were already enclaves of the elite class in the early years of the Colonial era. More recently, enclaves for the rich have come to be scattered around Hong Kong, often along the water front. Most of them are villas in low density developments in the New Territories, for instance, around Sai Kung in the eastern part of Hong Kong, Deep Bay in the west, and in Shatin and Taipo along the Tolo harbor as well as in Discovery Bay in Lantau Island close to the Disney Theme Park. Recent exceptions are the high density luxury developments on the waterfront of west Kowloon, next to the landmark development of the West Kowloon Cultural complex and the High Speed Railway Terminal, which are expected to be completed in the mid-2010s. By contrast, there hardly exists any spatial concentration of poor households. Only one percent of the LSBG had a concentration of over 30 percent of the poorest households (the lowest income decile group). Of the 17 such areas, the overwhelming majority are located in remote areas of the New Territories with few inhabitants, most of whom are elderly people. Of the seven poorest LSBG in the urban areas, six of them are old public housing estates built in the 1950s and 1960s (Shek Kip Mei, Ngau Tau Kok, Wong Chak Hang, Lam Tin, Ho Man Tin, and Valley Road), the remaining area is Nam Cheung in Shan Shiu Po having a high concentration of private sector dilapidated tenements. The overwhelming proportion of residents in such areas was poor and elderly people. These six old estates were all in the process of redevelopment, and by early 2010 they are replaced by new estates with much higher quality housing. The coefficient of variation of the percentage of the lowest income group among LSBG was 0.78 in 1996 and 0.63 in 2006. The corresponding figures for the percentage of the topmost income group were 1.36 and 1.25. The highest income group has a much higher variation among the LSBG than the lowest income group, which is in line with the aforementioned observations. Constrained by the geographic topology and high population density as well as a long history of state intervention in housing, the urban morphology of Hong Kong assumes a shape that is quite different from other cities of comparable economic status. Hong Kong has an unusually high proportion of public housing even by European standards. In 2006, nearly half (48 percent) of the housing stock had been developed by the Hong Kong Housing Authority (HKHA), a statutory organization under the Hong Kong Government. In 2006, only about two thirds of the flats developed by the HKHA (one third of the total stock) were rental housing whereas the other one third (16 percent of the total stock) was built-for-sale flats (called the Homeownership Scheme, HOS), most of which had been sold to better-off public housing tenants. Non-public housing tenants could also buy HOS flats but they had to pass a means test first. The HOS is organized as shared-equity housing and buyers could enjoy a 40 percent to 60 percent discount at purchase but had to repay the HKHA their share of the market value if the flats were re-sold on the open market. However, the overwhelming majority (87 percent) of HOS buyers have not sold their properties.

Residential Segregation in an Unequal City

Map 4.1

93

Distribution of the richest and poorest 10 percent of households Hong Kong, 2006

Note: High concentration of richest households (over 30 percent of households in top income decile group) High concentration of poorest households (over 30 percent of households in bottom income decile group). Source: Derived from Census 2006 (Census and Statistics Department 2007a).

Owing to considerations of efficiency, nearly all public housing developments are big. In 2009, there was an average of around 4,000 households with a mean population of over 10,000 people (more than double the average number of households in an LSBG) in each of the 155 public rental housing estates. Public housing is also very unevenly distributed around the city. The overwhelming majority (78 percent) of the LSBG in the 2006 Census did not contain any public housing at all whereas nearly all the remaining LSBG had more public than private housing. One in 10 LSBG were even public-housing-only areas. Although acquisition of public housing in Hong Kong is also means-tested, its relatively high quality (compared to lower end private housing) and cheap rent (only one third to half the market rent) makes it attractive to middle income households. In fact, public rental housing has been targeting households in the lower 40-percentile of the income spectrum and for a long time households who were affected by slum clearance did not have to be subject to a means test. Built-for-sale public housing flats were targeted at private tenants with income percentile of around 30 to 60, and public tenant buyers were exempted from the means test if they would relinquish their public flats. While sitting public tenants have to be subject to a means test 10 years after their move to public housing, the thresholds are quite lenient, at three times the income of new public housing tenants. Public-for-sale flats owners, on the other hand, did not have

94

Residential Segregation in Comparative Perspective

to be subject to subsequent means tests after they moved in. Hence, unlike most other countries, public housing in Hong Kong is not tenure for the poorest. Changing Socioeconomic Patterns in Hong Kong Hong Kong’s social and economic structures have been shaped by the city’s increasing integration into the global economy- and at the core of the wider Pearl River Delta (PRD) agglomeration, one of the most economically dynamic regions in the world. Hong Kong’s future is now inextricably linked to its position within the ever expanding metropolitan region of the PRD (Rohlen 2000). Hong Kong has been transformed from a manufacturing to a service sector city as labor intensive, manufacturing activities have shifted progressively to its Mainland backyard. Between 1984 and 1997, those employed in manufacturing fell from around 900,000 to less than 300,000 (Sung 2002). Between 1981 and 2006, the share of manufacturing industry fell from a dominating 49 percent of total employment to an insignificant 9.7 percent (Census and Statistics Department 1992, Census and Statistics Department 2007a). Employment in financial services, tourism, trading and logistics and professional and other producer services now accounts for 46 percent of total employment (Census and Statistics Department 2007a). It is thus not surprising to find a noticeable drop in the number of skilled and semi-skilled workers—from 51 percent in 1981 to 15 percent in 2006 (Census and Statistics Department 1982, Census and Statistics Department 1987, 2007a). Conversely, the managerial and professional groups saw the biggest increase over the same period, from 8.5 percent in 1981 to 27 percent in 2006 (Census and Statistics Department, 1987, 2007a). Such changes in the industrial and occupational structure lead Chiu and Lui to conclude that “Social polarisation is clearly observable in the occupational structure. And this polarisation is both relative and absolute” (2004, 1968). However, such conclusions are challenged by BorelSaladin and Crankshaw (2009) who argue that Chiu and Lui (2004) have arbitrarily treated middle income groups like “service workers and shop sale workers” as low wage, low skilled workers and consequently created an upward bias in the growth of the latter group. The growth of middle income service workers may have offset the loss of middle income manufacturing workers and polarization of occupational structure is not as serious. Nonetheless, even if Hong Kong’s occupational structure has not become more differentiated, its polarized income structure is evident. Not only did Hong Kong have the highest rate of income inequality in the developed world—the Gini coefficient of income distribution was 0.53 in 2006, but this figure has risen from 0.43 in 1981 (Census and Statistics Department 1982, 2007b). Between 1991 and 2001 workers in most sectors and at most levels experienced an increase in real incomes, but in general, higher paid workers gained more (Chiu and Liu 2004). Hence, “the observed overall inequality has been the result of the combined effect of a widening gap between the top and bottom incomes and the redistribution of the labor force from the manufacturing sector which had low inequality to sectors with significantly higher inequality” (Chiu and Liu 2004: 1875). The proportion of households who were at the extremes of the income distribution was on the increase while the proportion of middle income households has shrunk (Figure 4.1). The mean income of the richest 10 percent of households had also increased from HK$73,873 in 1996 to HK$87,926 in 2006 (19 percent increase in minimal terms and 24 percent in real terms) whereas for the poorest 10 percent of households, there was a corresponding decrease of 26 percent by nominal terms (24 percent by real term), from HK$2,792 to HK$2,056. The median household incomes between 1996 and 2006 were nearly the same (HK$17,500 in 1996 and HK$17,250 but because of deflation of 4 percent, median income in 2006 should be 3 percent higher than in 1996).

Residential Segregation in an Unequal City

Figure 4.1

95

Distribution of household income in Hong Kong, 1996 and 2006

Note: Nominal median incomes for 1996 and 2006 are nearly the same. Source: Census and Statistics Department 2007c.

So over the decade, although there were still a substantial proportion of middle income households, households at the extremes of the income distribution scale have increased. The richest also earned more while the poorest earned less. A greater income polarization thus took place between the richest and poorest families. New Immigrants from Mainland China have also exacerbated the problem of income polarization. Hong Kong accepts around 50,000 newcomers from Mainland China every year the majority of whom come for family reunion. Hence, 56 percent of new immigrants (official statistics only counts people from Mainland China who have stayed for less than seven years, after which they can become permanent residents) in 2006 were children under the age of 16 and their mothers (Census and Statistics department 2007a). Most of the new immigrants were less educated. Compared with 23 percent of the general population who had higher education, only 8 percent of new immigrants had a degree. Constrained by their level of education, familiarity with the local situation as well as problems of language fluency (local dialect Cantonese is not the mother tongue of many new immigrants), it is not surprising that they were less competitive in the job market. Nearly half of new immigrants who were in work could only find low income jobs in the service sector (compared with 27 percent of the general population) with median income only at 60 percent of that of the general population (Census and Statistics Department 2007d). Due to the expansion of higher education, the decade between 1996 and 2006 saw a big jump in the proportion of people who had completed higher education from 8 percent to 13 percent.

96

Residential Segregation in Comparative Perspective

Consequently, the occupation structure also has undergone a profound change with the professional groups (professionals and associate professionals) expanding relatively by one third over the decade. Yet as a result of the economic downturn after 1997, the relative proportion of managers and administrators dropped slightly from 13 percent to 11 percent over the decade. At the lower end of the occupational spectrum, with the continued trend of deindustrialization which began in the 1980s, the proportion of both skilled and semi-skilled manufacturing workers decreased in relative terms from 15 percent and 9 percent in 1996 to 11 percent and 7 percent in 2006 respectively, while the number of service workers and shop floor workers increased from 14 percent in 1996 to 18 percent in 2006. Further expansion of employment in the service sector and the contraction of the manufacturing sector suggest a continuation of the restructuring from a manufacturing to a service economy during the decade 1996 and 2006. However, the proportion of unskilled elementary workers during the period remained nearly the same (at respectively 12.5 percent and 12.4 percent). Hence, in Hong Kong the proportion of workers at the low end of the labor market (the unskilled workers) has remained stable over the decade of 1996 to 2006 but at the same time, workers at the high end employment have expanded. Whether the occupation structure has been polarized depends on the corresponding loss of relatively high waged skilled manufacturing workers relative to the gain in low waged service workers and shop floor workers. Yet, not all of the additional service workers in the 2006 census were low waged (Borel-Saladin and Crankshaw 2009). Given the relatively big increase in the proportion of professional groups, Hong Kong might possibly be stopping its polarization of the occupational structure and might be starting to engage in a gentle professionalization trajectory, with more of its workers having a higher education level and in better paid professional or white collar jobs. However, despite Hong Kong’s occupational polarization may have been halted, income divergence continues to worsen. While median income, measured in nominal terms remains nearly the same between 1996 and 2006 (real income increased by 16 percent largely owing to deflation), median household income of the lowest income decile group relative to the median income of the population has dropped from 16 percent in 1996 to 13 percent in 2006. On the contrary, corresponding figures for the highest income decile group have instead increased from 400 percent to 448 percent in the same period (Census and Statistics Department 2007c). There were also changes in the distribution of socially disadvantaged groups. The proportions of unemployed, elderly persons, and lone parent households have all increased relatively from 1996 to 2006. Lone parent households have doubled in relative terms (1.5 percent to 2.6 percent) over the decade, which may be partly attributed to the escalation of the divorce rate between 1996 and 2006.2 Hong Kong is also rapidly aging and the proportion of elderly people living alone or only with their spouse increased from 5.7 percent in 1996 to 7.3 percent in 2006. During that period, Hong Kong experienced the worst economic difficulties it has ever had since the Second World War. It was first hit by the Asian Economic Crisis in 1997, and then by the epidemic of SARS in 2003, so it is not surprising to find that unemployment rose from 2.3 percent in 1996 to 3.2 percent in 2006.3 Regarding housing, households who bought their homes in the private sector have risen from 33 percent to 37 percent during the decade, but those who rented from private landlords decreased 2  Based on the crude divorce rate which was 1.48 in 1996 and in 2006 it was 2.54 (Census and Statistics Department 2007a, Table 4). 3  The official unemployment rate in 2006 was 4.8 (Hong Kong Year Book 2006)—the difference is due to the way unemployment was recorded in the Census and the General Household survey from which the official rate is derived (http://www.yearbook.gov.hk/2006/en/fact_04.htm) Hong Kong Year Book 2006.

Residential Segregation in an Unequal City

97

slightly from 14 percent to 13 percent. With the slowdown of public rental housing construction and the loss of public housing stock through the sale of public flats to sitting tenants (the Tenant Purchase Scheme), only 31 percent of households remained public tenants in 2006, compared with 35 percent in 1996. Yet the biggest change in tenure is in public (assisted) homeownership, which has seen a more than 50 percent relative increase during the decade. This is a result of the high volume of assisted homeownership flats (Home Ownership Scheme) supplied as well as the launch of the Tenant Purchase Scheme in 1998 in which 117,137 flats were sold to sitting tenants (Hong Kong Housing Authority 2010) until the scheme came to an abrupt stop in 2002. Spatial Segregation 1996–2006 The segregation indices and Gini segregation index displayed in Table 4.1 enable us to look at both the patterns of spatial segregation and their development over the decade. Most attributes commonly associated with deprivation are fairly evenly distributed among the census tracts (Large Street Blocks LSBG) in Hong Kong. The pattern does not appear to be skewed toward the poorly educated across Hong Kong on a large scale. Only 16 percent in 1996 and 17 percent in 2006 would have had to be moved across LSBG boundaries in order to equalize their distribution. A similar observation holds for the economically inactive at respectively 12 percent and 11 percent. For the disadvantaged groups, which are of relatively smaller size, their distribution across the LSBG is more uneven. Around one third of the poorest households (the lowest income decile group) in 1996 and one quarter of them in 2006 and a similar proportion for unskilled elementary workers (26 percent in 1996 and 28 percent in 2006) would have had to move in order to equalize their distribution. Table 4.1

Segregation patterns in Hong Kong 1996–2006 (education, occupation and employment) % in population 1996

2006

Dissimilarity index 1996

2006

Gini segregation index Change 1996

2006

Education Below primary education University or above

41.3

32.9

0.16

0.17

0.22

0.23

8.2

13.2

0.39

0.36

0.52

0.48

↓

5.2

6.5

0.39

0.35

0.52

0.46

↓

↑

Occupation Professional Associate professional

12.6

17.0

0.22

0.20

0.30

0.27

↓

Manager and administrators

12.7

10.9

0.42

0.41

0.56

0.55

↓

Service workers and shop floor workers

14.3

17..5

0.16

0.17

0.18

0.18

↓

Skilled manufacturing workers

15.3

10.7

0.25

0.24

0.35

0.33

↓

Semis-skilled manufacturing workers Unskilled elementary labor

8.9

6.6

0.24

0.23

0.34

0.32

12.5

12.4

0.26

0.28

0.34

0.37

↓ ↑

98

Residential Segregation in Comparative Perspective

% in population

Dissimilarity index 1996

2006

Gini segregation index Change

1996

2006

1996

2006

39.3

34.8

0.12

0.11

0.17

0.15

↓

2.3

3.2

0.20

0.19

0.28

0.27

↓

Total persons (thousands)*

6,023

6,449

Average persons per LSBG

3,871

3,932

Total number of LSBG

1,556

1,640

Employment Economically inactive Unemployed

Note: * Number of persons in this analysis excludes domestic helpers and people who were in institutions. A total of 388,561 and 415,327 were excluded in 1996 and 2006 respectively. Source: Author’s Analysis of Census 1996 and 2006 (Census and Statistics Department 1997, 2007a).

Attributes associated with affluence and tenure are more unevenly distributed spatially as revealed in Table 4.2. For example, over half of the richest households (the top 10 percent of the income spectrum) (55 percent in 1996 and 57 percent in 2006) and a similar proportion of managers and administrators (42 percent in 1996 and 41 percent in 2006) as well as the group of professionals (39 percent in 1996 and 35 percent in 2006) would have had to move to equalize their distribution. Nearly three quarters of the private homeowners (72 percent in 1996 and 71 percent in 2006) would have to move to equalize their distribution and so would private tenants (55 percent in 1996 and 53 percent in 2006). A very high proportion of public tenants (95 percent) and assisted homeownership scheme owners (91 percent) in 1996 (respectively 87 percent and 86 percent in 2006) would have to move in order to achieve equality of spatial distribution. A change in the Gini segregation coefficients of the relevant population groups also largely matches the change in the segregation indices, More revealing is the direction of change in the pattern of segregation over the decade. Notwithstanding that income inequality, measured by the coefficient of variation (around the mean), has deteriorated from 0.55 in 1996 to 0.63 in 2006, yet in terms of spatial distribution, the poorest households were spatially more evenly distributed in 2006 than in 1996 (dissimilarity index from 0.31 to 0.24 and Gini segregation index from 0.43 to 0.34). Similar patterns are observed among the unemployed (dissimilarity index from 0.35 to 0.31 and Gini segregation index from 0.48 to 0.44) and the elderly households (dissimilarity index from 0.20 to 0.19 and Gini segregation index from 0.28 to 0.27) though the changes in other attributes associated with deprivation and disadvantage are rather mixed. The dissimilarity index for rich households has slightly increased, from 0.55 to 0.57, and the Gini segregation index 0.71 to 0.73, while the spatial distribution of professionals and associate professionals has decreased. (with dissimilarity index respectively from 0.39 and 0.22 to 0.35 and 0.20 and Gini segregation index 0.52 to 0.46 and 0.30 to 0.27). Hence, although income inequality in Hong Kong worsened over the decade between 1996 and 2006, spatial inequality did not. On the other hand, dissimilarity indexes of public tenants and homeowners were reduced respectively from 0.95 and 0.91 in 1996 to 0.87 and 0.86 in 2006 while Gini segregation index from 0.99 to 0.98 for public tenants and remained the same at 0.84 for homeowners. Tenure has become spatially marginally more evenly distributed but the degree of segregation is still very high.

Residential Segregation in an Unequal City

Table 4.2

99

Segregation patterns in Hong Kong 1996–2006 (income, family type and housing) % in population

Dissimilarity index

Gini segregation index

1996

2006

1996

2006

1996

2006

10.0 10.0

10.0 10.0

0.55 0.31

0.57 0.24

0.71 0.43

0.73 0.34

5.7 1.5

7.3 2.6

0.35 0.34

0.31 0.36

0.48 0.48

0.44 0.49

Private home ownership Private rental Public home ownership Public rental

33.3 13.9 10.3 35.3

36.7 12.7 15.8 31.0

0.72 0.55 0.91 0.95

0.71 0.53 0.86 0.87

0.84 0.71 0.98 0.99

0.84 0.70 0.94 0.98

Total households (thousands)

1,855

2,226

Average household per LSBG

1,192

1,357

Total number of LSBG

1,556

1,640

Change

Household income Top decile group Low decile group

↑ ↓

Family type Elderly households Lone parent households

↓ ↑

Housing ↓ ↓ ↓ ↓

Source: Author’s Analysis of Census 1996 and 2006 (Census and Statistics Department 1997, 2007a).

Urban Planning System and Housing Policy that Keep the City from Ghettoization The lack of urban ghettos in Hong Kong is believed to be the result of the housing and urban planning system in which public housing in Hong Kong was not the housing sector for the poor which is often the case in many other countries, like the United States, Australia and the UK. Hong Kong’s public housing (rental plus built-for-sale assisted homeownership housing) accommodates nearly half of the population, a deliberate effort in preventing the formation of a spatial concentration of the disadvantaged by maintaining an appropriate social mix that would itself be an effective policy in preventing spatial segregation in a city of worsening social inequality. Public housing in Hong Kong did not start as tenure for the poor in which the overwhelming majority of residents belong to the lower middle income group. Not only does the Hong Kong Housing Authority uphold a policy of social mix within their estates, their program of redeveloping old public housing estates has reinforced a more resident mix. The redevelopment program replaces estates which were built before the 1970s with a mix of modern rental blocks and more upmarket built-for-sale blocks. This effectively changes the mix of residents from a concentration of older and poorer people in the dilapidated rental blocks to a mix of rental and homeownership blocks in which a substantial proportion of new residents are wealthier middle class families who are also comparatively younger.

100

Residential Segregation in Comparative Perspective

To examine whether public housing creates a dampening effect on spatial segregation, a comparison is made between areas with public housing and those without. Four groups of areas (areas with and without public housing respectively in 1996 and 2006) are plotted in Kernel density plots. Kernel density plots use a smoothing technique to represent a histogram of a frequency plot.4 For low income households (Figure 4.2.1), elderly households (Figure 4.2.2) and people who are economically inactive (Figure 4.2.3), their distributions in areas with and without public housing are almost identical in shape. For the unemployed (Figure 4.2.4), elementary workers (Figure 4.2.5), single parents (Figure 4.2.6) and people with low education (Figure 4.2.7), there is substantial overlapping in their distributions but public housing areas tend to have a higher concentration of the relevant groups. Accordingly, although there is no definite statistics to quantify the differences, there is sufficient evidence that the social mix in public housing estates is largely similar to areas with a low level of public housing. Public housing areas with a high concentration of poor and disadvantaged households are also not evident. If we compare the distribution of public housing areas between 1996 and 2006, all the above attributes, with the exception of single parent households and the economically inactive, experienced a more even distribution. For variables that are associated with affluence, like the richest households (Figure 4.2.8), professionals (Figure 4.2.9), managers and administrators (Figure 4.2.10), differences in the concentration of the relevant groups between areas with and without public housing are more apparent. Areas with a higher concentration of public housing have a lower concentration of the relevant groups. Yet, distributions of more “affluent” groups show a much “flatter” shape and have a longer “tail” which indicates a more even spread of affluent households across areas that have no public housing. However, there is a much narrower difference in the distribution of the lower middle class group of associate professionals between areas in which private housing dominate and areas that are public housing dominated. At the same time, Hong Kong’s urban planning and land policy indirectly contributes to the creation of a social mix in the private sector. In order for the Government to reduce its upfront investment in developing the infrastructure, it has been a common practice in Hong Kong since the 1960s and 1970s for land to be leased to developers with only very minimal infrastructure (and sometimes none). The developers then have to shoulder the cost of construction of infrastructure (roads, community facilities and sometimes even land formation or reclamation) and in return they can receive a concession in the land premium or extra floor space beyond the limits set by the plot ratio. It would be in the interest of both the government and the property developer to achieve economies of scale by developing plots of land that are big enough. It is in this context that the scale of multi-owned residential projects is usually massive. The largest residential development project, Kingswood Villa in the New Territories, houses more than 15,000 households. On average, over half of the households in private housing in Hong Kong live in residential projects of over 1,000 households. Thus, given the high investment for big projects, property developers often develop the projects in phases to allow better circulation of capital. Hence, it would be a prudent market strategy not to concentrate on too narrow a market sector in order to spread the risk of the property cycle. It is, therefore, not uncommon to find big residential projects which contain flats 4  A kernel density chart can be thought of as a form of histogram which shows the distribution of the variable in question as if it were continuous. As a continuous distribution its shape is therefore not influenced by the number of ‘bins’ chosen to categorize the data as is the case with a histogram. They are extremely useful for visualizing the overall shape of a set of closely related distributions (such as here) which would be difficult to examine using a histogram. The examples here use stata’s kdensity plot command with its default Epanechnikov kernel function (http://www.stata.com/help.cgi?kdensity). Kernel density plots were generated with the statistical software R 2.11.1.

Residential Segregation in an Unequal City

101

of differing sizes (and as a result a wide range of prices) in order to minimize the risk of price fluctuation during the long period of development. For instance, in City One in Shatin (which accommodates over 10,000 households), the largest flats are 105m2 and cost US$850,000 (in early 2011), while the smallest ones are only 38m2 and cost US$280,000 (Centaline 2011). As a result, the median monthly income of different sections of the neighborhood ranges from US$7,600 to US$2,550 (Centamap 2008). It generates households with mixed income levels. However, more analyses of other similar neighborhoods would have to be conducted before we can have a fuller picture of the contribution of land and urban planning policy that works to keep spatial segregation from developing. Conclusion Hong Kong is a socially diverse city. In 2006, one in eight people were unskilled workers, while one in six people were managers or professionals. Hong Kong is also highly polarized in terms of its income distribution. The share of overall income of the richest 10 percent of households in 2006 was 52 times greater than their poorest counterparts in the bottom 10 percent of the income distribution, up from 38 times 10 years earlier. Moreover, this may seriously underestimate the level of inequality as financial assets are not included in this study. The richest households as well as people who are at the higher end of the occupation hierarchy (the managers, administrator and professionals) also tend to live in private owner-occupied housing and are spatially more segregated from the rest of the population. Despite these stark contrasts, there is little evidence of large-scale residential segregation of disadvantaged groups. The poor and the disadvantaged groups are still fairly evenly distributed spatially and across tenure. With the exception of exclusive areas for the very rich, most residential areas in Hong Kong are socially mixed. It is not uncommon to find middle class households living in close proximity with disadvantaged households like the low income households, the unemployed and those who are economically inactive. Notwithstanding the public housing sector, which accommodates nearly half of the population and is very unevenly distributed across the city, is by no means similar to its counterparts in USA and UK as tenure of the poor and disadvantaged, Although the poor, the unemployed and lone parents are over-represented in the public rental sector, a very similar distribution profile of the disadvantaged groups is found in the private housing sector areas. Hence, Hong Kong is still relatively homogenous in terms of its socio-spatial divisions A comparison between 1996 and 2006 has shown a mixed picture in the changing patterns of segregation over the decade—some attributes that are associated with poverty and disadvantage had experienced a reduction (poor households, the unemployed, economically inactive, elderly households) while others experienced an increase (lone parents, unskilled elementary workers and the poorly educated). Similar observations are made for attributes associated with affluence. For instance while the highly educated group, professionals and semi-professionals are more evenly distributed across spatial areas across the decade, the richest households have become more concentrated spatially. Yet, such changes should only be taken as a snapshot of the trend as a decade is too short to signify any long term change in spatial segregation. An extensive but spatially highly segregated public housing sector would have exacerbated the problem if the public sector in Hong Kong were to exhibit social characteristics and a tenants’ profile similar to its counterparts in western countries like the United States and UK. Yet, urban ghettos have not been found and in fact, Hong Kong has even become spatially slightly less segregated recently. The provision of affordable and reasonably good quality public rental housing,

102

Residential Segregation in Comparative Perspective

which attracts residents from a wide income spectrum, acts as an implicit policy in enhancing social mix within the public housing sector. The extension of public housing from rental to builtfor-sale assisted homeownership flats in the mid 1970s for better-off public tenants as well as middle income private tenants further advances the momentum of such a policy. The land and urban planning policy also plays a role in inhibiting factors that might lead to development of ghettos and in enhancing social mix. The government of Hong Kong commonly practiced shifting the burden of infrastructure construction to private developers and offered them large pieces of raw land with incentives like additional plot ratios or more relaxed planning restrictions. Creating a socially mixed development was often the strategy many property developers have adopted to contain the financial risk of their big investments. Thus Hong Kong’s high density living, town planning practice and strong public housing legacy have helped shape the residential segregation pattern peculiar to the city. The city’s segregation pattern is a product of market forces (labor and housing markets) and state policies and public institutions. Yet, the quantitative analysis of spatial differentiation in this chapter is unable to address whether an apparently not-so-segregated city is really a shared rather than a divided one. The adjacency of rich and poor in typically high rise, high density environments may only have created an illusion of meaningful propinquity and social interaction but not an indication of genuine integration, especially when residential enclaves for the richest people do exist extensively. Although the state may not have deliberate policies to mitigate social and spatial segregation, it has nevertheless created de facto social mix policies which have benefitted the poor and the disadvantaged. Hong Kong may have escaped from the problem of urban ghettos found in some residualized public housing estates in Europe or physically decayed neighborhoods in some cities in the United States. With socially mixed neighborhoods served with good public facilities and well-connected public transport, not only would the quality of life of the disadvantaged groups be improved, they may also be better integrated into the society as well as the labor market. A few new developments in Hong Kong may have a subtle impact on the future of residential distribution in Hong Kong. While the dilapidated buildings in the inner city may have triggered a process of urban decay and led to an increase in the concentration of disadvantaged groups, it may also induce intensive gentrification in some inner city areas. The displacement of poorer households to make way for an expanding middle class has occurred in Hong Kong as in other cities. However, gentrification in Hong Kong, because of its land use policy, high density and topography, has been more typically small scale and pepper potted-as opposed to the larger scale developments now in train. Older, lower rise apartment blocks in inner city neighborhoods have been replaced by higher rise, up market tower blocks. In this way, gentrification typically produces much higher middle class population densities. Thus, urban redevelopment in Hong Kong may have injected a significant degree of ‘social mix’ across a large number of poor neighborhoods. Yet, at the same time, such poor households, which are displaced by urban redevelopment, would be relocated to public housing. This would further residualize the already aging tenant profile in public housing and reverse the neutralizing role of the sector toward spatial segregation. At the same time, under the Tenant Purchase Scheme, better and well located public housing estates have been sold out whereas new estates are developed in more remote areas where new lands are available. Such new estates in the peripheral areas, which are unpopular owing to their relative remoteness, are attractive to disadvantaged groups, like elderly people and new immigrants, who could not afford to wait longer for public housing. Hence problems of social exclusion and spatial remoteness combine to reinforce the gap between these deprived groups and mainstream society. Overall, this development may weaken the role of public housing as a counteracting force against the development of ghettos.

Distribution of selected variables by areas with and without public housing in Hong Kong

Source: Author’s analysis of Census 1996, 2006 (Census and Statistics Department 1997, 2007a).

Figure 4.2

Figure 4.2

continued

Figure 4.2

continued

Figure 4.2

continued

Figure 4.2

continued

Figure 4.2

continued

Residential Segregation in an Unequal City

109

Acknowledgment The author would like to acknowledge the funding support of the CHASS Research Grant Projects of the City University of Hong Kong (Project #9610120). References Berry, J. and Land Rees, P.H. 1969. “The Factorial Ecology of Calcutta.” American Journal of Sociology, 74, 445–91. Borel-Saladin, J. and Crankshaw, O. 2009. “Social Polarisation or Professionalisation? Another Look at Theory and Evidence on Deindustrialisation and the Rise of the Service Sector.” Urban Studies 46(3), 645–64. Census and Statistics Department. 1982. Hong Kong Population census 1981 Main Report. Hong Kong: Hong Kong Government Printer. Census and Statistics Department. 1987. Hong Kong Population By-census 1986 Main Report. Hong Kong: Hong Kong Government Printer. Census and Statistics Department, 1992. Hong KongAnnual Digest of Statistics 1991. Hong Kong: Hong Kong Government Printer. Census and Statistics Department. 1997. Hong Kong Population By-census 1996 Main Tables. Hong Kong: Hong Kong Government Printers. Census and Statistics Department. 2007a. Hong Kong Population By-census 2006 Statistical Tables [Online]. Available at: http://www.bycensus2006.gov.hk/data/data3/statistical_tables/ index.htm#C1 [accessed: 25 March 2007]. Census and Statistics Department. 2007b. Hong Kong 2006 Population By-census Thematic Report: Household Income Distribution in Hong Kong [Online]. Available at: http://www. censtatd.gov.hk/products_and_services/products/publications/statistical_report/population_ and_vital_events/index_cd_B1120045_dt_latest.jsp [accessed: 25 June 2007]. Census and Statistics Department. 2007c. Domestic Households by Monthly Domestic Household Income, 1996, 2001 and 2006, Table D102 Statistical Tables of the 2006 Population By-census Main Page [Online]. Available at: http://www.bycensus2006.gov.hk/en/data/data3 /statistical_ tables/index.htm#D1 [accessed: 10 November 2010]. Census and Statistics Department. 2007d. Hong Kong 2006 Population By-census Thematic Report: Persons from the Mainland Having Resided in Hong Kong for Less than 7 Years [Online]. Available at: http://www.censtatd.gov.hk/products_and_services/products/publications/ statis tical_report/population_and_vital_events/index_cd_B1120049_dt_latest.jsp [accessed: 10 March 2011]. Centamap. 2008. [Online map]. Available at: http://hk.centamap.com/gc/home.aspx [accessed: 16 August 2009]. Centaline. 2011. [Online map]. Available at: http://hk.centamap.com/gc/home.aspx?ck= glist&ft3=cpal&source=C [accessed 18 March 2011]. Chiu, S.W.K. and Lui, T.L. 2004. Testing the Global City-Social Polarisation Thesis: Hong Kong since the 1990s. Urban Studies 41, 1863–88. Duncan, O.D. and Duncan, B. 1955. “A Methodological Analysis of Segregation Indices.” American Sociological Review 20, 210–17. Forrest, R., La Grange, A. and Yip, N.-M. 2004. “Hong Kong as a Global City? Social Distance and Spatial Differentiation.” Urban Studies 41, 207–27.

110

Residential Segregation in Comparative Perspective

Granovetter, M. 1995. Getting a Job: A Study of Contacts and Careers. 2nd edn. Chicago: University of Chicago Press. Hamnett, C. 1994. “Social Polarisation in Global Cities: Theory and Evidence.” Urban Studies 31(3), 401–24 Hong Kong Housing Authority. 2010. Revitalizing the Home Ownership Scheme Secondary Market, Memorandum for the Subsidised housing Committee of the Hong Kong Housing Authority [Online: Hong Kong Housing Authority, paper number SCH 32/2010, 11 May 2010]. Available at: http://www.thb.gov.hk/eng/policy/housing/policy/ consultation/shc3210.pdf [accessed: 15 September 2010]. Hong Kong Year Book. 2006. http://www.yearbook.gov.hk/2006/en/fact_04.htm [accessed: 20 August 2010]. James, D.R. and Taeuber, K.E. 1985. “Measures of Segregation.” Sociologicallvlethodology [Online], 14, 1–32. Available at: http://www.npc.umich.edu/publications/u/working_ paper09-05.pdf [accessed: 1 December 2010]. Lo, C.P. 2005. “Decentralization and Polarization: Contradictory Trends in Hong Kong’s Postcolonial Social Landscape.” Urban Geography 26(1), 36–60. Marcuse, P. and van Kempen, R. eds. 2000. Globalizing Cities: A New Social Order. Oxford: Blackwell. Marcuse, P. and van Kempen, R. eds. 2002. Of states and Cities: The Partitioning of Urban Space. Oxford/New York: Oxford University Press. Musterd, S. 2005. “Social and Ethnic Segregation in Europe: Levels, Causes, and Effects.” Journal of Urban Affairs 27(3), 331–48. Muster, S. and Andersson, R. 2005. “Housing Mix, Social Mix, and Social Opportunities.” Urban Affairs Review 40(6), 761–90. Musterd, S. and De Winter, M. 1998. “Conditions for Spatial Segregation: Some European Perspectives.” International Journal of Urban and Regional Research 22(4), 665–73. Musterd, S. and Ostendorf, W. 1998. Urban Segregation and the Welfare States, Inequality and Segregation in Western Cities. London: Routledge. Reardon, S.F. 1999. SEG: Stata Module to Compute Multiple-Group Diversity and Segregation Indices [Online]. Available at: http://econpapers.repec.org/software/bobocode/s375001.htm [accessed: 10 March 2010]. Reardon, S.F. 2006. “A Conceptual Framework for Measuring Segregation and Its Association with Population Outcomes in Michael Oakes.” In Methods in Social Epidemiology, ed. J.S. Kaufman. San Francisco, CA: Jossey-Bass, pp. 169–92. Rohlen, T.P. 2000. Hong Kong and the Pearl River Delta: “One Country, Two Systems” in the Emerging Metropolitan Context. Stanford: Shorenstein Asia-Pacific Research Center, Stanford University. Smets, P. and Salman, T. 2008. “Countering Urban Segregation: Theoretical and Policy Innovations from Around the Globe.” Urban Studies 45(7), 1307–32. Sung, Y.-W. 2002. Redefining Hong Kong’s Strategy of Growth and Development. Shatin. Department of Economics, Chinese University of Hong Kong. van Kempen, R. 2007. “Divided Cities in the 21st Century.” Journal of Housing and the Built Environment 22, 3–31. Wilson, W.J. 1987. The Truly Disadvantaged: The Inner City, the Underclass, and the Public Policy. Chicago: University of Chicago Press.

Chapter 5

A Portrait of Residential Differentiation in Taipei City (1980–2010) Chia-Huang Wang and Chun-Hao Li

Introduction There are two motives in this chapter. First, as the theme of the book is a comparison of urban residential segregation around the world the case study of the residential distribution pattern in Taipei City is part of the project to examine the similarities, differences, and mechanisms of residential segregation or differentiation in the cities covered in this book. Second, although the research on demographic mobility about Taipei City is abundant in Chinese and English research literature, so far we have found few publications on the residential segregation/differentiation of Taipei City or the urban areas of Taiwan. By writing this chapter, we hope to motivate scholars interested in the spatial and urban issues of Taipei City in particular and other cities in Taiwan and East Asian countries in general, to conduct more case and comparative studies, and to enrich the research agenda. Thus, this chapter intends to explore whether there has been residential segregation in Taipei City and tries to find the explanations. We intend to discuss the following questions: Has there been significant residential segregation in Taipei City? If so, what are the possible causes and contributing factors? If not, what are the factors that have prevented residential segregation from forming in Taipei City? Is there a better term to describe the residential distribution patterns in Taipei City? What are the roles of the Taiwanese developmental state and the Taipei City Government in shaping the residential pattern in Taipei City? What theories, models, or theses better describe and explain our research findings? The chapter is divided into six parts. The first part explains the research motives and research questions, establishing the conceptual framework to direct our discussion and reasoning. The second part describes the data and methods used in our study, and presents a short history of Taipei City. The third part summarizes our research findings, focusing on the socioeconomic status of the city’s residents, the housing prices, and the residential distribution patterns in Taipei City. The residential distribution of minorities in Taipei City is also briefly described. The fourth part describes the city’s status in the industrial policy of the central government, the developmental strategy of the city government under the mayors since the early 1990s, the intergovernmental relations of the city, as well as the spatial planning and housing policy, all of which shape and condition the residential distribution patterns in Taipei City. The fifth part examines the descriptive and explanatory power of theoretical propositions based on the investigations of American and European urban development, in view of our research findings, on the one hand, and the developmental state theory and the nested city approach, on the other. Finally, the last part summarizes the major points and arguments of this chapter.

Residential Segregation in Comparative Perspective

112

To help direct our analysis and discussion, we propose a conceptual framework, which is shown in Figure 5.1. We assume that the residential distribution patterns in Taipei City have been shaped by the policy mix (spatial planning framework, and industrial and housing policies) of the developmental state, the socioeconomic status of residents, as well as racial prejudice and discrimination. The spatial planning regime, industrial policies, and housing provision policies of the developmental state are the most important set of factors in our conceptual framework; they influence the socioeconomic status and socioeconomic opportunities of the city’s residents through economic planning, industrial policies, educational systems, and developmental strategy. The degree of racial/ethnic prejudice and discrimination is conditioned by the policies and legislation concerning racial/ethnic equality. The thicker arrows indicate stronger relations while the thinner and dotted ones weaker connections. The conceptual framework does not include all the possible causes of the residential distribution patterns in Taipei City. It is based on the data and documents currently available and focuses on the macro factors (especially industrial and housing policies) that interest us. Micro factors, such as personal choices and preferences, are not included in our study. This does not mean that we intend to ignore them, to leave out evidence contrary to our argument, nor do we think they are insignificant. We leave them to the scholars who hold micro perspectives and focus on individual factors.1

Figure 5.1

A conceptual framework of the residential distribution patterns in Taipei City

Source: Drawn by the authors.

1  See, for example, Chang et al. (2003) for a quantitative analysis of residential mobility in Taipei City.

A Portrait of Residential Differentiation in Taipei City (1980–2010)

113

Data, Method and a Short History of Taipei City Data and Method To support our reasoning and present the patterns, we adopt secondary data analysis. The first type of data derives from the censuses of 1990 and 2000 conducted by the Directorate-General of Budget, Accounting and Statistics (DGBAS), the Executive Yuan.2 The second type is based on the database of Taipei City Government to complement the first type of data. In analyzing the census data, we take borough (Li in Chinese) as the fundamental analytical unit, using the dissimilarity index to measure the pattern of residential differentiation.3 The strength of this database and analytical strategy is that we can capture more detailed distribution patterns and conduct statistical tests to explore the significance of the data. However, there are several weaknesses in the census data. For one thing, the variables surveyed in these two censuses are not identical. For example, the 1990 census did not collect any data about the residential distribution of aborigines. Thus, a longitudinal comparison of their residential patterns from the 1980s to 1990s is impossible. Moreover, data of personal or household incomes were not collected. To overcome the lack of consistent data, we have used the data annually compiled by Taipei City Government.4 The data allow us to cover the latter half of the 2010s and make it possible for us to present a three-decade longitudinal study.5 To present our data graphically, we used Arc-GIS to chart the residential distribution patterns. A Short History of Taipei City Map 5.1 shows the shape of Taipei City, surrounded by Xinbei City (called Taipei County before January 1, 2011). Both cities constitute Metropolitan Taipei and look like a poached egg. Taipei City is the yolk or center of metropolitan Taipei, whereas Xinbei City is the egg white. The industrial and socioeconomic interaction (e.g. commuting and consumption) between the two cities has been intertwined and intensive. Map 5.2 indicates the contemporary administrative districts of Taipei City and the distribution of higher educational attainment (undergraduate and graduate).6 Roughly speaking, the center of gravity of Taipei City developed in the western part (Districts 4, 6, and 7) of the city under the rule of Qing Dynasty (1683–1895), but several clusters were formed in other parts of the city. During the Japanese colonial period (1895–1945), not only were the governor-general’s office and the agencies of the colonial government located in the western part, but economic (business and consumption) activities were concentrated in the same region.

2  The census is implemented decennially in Taiwan. The 2010 census data were not available at the time of writing this chapter. The Executive Yuan is the top administrative organ of the government and the Premier is the leader of the cabinet. The president is the national leader, who commands the Executive Yuan and nominates the leaders of other Yuans, such as the Control Yuan and the Examination Yuan. 3  Currently, there are 456 boroughs in Taipei City and the average population size of a borough is 5700. 4  This set of data is displayed in Chinese and compiled by the Department of Budget, Accounting, and Statistics (DBAS), the Taipei City Government. The e-database is accessible at http://www.dbas.taipei.gov. tw/np.asp?ctNode=6151&mp=120001. 5  The contribution of this set of data is limited, however, because the data before 2004 were not divided into administrative districts. 6  There were 16 administrative districts, but they have been regrouped into 12 since 1990.

114

Map 5.1

Residential Segregation in Comparative Perspective

Metropolitan Taipei (Taipei City and Xinbei City)

Source: Drawn by the authors.

Since the relocation of the Kuomintang (KMT or the Nationalist Party) Government to Taiwan, Taipei City has become the “temporary” capital of the Republic of China. The president’s office and most central governmental ministries were established in the western part, which became the central business district (CBD) of Taipei City. Taipei City was promoted to the status of municipality in 1967 and incorporated part of the townships of Taipei County (Xinbei City since January 1, 2011) in 1968. These townships are Districts 12, 11, 10, 9, and 8, which constitute the eastern part of Taipei City. The CBD has moved toward the eastern part of Taipei City (Districts 2 and 3) since the late 1980s, because of the expansion of economic activities, the growth of population, and the City Government’s spatial development planning. Moreover, some sub-CBDs emerged from the old city clusters in the northeastern part (Districts 10, 11, and 12) and southeastern part (Districts 8 and 9) of the city. The relocation of the City Government and City Council to the new premises in District 2 (1994 and 1990, respectively) symbolizes the socioeconomic development in the eastern part and the emergence of a new CBD. The western part experienced social and economic decline for a while, but signs of revival have begun to emerge, partly because of the mass rapid transit system (MRT) and urban revival strategy (e.g., axial reversal) of the City Government. Thus, Taipei City has evolved from a monocentric city to a multi-centric metropolis (Chou 2005, Jou 2005).

A Portrait of Residential Differentiation in Taipei City (1980–2010)

Map 5.2

Higher vs. non-higher education in Taipei City (1990 and 2000)

115

Source: Based on the 1990 and 2000 censuses of Directorate-General Budget, Accounting & Statistics, Republic of China in Taiwan; drawn by the authors.

116

Residential Segregation in Comparative Perspective

Major Findings Higher Educational Attainment Map 5.2 shows the long-term distribution patterns of residents with higher education (graduate and postgraduate) in the boroughs of Taipei City during the 1990s. The two maps have been produced taking into account contiguity between high and low values of the higher education variable; therefore, the areas come out on these maps not only in respect of the rate of persons with higher education they contain, but also in respect of what happens in their neighboring areas. The areas depicted with the darker shades are those with high rates of higher education graduates that are part of broader areas with the same characteristic. The upper map (1990) shows that residents with higher education tend to concentrate in Districts 1, 2, 3, and part of 5, whereas those with lower education in Districts 6, 7, 9, as well as part of 8, 11, and 12. That is, the residents with higher education concentrate in the central part or CBD of Taipei City, while those with lower education in the western and eastern parts of the city. The lower map (2000) shows that residents with higher education tend to concentrate in Districts 10, 11, 12, and, (to a lesser extent) 8. The dissimilarity index of residents with higher education decreased slightly from 0.26 in 1990 to 0.24 in 2000. Nevertheless, we do not have reliable data to indicate whether these changes are explained by intra-city relocations or by in-migration from other cities of Taiwan. These two maps and the dissimilarity indices indicate that the segregation between groups of different education qualifications is quite moderate. Residential differentiation or clustering would be a more appropriate term to describe the residential distribution pattern in Taipei City. Moreover, the Pearson’s correlation between higher education and average annual household income is high and significant both in 2001 (r = 0.82) and 2006 (r = 0.90). Thus, it is reasonable to use higher education as a measure to indicate socioeconomic residential differentiation in Taipei City. Housing Price and Yearly Average Household Income We examine two sets of indicators in this section. The first is housing prices and the second yearly average household income in Taipei City. Table 5.1 shows the housing prices of the administrative districts of Taipei City from 2008 to 2010. Firstly, the average housing prices of Taipei City increased by 17.6 percent, from 599,000 to 705,000 NT Dollars per surface unit (about 3.3 square meter or 36 square feet) during these three years. Secondly, the higher housing prices tend to concentrate in the administrative districts in the central region of the city, such as Districts 3 and 5, while the lower ones in Districts 6 and 7 (the old CBD and the western part of Taipei City). The housing prices in Taipei City have been unreasonably high in the last decade, if one considers the presence of low-income households (Chang et al. 2001). The housing prices in Taipei City have been significantly high when one considers the distribution of yearly average household income per household (see below). Table 5.2 shows the housing prices in Xinbei City. It is apparent that the housing prices in Xinbei City (in each administrative district and on average) are far lower than those in Taipei City. Xinbei City has provided an affordable alternative residential opportunity for people who need to work in Taipei City or other northern Taiwan cities but cannot afford to buy a house in the City. Nevertheless, the housing prices in the districts of Xinbei City have been rising, partly because of the demand of the residents and partly because of the speculations of property developers. The average housing price from 2008 to 2010 increased by 30 percent, a rate higher than that of Taipei City. The housing prices are especially high in the districts directly adjacent to Taipei City, partly

A Portrait of Residential Differentiation in Taipei City (1980–2010)

Table 5.1

117

Housing prices in Taipei City (2008–2010, thousand new Taiwan dollars and percentage) 2008

2009

2010

Increase 2008–2010 (%)

1 Songshan 2 Xinyi 3 Daan 4 Zhongshan 5 Zhongzheng 6 Datong 7 Wanhua 8 Wenshan 9 Nangang 10 Neihu 11 Shihlin 12 Beitou

625 675 828 730 725 467 489 397 498 546 782 428

864 840 1,099 610 653 491 401 389 526 470 663 431

951 951 1,152 750 901 556 440 464 634 569 699 476

39.2 40.8 39.1 2.7 24.2 19.0 -10.0 16.8 27.3 4.2 -10.6 11.2

Average

599

576

699

17.6

Year      District

Source: http://www.a-life.com.tw/reported2.php?rec=114.

Table 5.2

Housing prices in New Taipei City (2008–2010, thousand new Taiwan dollars and percentage)

Year      District

2008

2009

2010

Increase 2008–2010 (%)

1 Banqiao 2 Shanchong 3 Zhonghe 4 Yonghe 5 Xinzhuang 6 Xiandian 7 Shulin 8 Yingge 9 Shanxia 10 Danshui 11 Xizhi 13 Tucheng 14 Luzhou 16 Taishan 17 Linko 18 Shekeng 23 Bali

254 262 262 351 219 323 171 126 149 181 214 218 21 171 155 183 175

287 266 272 354 256 319 173 123 14 174 20 229 22 17 159 165 17

329 349 314 464 39 345 184 13 223 234 255 247 323 29 217 23 178

29.5 33.2 19.8 32.2 78.1 6.8 7.6 3.2 49.7 29.3 19.2 13.3 53.8 69.6 40.0 25.7 1.7

Average

213

216

277

30.0

Source: http://www.a-life.com.tw/reported2.php?rec=114.

118

Residential Segregation in Comparative Perspective

because buying or renting houses in the neighboring districts could help residents save commuting costs and access the amenities in Taipei City. The second set of economic indicators is the yearly average income per household. Figure 5.2 shows three income groups. The yearly average household incomes of residents in Districts 3 and 5 were higher than those of their counterparts in other districts in 2002, 2005, and 2009. Districts 1, 2, 3, and 5 (the central part or CBD of Taipei City) constitute the top group while Districts 6, 7, and 9 (the western and eastern peripheral areas) are at the bottom. The rest of the districts constitute the middle group. According to our statistical analysis, the variation is significant. The η2 (eta-square) values are 0.76, 0.76, and 0.74 in 2002, 2005, and 2009, respectively.

Figure 5.2

Average yearly household income in Taipei City*

Note: * For the names and locations of Districts 1 to 12 on the horizontal axis, see Table 5.1 and Map 5.2. Source: Based on the online database of the Department of Budget, Accounting & Statistics, Taipei City Government: http://163.29.37.101/pxweb2007-tp/dialog/statfile9.asp; drawn by the authors.

Moreover, a governmental research (CPAMI 2011) indicates that, from the first quarter of 2007 to that of 2011, the ratio of housing prices to average yearly household income has increased from 8 to 13, which has always been far higher than those of Xinbei City (roughly from 6.5 to 9) and other cities in Taiwan (roughly from 6.5 to 8.5). More than half of the residents’ monthly household income is spent on housing mortgages in Taipei City, compared to 35 percent in Xinbei City and to lower than 35 percent in other cities in Taiwan. The description and reasoning about housing prices, household income, and the ratios of housing price to yearly household income and of housing mortgage to monthly household income above are partially supported by a quantitative research of Hsu and Guo (2006: 243), who find that higher income households prefer to live in the CBD to save commuting costs and enjoy convenient urban amenities, and that numerous households have moved to the northern and eastern parts of Taipei City and some districts in Xinbei City with the MRT stations, which indicates that those households are willing to balance commuting cost with lower housing costs and other advantages. These patterns are very important, because the districts with higher housing prices happen to be the places where yearly average household income is higher. In the descriptions below, we will

A Portrait of Residential Differentiation in Taipei City (1980–2010)

119

see that the districts (or boroughs) where residents with higher educational attainment are clustered happen to be the places where housing prices and yearly average household income are higher. However, we are not able to conduct correlation analysis because these sets of data are compiled from different sources. Occupational, Employment, and Industrial Structure Figure 5.3 displays the structure and change of occupations in Taipei City between 1993 and 2010. The percentage of “professionals” and “technicians and associate professionals” grew steadily from the early 1990s to 2010. In contrast, the percentage of “Service Workers” and “Craft and Related Trades Workers” gradually declined. In other words, professionalization seems to be the major trend and our findings correspond to the comparative studies of East Asian cities held by Tai (2006, 2010), who finds that the data on socioeconomic status for Taipei, Seoul, Hong Kong and Singapore indicate professionalization, rather than polarization. Thus, we do not think that social polarization or an hourglass-shaped structure in terms of occupation categories is of relevance in Taipei City. The structure of the occupational distribution looks like an inverse trapezoid.

Figure 5.3

Occupational distribution of residents in Taipei City

Source: Based on the online database of the Department of Budget, Accounting & Statistics, Taipei City Government: http://163.29.37.101/pxweb2007-tp/dialog/statfile9.asp; drawn by the authors.

In terms of employed population, the service industry (trade, transportation and storage, accommodation and eating/drinking place, finance and insurance and other services) accounts for 80.6 percent, while the industrial and manufacturing industry (mining and quarrying, manufacturing, electricity and gas supply, water supply, and construction) accounts for 19.3 percent in 2010. (Taipei City Government 2010: 17, 26–7). With regard to the industrial structure, the majority (92 percent) of businesses registered in Taipei are in the service industry and wholesale and retail has been the leading sub-sector (56 percent) in the service industry. This set of data seems to correspond to the pattern shown in Figure 5.3. We assume that this type of growth is related to the industrial development of the city under the industrial policy of the developmental state and the development strategy of Taipei City Government, but we need more data to confirm this hypothesis.

120

Residential Segregation in Comparative Perspective

Minorities: Aborigines, New Immigrant Spouses, and the Post-war Mainland Chinese Refugees Since the Qin Dynasty, the aborigines in Taiwan have been suppressed and constituted the major minority group under the Qin Dynasty, the Japanese colonization, and the Kuomintang rule. After decades of policies of assimilation based on Chinese ethnic identity and thanks to the democratization since the late 1980s, the aborigines began to struggle for the revival of their culture.7 Thus, the central and local governments have implemented policies to improve the political and socioeconomic situation of the aborigines since the mid-1990s, but the majority of the aborigines are still socioeconomically disadvantaged. For example, a survey conducted by the Taipei City Government in 2006 finds that the proportion of aborigines with higher educational attainment is only 17.3 percent, which is clearly lower than that of the general population (more than 50 percent in 2008–2009). The unemployment rate of the aborigines is 4.8 percent while that of Taiwanese 3.9 percent.8 Due to the limited space, we are not able to display the map of the aborigines’ residential distribution in Taipei City, but our data from the 2000 census indicate that the aborigines in Taipei City tend to concentrate in the eastern and southern part of the city; some of them are scattered in the central and western parts. The dissimilarity index of aborigines was 0.31 in Taipei City in 2000. There have been some clusters of aborigines in the peripheral areas in Taipei City, but we can hardly conclude that ethnic Chinese and the aborigines are residentially segregated in Taipei City. Moreover, we have to point out that many Taiwanese are the descendants of Chinese and the aborigines, and that many aborigines have been living in the highly urbanized areas in Taiwan and intermingling with ethnic Chinese. We do not display the residential distribution map of the new immigrant spouses (mostly female) through cross-border marriage either, due to the limit of space. However, some points are noteworthy stemming from the online database of Taipei City Government and the 2000 census. Firstly, the numbers of the new immigrant spouses are not important, but in many cases they are socioeconomically disadvantaged.9 The majority of them need the government and volunteer groups to help them. The policy issues include domestic violence, cultural (and language) adaption of themselves and their children, discrimination, caring the elderly, and employment. Secondly, the number of mainland Chinese spouses is far bigger than of those from Southeast Asia in every administrative district in Taipei City, and especially in Districts 7, 8, and 2. This is reasonable when we consider the socio-cultural affinity of Mainland China and Taiwan as well as the increasing socioeconomic exchange across the strait. Finally, there seems to be a bifurcated pattern in the distribution of Mainland Chinese spouses in Taipei City. Many of them married natives without higher educational attainment or who could not afford the high housing prices in Taipei City, but quite a few married natives do live in the districts with high housing prices. The residential locations of those new immigrant spouses are not 7  For further information about the aborigines in Taiwan, see the homepage of the Council of Indigenous People, ExecutiveYuan: http://www.apc.gov.tw/portal/index.html?lang=en_US&CID=B20D85423B8EED93. 8  The 2006 survey interviewed 49 percent of aboriginal residents in Taipei City (total 11,551). Available at: http://w2.native.tcg.gov.tw/9501.pdf. The ratio of higher education of the total population in Taipei City is obtained from the database of the DBAS, Taipei City Government. As of 2010, there are 13,745 aborigines and 4,334 aboriginal households in Taipei City; the number of aboriginal population is roughly 0.5 percent of the city’s population. 9  In 2010, there were 33,766 new immigrant spouses in Taipei City and the percentage of this group in the city’s population was 1.29 percent. The number of new immigrant spouses from Mainland China was 30,557 and those from Southeast Asia 3,209.

A Portrait of Residential Differentiation in Taipei City (1980–2010)

121

based on their own status (many of them came from the socioeconomically disadvantaged regions of Mainland China and Southeastern Asian countries), but on the socioeconomic status of their husbands, who usually occupy lower socioeconomic status and want to marry new immigrants to find helpers for domestic labor (taking care of the elderly) or to assist them with their businesses. Regarding the Mainland Chinese refugees moving to Taiwan after 1949, most were military personnel (officers and soldiers), civil servants, students, and dependents, but there are no reliable data. The estimates range from one to four million. Some of them lived in the villages built for the dependents of military and civil personnel, but many more lived independently and married local Taiwanese. Those villages were set up to take care of the dependents of military and civil servants, rather than to segregate them from local Taiwanese, because of ethnic prejudice or discrimination. During the Cold War era, some villages were established in the less urbanized areas to minimize the casualties in case of an attack from Mainland China. After half a century, the majority of the villages are so old that they need to be reconstructed. Many new residents who are not of Mainland Chinese origin moved into the new communities. At the same time many people of the second generation have married local Taiwanese and raised their children. In other words, all residents moving to Taiwan after 1949 and those born in Taiwan are citizens of the Republic of China in Taiwan, although the People’s Republic of China (PRC) still claims that Taiwan is a province of the PRC. There have been some Chinese and English studies trying to argue that Mainland Chinese (the first generation and their offspring) and local Taiwanese are different in terms of national identities, ethnic culture, and socioeconomic status. Indeed, there was preferential treatment for Mainland Chinese and their children in terms of civil service examinations and recruiting, and the policy of suppressing Taiwanese dialects, which forced all Taiwanese to speak mandarin. However, those policies and measures belong to the past and we find no evidence to argue that there is residential segregation between Mainland Chinese (first and second generations) and Taiwanese based on ethnic prejudice. The Taiwanese Developmental State and the Residential Differentiation in Taipei City Industrial Policy, Developmental Strategy and Taipei City Government As “Temporary Capital” of the Republic of China, Taipei City has always been the command and control center of Taiwan’s economy. In view of the need of industrial structural transformation and upgrading, the central government and Taipei City Government have been planning to strengthen the city’s commanding status and international competitiveness. Thus, knowledge-intensive (e.g. biotechnology) and high-end tertiary industries are the strategic sectors in the city’s economic development. In the 1980s, the Ministry of Economic Affairs (MOEA) started constructing the Taipei World Trade complex to promote international trade and exhibitions. In the 1990s, the central government drew the Asian Pacific Operational Center (APROC) Plan, designating Taipei City as the base of Taiwan’s international financial hub and media center. Incentives were provided to attract transnational corporations to establish headquarters in Taiwan (Wang 2003). In the early 2000s, Taipei City under Mayor Ma Yin-Jeou did not occupy an important status in the economic plans and industrial policies of the central government. The main reason is that Ma was one of the political stars of the Kuomintang and the potential candidate in the coming presidential campaign of 2008. The Democratic Progress Party (DPP), i.e. the ruling party at that

122

Residential Segregation in Comparative Perspective

time, did not want to give Ma the opportunity to gain popularity. The political contradictions between the central and city governments changed when a KMT member became the third mayor voted by the residents of the city in 2006 and Ma won the presidential campaign in 2008. Despite the political contradictions, the industrial and economic development of Taipei City has not been totally blocked. On the one hand, the MOEA promoted the construction of Nangang Software Park in District 9 of Taipei City, to build a software R&D base and the Asian-Pacific Software Development Center.10 The transportation systems, such as the mass rapid transit (MRT), the railway, the high speed railway, and superhighways in Metropolitan Taipei, have been continuously improved and constitute a dense web facilitating mass transportation and the use of automobiles, which has been instrumental to the commuting flows among cities in northern Taiwan. On the other hand, the designation of the Xinyi Planning Area (in District 2), which is located in the eastern part of the city center, as a secondary CBD illustrated the development strategy of Taipei City. The plan of developing the Xinyi Area slowed down in the late 1980s, but resumed in the 1990s when the Taipei City Government relocated to its present location in Xinyi Planning Area in 1994. The relocation of the city government signified the eastward development of the political and administrative center of gravity. Since then, the development strategy of the Taipei City Government has started to take shape with the aim to transform Taipei City into an internationally competitive city, to attract professionals, (white- and blue-collar) guest workers, tourists, transnational corporations, and foreign investment. Since the late 1990s, the Taipei City Government has been promoting high-end service and hightechnology industries. The high-end service industrial sector includes FIRE (finance, insurance, and real estate), MICE (meeting, incentive travel, convention, and exhibition), as well as creative and tourist industries. The Department of Taipei Information and Tourism has been established and operating since 2007, in order to promote tourism and build city image through cultural industries and festivals (e.g. Taipei International Flora Exhibition). To promote high-technology industry, the city government also built the Neihu Technology Park (in District 10) and planned to construct the Beitou-Shihlin Technology Corridor (Districts 12 and 11), which is connected with the Nangang Software Park and the Neihu Technology Park. This technology corridor is planned to generate clustering effects and complement the commercial and business development zone in the central part of the city (Wang 2006: 287–90). The city government was also coordinated with the central government’s policy of building Nangang Software Park by constructing the Nangang Economic and Trade Park, in order to integrate Taiwan into the Asian Pacific Logistic and Operating Hub.11 Finally, the City Government restructured the Office of Public Housing into the Urban Redevelopment Office and merged it with the Department of Urban Development in 2004. Since then, the city government has been promoting urban renewal and providing developers with incentives (e.g., plot ratios) to promote the city’s economic development and shape new landscape with sufficient amenities. The merging of the Public Housing Office and the policy of promoting urban renewal suggest two things. First, the housing policies of the central government and the city government have not been successful in balancing the supply-and-demand of the housing market in Taipei City. Second, the city government gradually relies on the private sector to adjust the supply10  See the webpage of the park for its mission, development plan, and other details: http://www. centurydev.com.tw/nksp%202/default.asp. 11  See the webpage of Nangang Software Park at http://www.centurydev.com.tw/nksp%202/introduction/ intro.asp.

A Portrait of Residential Differentiation in Taipei City (1980–2010)

123

and-demand of the housing market, despite the fact that the central and city governments have announced that they will increase the provision of public housing.12 According to our calculation based on data provided by the Department of Urban Development, Taipei City Government, there are 55,430 units/households of public housing as of 2011 and more than half of them concentrate in Districts 1 (12,780), 3 (7,149), 7 (8,674), and 10 (6,755). The number of total households in Taipei City are 983,237 in 2011 and thus the ratio of public housing is only 5.6 percent, which is far lower than those of Hong Kong (29 percent), Singapore (8.7 percent), and other advanced countries (e.g. United States, Holland, Japan, etc.).13 While the eastward development has been accelerating, Taipei City Government also tried to revive the declining western part (mainly Districts of 6 and 7). In his term as Taipei City’s mayor (1998–2002, 2002–2006), Ma Ying-Jeou, the president of the Republic of China in Taiwan (2008–2012), promoted the “Axial Reversal” policy characterized by the Fashion Street Project and various other measures of urban redevelopment. We do not have reliable and systematic data to examine the effectiveness of the Axial Reversal policy, but the news releases of the City Government indicate that the revenues of the businesses assisted by the City Government have grown from 9.7 billion NT Dollars in 2008 to 14 billion NT Dollars in 2010. The average housing price per surface unit has risen from 191,100 NT Dollars to 430,000 NT Dollars. Moreover, six out of the seven urban renewal foundation sites are located in the western part of the city.14 Nevertheless, the effects of the Axial Reversal policy on residential segregation are not clear. While the policy has reinvigorated the development momentum of the western part, the lower and middle classes with lower incomes might not be able to afford the skyrocketing housing price and their residential options have been limited. Inter-Governmental Relations of Taipei City The development of Taipei City has been conditioned by inter-governmental relations, which structure the power and resources of the city government. In what concerns the relations between the different levels of administration, the social and economic development of Taipei City has been shaped by the Taiwanese developmental state since the relocation of KMT to Taiwan. The Kuomintang was founded by Dr Sun Yat-Sen, the Founding Father of the Republic of China. The Kuomintang Government had been ruling Mainland China from the late 1910s to 1949 when it was defeated by the Chinese Communist Party in the civil war and moved to Taiwan. From 1949 to 2000, the KMT government controlled the state apparatus following the Leninist doctrine, penetrating the local governments and almost every aspect of the society. The central government controls the fiscal resources and important personnel appointment (e.g. mayors, police commissioners, as well as directors of department of budget, accounting, and statistics) of local governments. After the democratization and liberalization since the late 1980s, Taipei City and other city/ county governments have gained more power, such as the opportunity to establish autonomous statutes to promote local economic and social development. The election of mayors by public vote since the early 1990s, has strengthened the legitimacy of mayors or local political leaders 12  For the public housing policy of the city government, see the homepage of the Department of Urban Development at http://www.udd.taipei.gov.tw/public-housing/Page.aspx?categoryId=2 (in Chinese). 13  For the detailed information of the public housing in Taipei City, see the webpage of the Department of Urban Development at http://www.udd.taipei.gov.tw/webhouse/area.aspx?AN=SY (in Chinese). 14  For the details of the news releases, see the webpage of the Secretariat of Taipei City Government at http://www.tpech.gov.tw/ct.asp?xItem=2026311&ctNode=22621&mp=101001.

124

Residential Segregation in Comparative Perspective

in interacting with the central government. Nevertheless, they still rely on the distribution of the centrally-allotted tax revenues and special subsidies dedicated to special local economic development plans and constructions. According to the laws, regulations, and formulae set by the central government, Taipei City and Kaohsiung City, the only two municipal entities before the beginning of 2011, enjoyed more than 43 percent of the centrally-allotted tax revenues, while the rest was allocated to other local governments of cities/counties (39 percent) and townships (12 percent). The central government held 6 percent for comprehensive adjustment. Since 2011, some counties are merged or promoted into three municipal entities. This change means that Taipei City and Kaohsiung will no longer enjoy the lion’s share of the tax revenue because there will be five cities to share the pie of the tax revenue. Together with other municipal authorities and counties, Taipei City will face more fiscal pressure and rely more on the subsidies of the central government. For example, Taipei City Government held the Taipei International Flora Exposition (from November 6, 2010 to April 25, 2011) receiving 3.1 billion NT Dollars from the central government, which is a third of the total budget. Kuomintang rule ended in 2000 when Chen Shui-Bian won the presidential election and the Democratic Progress Party (DPP) became the ruling party for eight years. The KMT won the presidential election in 2008 and became the ruling party again. This political transition is very important, because there were political conflicts and struggles between Taipei City under the rule of Ma Yin-Jeou (1998–2002, 2002–2006), a Kuomintang mayor, and the central government controlled by the DPP. The development strategy of Taipei City was not able to obtain sufficient resources and support from the DPP government, mainly because Mayor Ma was the leading political star of the KMT to run for the presidential campaign of 2008 (Wang and Huang 2009). In the local elections in 2006, the KMT candidate defeated the DPP opponent and became the third elected mayor of Taipei City since the early 1990s. The central government still refrained from providing sufficient resources to Taipei City. For example, Taipei City obtained only 60 million NT Dollars for the Deaf Olympics, 2009, but Kaohsiung City under a DPP mayor received about six billion dollars for the World Games, 2009.15 The political contradictions between Taipei City Government and the central government have been alleviated, mainly because Ma and the KMT won the presidential elections in 2008 and 2012. Nevertheless, this does not mean that the vertical political and administrative relations between Taipei City Government and the central government have always been harmonious. The controversy of urban renewal policy and its implementation has created new political contradictions between the central government and the city government. With regard to the horizontal inter-governmental relations, the Taipei City Government has been promoting the initiative of cooperation among cities and counties in the northern Taiwan region, partly because of the need to improve metropolitan governance, and partly because of the strategy of circumventing the constraints of the DPP government. The Northern Taiwan Development Committee was established in 2005 and originates from individual cooperation projects between Taipei City and other cities in northern Taiwan. A framework of division of labor among the eight cities and counties was built to deal with the issues of metropolitan governance, such as disaster prevention and public security, environmental protection and resources sharing, industrial development, leisure and tourism, health and social welfare, transportation and communications,

15  For the intergovernmental conflicts between the Taipei City Government and the central government under the DPP from 2000 to 2008, see Wang and Huang (2009: 106–08).

A Portrait of Residential Differentiation in Taipei City (1980–2010)

125

culture and education, as well as minorities (aborigines and new immigrant spouses). Various forums and conferences have been held to examine the policy effects and make improvements.16 The horizontal inter-governmental relations between the Taipei City Government and other city/county governments of northern Taiwan do not mean that they can challenge the central government. The Taipei City Government and the Northern Taiwan Development Committee still need the support and resources of the central government. As the Department of Urban Development indicates, the Northern Taiwan metropolitan cooperation needs to solve three difficulties: weak legal basis, obscure strategic roles, and insufficient fiscal resources (Department of Urban Development 2009: 116–18). In other words, the vertical inter-governmental relations are still important in conditioning the city/county governments than the horizontal ones. The development strategy of Taipei City Government and its effect in shaping the residential distribution are highly conditioned by the industrial policy and economic planning of the central government. Residential Differentiation in Taipei City: Industrial Policy, National Spatial Planning, and Housing Policy As described above, Taipei City has been the main focus and control center of Taiwan’s post-war industrialization and economic development. However, we have no systematic data or evidence to conclude that the absence of residential segregation in Taipei City from the 1980s to 2000s is the direct effect of the developmental state’s industrial policy. The construction of high-technology parks and the developmental strategy of Taipei City since the mid-1990s might have created a number of job opportunities, which attracted certain professionals to move into Taipei City. This assumption is based on the low dissimilarity index of higher education and the growth of certain professional categories (Figure 5.3). Nevertheless, the government has not conducted comprehensive surveys or compiled databases to help us trace whether the new residents of Taipei City in the past three decades have been pulled in by the industrial policy of the central state. Those who benefited from the job opportunities created by the new industrial policy and development strategy could have moved into or out of Taipei City (and commute between Taipei City and other neighboring cities). Conversely, we can also hypothesize that some residents in Taipei City have moved out of Taipei City and residents refrained from moving into Taipei City, because they are incapable of taking the jobs resulting from the new industrial policy and development strategy, and because there are other factors that have made them decide to move out of Taipei City or continue to be commuters. However, we do not have solid surveys or data to support this hypothesis, either. There is some partial evidence, which will be discussed in the next section. In contrast to the industrial policy, the spatial planning and housing policy of the government has been more significant in shaping the residential distribution in Taipei City. The industrial policy of the Taiwanese developmental state and the developmental strategy of Taipei City Government have contributed to the industrialization and economic development of Taipei City. However, the industrial policy and development strategy have not been well coordinated with appropriate spatial planning, (public) land use, urban renewal, and housing policies, thus aggravating the problem of land/housing speculations.

16  For the details of the committee and the northern Taiwan metropolitan development plan, see the web pages of the committee: http://www.ntdc.org.tw/ and http://www.ntdc.org.tw/UploadFiles/relatedownload/01. %E7%B8%BD%E8%AB%96.pdf.

126

Residential Segregation in Comparative Perspective

In the post-war years, the central government in Taiwan formulated a series of spatial development plans, such as the “Plan for the Overall Development of Taiwan Area” (1979–1996), the “National Comprehensive Land Development Plan” (1996–2009), and the “Strategic Plan for National Spatial Development.” In the new spatial development plan drawn by the Council for Economic and Planning Development, the concept of city-region is introduced to assert inter-city cooperation to form economies of scale, in order to strengthen Taiwan’s international competitiveness (CEPD 2010: 4). These spatial development plans constitute a comprehensive blueprint for land use and urban development in Taiwan. However, the central government has not been able to enact a law to regulate the comprehensive spatial planning and land use at different scales. The consequences are very serious, because without a comprehensive framework with a legal basis to guide urban planning and land use (urban, agricultural, and industrial), the development of land is delegated to local (city and county) governments, which has been the hotbed of local land/ housing speculations.17 Taking advantage of this kind of spatial framework and housing policy, local politicians and (construction and financial) conglomerates were able to cooperate and engage in real estate speculations by buying agricultural and industrial land at low cost and transforming them into residential or commercial land with higher prices for sale. The land tax and land value-added tax are relatively low, because they are based on the assessed and publicly announced land values, rather than the real market values. This means that, the developers and financial conglomerates capture huge profits from the gap between low land costs and high housing prices. This speculation strategy is based on a symbiotic mechanism of local politicians and conglomerates. Local politicians provide information of urban planning and (re)zoning to conglomerates and manipulate zoning in exchange for large amounts of political contributions from the conglomerates. The conglomerates obtain precious information, buy the targeted land, and get huge profits by constructing residential/commercial buildings or by speculations (Chen 1995). This type of land alchemy has been one of the major forces driving up housing prices in Metropolitan Taipei since the late 1980s. Land/house speculation worsened from the late 1980s to early 1990s, partly because of the bubbles generated from lower interest rates and financial deregulation under the pressure of the United States, which led domestic surplus money and foreign capital into real estate and stock markets (Chen 2011: 310). During the 1990s, developers and investors took advantage of the revision of the master plan and detailed plans of Xinyi Planning Area, thus generating new waves of speculation by building mansions and investing in consumption-entertainment complexes, such as movie theaters, malls, and high-rise office buildings in the Area (Jou 2005: 121–35). Since the beginning of the new century, new waves of land speculations in Taipei City (and Metropolitan Taipei) have been the consequences of the new speculation strategy of property developers, financial conglomerates and real estate investors, as well as the decrease of the inheritance tax rate (down from 50 percent to 10 percent), the mortgage loan interest rate (down to 2 percent), and the land value-added tax rate (cut by half) to stimulate economic growth (Chen 2011: 311–12), Some developers adopted a new strategy of speculation and built super-luxurious houses for the super-rich or upper class. Thus, the financial incentives and the new speculation strategy have been driving up the land/house prices of neighboring areas again. Middle and lower classes are increasingly unable to afford the rapidly rising housing prices and burden of loans in the city. One of their options, and maybe the only one, is moving out of Taipei City and commute between Taipei City and other cities in northern Taiwan (Chen and Huang 2008). 17  For the details of Taiwan’s national spatial planning, see Wang (2011).

A Portrait of Residential Differentiation in Taipei City (1980–2010)

127

The new waves of land speculations have been increasingly serious due to the special housing market dynamics in Taipei City and the government’s policy of auctioning public land held by the National Property Administration under the Ministry of Finance and other central ministries (e.g. Ministry of Defense). On the one hand, the housing market in Taipei City has been characterized by high demand and insufficient supply since the mid-2000s. The skyrocketing demand originates from the new residents who need to buy or rent a house in the city, the old residents who need to find bigger or newer houses, and the investors. High demand and insufficient supply constitute one of the main forces pushing up housing prices in Taipei City (CPAMI 2011). On the other hand, the policy of selling public land of the National Property Administration has been instrumental to driving up the housing prices in Taipei City. The official claim is that selling public land helps the government to raise money to solve the worsening fiscal problem and promote urban renewal. However, the investigation of the Control Yuan finds that the policy is inappropriately implemented, because the government has not established a complete legal framework to regulate how the public land will be used by the buyers, which gives speculators good opportunities to grab huge profits by driving up the land/house prices.18 According to the investigation, the ratios of the surface area of public land sold in Taipei City and Xinbei City to the national total from 2000 to 2009 are 15.3 percent and 5.7 percent respectively, but the ratio of the price in the two cities to the national total is 60.9 percent. The ratio of Taipei City alone reaches 49.7 percent, which suggests that the land value has been artificially inflated. From 2003 to 2009, 199 lots of public land were sold, but the buyers of 62 of them never applied for construction licenses. Thirty-two of the 62 lots were re-sold to others. Thirteen of the remaining 137 lots got construction licenses, but had not yet been developed at the end of 2009. The investigation report indicates that the longer the sold public lands had not been developed, the higher the frequency of re-selling. That is, they were the targets of land speculation. There are 11 cases of selling of public land in metropolitan Taipei, and the buyers are mostly insurance companies (Control Yuan 2011). Moreover, the public land that has been sold in Taipei City is mostly located in the CBDs or fast developing blocks, which help developers and investors drive up housing prices (Chen and Xiong 2010: 83–5). This does not mean that the Taiwanese state has always been passive in regulating the housing market. Prior to the mid-1980s, there were fragmented policies of providing public housing to public servants, teachers, dependents of military personnel, and victims of natural disasters. The state was active in public housing by making comprehensive plans and preferential loans from the mid-1970s to the late 1980s to solve the residential problem resulting from the rapid urbanization and fast industrialization (such as the passage of the Public Housing Act in 1989), but the policy was not very effective in providing affordable housing to lower income households and individuals (Mii 1988, La Grange et al. 2006: 62–66). The state has been intervening in the housing market by providing financial incentives (low-interest rate mortgage and subsidies for mortgage interests), thanks to the democratization and the pressures of the three waves of social movement demanding the state to reform the housing policy since the late 1980s (Chen 2011: 312–15), on the one hand, and to the considerations of pursuing economic growth by promoting the construction and real estate industries and to those of attracting votes by providing subsidies to mortgage loan interest rates, on the other (Chen and Li 2012: 215–17). Nevertheless, the state has been playing an indirect role by formulating plans and establishing laws, providing subsidies, allocating public lands, whereas 18  The Control Yuan is a branch of the state and its function is somewhat similar to that of Ombudsman in Sweden. For the history and functions of the Control Yuan, see the homepage of the Control Yuan: http:// www.cy.gov.tw/mp21.htm.

128

Residential Segregation in Comparative Perspective

the private sector participated in construction and marketing (Doling 1999, Li 2006). Since the late 2000s, the KMT government has been formulating various policy packages to provide low-price houses for lower income households and individuals to buy and to check the real estate market by levying the luxury tax, but so far the new policies have not been very successful. Therefore, the macroeconomic and political dynamics at different periods aside, the fundamental problem is that the government has not constructed a set of comprehensive, systematic, and consistent public/social housing policies. The housing policy characterized by promoting homeownership and the strategy of using the housing policy to stimulate economic growth have prevented the Taiwanese developmental state from being able to control the land/ housing speculations effectively. The public housing policy regime in Taiwan has not been as efficacious in balancing the housing market as those of Hong Kong and Singapore, on the one hand, and those of the United States and several European countries, on the other (Chen 2011: 307–09). In Taiwan, even public housing has been highly commodified (La Grange et al. 2006: 63); subsidies for the lower-income households to rent and other disadvantaged groups (young couples, aborigines, and the working class) to apply for preferential mortgage loans have not been very effective in restraining the skyrocketing housing prices. The central state and the Taipei City Government (such as the Department of Social Welfare and the Council of Aboriginal Affairs) have provided a variety of subsidies to help the households and aborigines who need emergency aid and opening new businesses, but most of the subsidies are assisting measures, which are not able to structurally and systematically improve their situation in finding affordable houses. Theoretical Implications: The Applicability of Social Polarization Thesis, Residential Segregation, and the Developmental State Theory Theories in Taiwan The Social Polarization Thesis In view of our findings, it seems that the descriptive and explanatory power of the hypothesis of social polarization (widening income gaps, growth of upper and lower classes, shrinking of the middle class, and the concentration of new immigrants in low-paid industrial/service sectors) of the world city ( Friedmann and Wolf 1982: 320–23, Friedmann 1986) and global city theory (Sassen 2001) is limited. The thesis of social polarization is attractive because it is easy to understand and its metaphorical image is strong. Thus, many scholars have conducted case/comparative studies to examine the hypothesis. The major examples include the foreign households in Paris (1982–1990) (Rhein 1998), Toronto urban region (Walks 2001), Zurich (Koll-Schretzenmayr et al. 2005), the paid domestic workers in London (Cox and Watt 2002), London’s migrant labor from the 1980s to 1990s (May et al. 2007), Cape Town (Lemanski 2007), Singapore (Baum 1999), and Hong Kong (Chiu and Lui 2004, Lee et al. 2007). However, some scholars point out that the definition of social polarization is imprecise and professionalization is more salient than polarization in Randstad, Holland (Hamnett 1994) and sites nested in East Asian developmental states (Tai 2005, 2006 and 2010). Woodward (1995) argues that social polarization is more like a metaphor than a precise concept for case/comparative studies, while Maloutas (2007) reminds us that the concepts of social polarization, segregation, and gentrification should be differentiated clearly. Furthermore, Nørgaard (2003) indicates that Sassen’s reading of data tends to be selective and her conclusion is thus problematic.

A Portrait of Residential Differentiation in Taipei City (1980–2010)

129

Others have emphasized that contexts and local processes are important in describing and explaining social inequalities in urban development. These are overshadowed by the oversimplified and metaphorical idea of social polarization, such as in the study of Sydney, Australia by Baum (1997), the investigation of Athens (1990s) by Maloutas (2007), the role of the welfare state by Hamnett (1996), and the study of Shanghai by Li and Wu (2006). We agree with the critics of social polarization that local and national contexts and processes are very important in describing and explaining social inequalities in cities or urban development. As with other studies on many cities listed above, our research findings indicate that the social polarization hypothesis does not give us a valid explanation of the residential differentiation patterns in Taipei City. Our discussion in the next section will show how the local contexts and processes of Taipei City have been shaping the residential differentiation and the relative homogeneity in terms of socioeconomic status and ethnic groups/races in the city. Models of Residential Segregation There is a variety of ways of typifying the theories of residential segregation, and most of them focus on the correlations among races and ethnic groups, socioeconomic factors, and prejudice and discrimination. Galster and Keeney (1988) summarize relevant researches and indicate that there are three residential segregation theories: class theory based on socioeconomic indicators, selfsegregation theory based on racial preferences, and (housing and labor markets) discrimination theory. According to Charles (2003, 176–97), spatial assimilation and place stratification are the two leading theories of (racial) residential segregation in the American research literature of urban development; the former focuses on socioeconomic status differences, while the latter emphasizes the importance of prejudice and (housing) market discrimination in shaping residential segregation. Iceland and Scopilliti (2008: 80–81) indicate three theories of (immigrant) residential segregation. The first is the spatial assimilation model or theory, which contends that acculturation and socioeconomic status are significant in explaining residential segregation. This theory predicts that residential segregation would decline because individual and residential mobility would improve over time with continuous adaption and upward social mobility. The second is the ethnic disadvantage (place stratification) model or theory, which underlines that prejudice and discrimination constitute structural barriers against assimilation. Thus, the improvement of socioeconomic status would not help alleviate residential segregation. The third theory argues that the divergence of different racial and ethnic groups in residential segregation would be better described by the concept “segmented segregation,” which takes both individual and structural elements into consideration and focuses on their interaction. In view of our findings and the relative ethnic homogeneity of Taiwanese society, it is difficult to conclude that racial/ethnic prejudice and housing market discrimination have played a significant or determining role in shaping the residential distribution pattern in Taipei City. Rather, socioeconomic status has been more significant in contrast to the ethnic prejudice/discrimination factor. It is true that there are prejudices and discrimination against the aborigines and new immigrant spouses, but we have not found that prejudice and discrimination have constituted a systematic mechanism in shaping their residential distribution patterns. Spatial assimilation theory cannot fully explain the concentration of higher education/higher income in the central part and in some niches of Taipei City, either. It is based on research on the American society, and tends to be deterministic and over-optimistic. It would be difficult to apply it directly in racially/ethnically homogeneous societies. The politico-economic symbiotic

130

Residential Segregation in Comparative Perspective

mechanism based on land/housing speculations in Taipei City and other cities in Taiwan is so special that spatial assimilation theory would not be very useful. The Developmental State Theory and the Nested City Perspective Since we focus on socioeconomic status and its relations to residential differentiation, place stratification theory is not very useful here. Thus, we need to find a more convincing and contextualized theory to explain the residential differentiation patterns in Taipei City. According to the view of the nested city (Hill 2004), a city’s function, position, and development within a hierarchical urban system is conditioned by multi-level relations, local contexts, and specific processes, rather than totally determined by globalization forces. Thus, the (world or global) cities or city-regions of East Asia have been nested in the urban regime conditioned and shaped by the institution of the East Asian developmental state (Hill and Kim 2000, Wang 2003). The literature on the East Asian developmental state is huge and we are not going to enter the debate,19 but we want to argue that the developmental state has been playing significant roles in the industrial development of and residential distribution in cities in Taiwan. We intend to examine whether the spatial planning, industrial policy, and housing policy have been the major forces shaping the residential distribution pattern in Taipei City. Nevertheless, we would like to point out that, like other East Asian developmental states, the Taiwanese developmental state is no longer as powerful and resourceful as it was in the post-war era of economic take-off. It has evolved into a neo-developmental state, partly because democratization, (limited) liberalization, and privatization have weakened its political power and policy tools. Nevertheless, the Taiwanese state continues to formulate economic plans and strategies of industrial upgrading to maintain economic growth and promote industrial structural transformation. Rather than intervening directly in the economy and industrial development, the Taiwanese state has used more indirect policies and strategies to promote high-end service and high-tech industries. While evolving into a neo-developmental state, the Taiwanese developmental state has been characterized as increasingly predatory in the past two decades. From the mid-1990s to the end of the 1990s, the KMT Government under President Lee Teng-Hui (1988–2000) was engaged in political and ideological struggles against domestic political opponents and Mainland China, thus losing the opportunity of promoting Taiwan’s economic and industrial transformation by taking advantage of the rising China’s market and resources. The predatory effect was especially serious from 2000 to 2008 during the rule of Chen ShuiBian and the DPP. On the one hand, the DPP government’s industrial policy had been inconsistent and the economic bureaucracy unstable under the DPP. The main cause is that the president enjoys huge constitutional power in nominating the Premier and influencing the appointment of the cabinet members regardless of the political distribution of seats in the Legislative Yuan. During Chen’s two terms (2000–2004, 2004–2008), six premiers were nominated and the cabinet partially or wholly restructured several times. On the other hand, Chen, his family, and many DPP members serving on government posts were involved in many cases of corruption and scandals. These factors have weakened the capacities of the Taiwanese neo-developmental state. Even though the KMT won the presidential election in 2008 and regained the control over the state apparatus, the predatory effects have not been addressed immediately, partly because the problems caused by the DPP are too serious to be resolved quickly, and partly because the leadership and administrative capabilities of the KMT government have proved to be limited, failing to satisfy the 19  For a rough literature review and the debates about the East Asian development, see Wang (2011).

A Portrait of Residential Differentiation in Taipei City (1980–2010)

131

majority of voters (e.g. reforming the spatial planning regime and housing provision problem). We should take these political phenomena into consideration when we hold the views of developmental state and nested city hierarchy in exploring the residential distribution of Taipei City. Thus, the residential differentiation patterns in Taipei City since the 1980s are the mixed result of a set of macro and structural factors, especially the industrial policy of the Taiwanese neodevelopmental state not coordinated well with the spatial planning system and housing policy, the development strategy of the city government, as well as the predatory effects and the contradictory vertical intergovernmental relations during the DPP’s rule. Conclusion This chapter has sought to explore whether there has been residential segregation in Taipei City, to explicate the mechanism and dynamics of residential distribution patterns, and to examine the descriptive and explanatory power of social polarization thesis, models of residential segregation, and the developmental state theory. We constructed a conceptual framework and adopted secondary analysis to direct the reasoning of our research. Under the framework, our data based on the census indicate that residential differentiation is a better term in describing the residential distribution in Taipei City. The social polarization thesis is not only problematic, but also incapable of describing or explaining the residential distribution patterns in Taipei City. The models of residential segregation based on the American context are not applicable in the relatively ethnically/racially homogeneous contexts of Taipei City. In view of their limitations and the nested view of urban development, we argue that the developmental state’s spatial planning framework, the central government’s industrial policy and development strategy of city government, as well as the housing policy have been more influential in shaping the residential distribution in Taipei City. The findings confirm the perspective of nested city and the importance of local/national contexts. Based on our findings, we would like to underline four points. Firstly, the industrial policy of the central government and the development strategy of Taipei City have tried to establish Taipei City as the control and command center of Taiwan’s economic development and the hub of global FIRE and MICE industries. The industrial policies and development strategies have established specific urban settings in Taipei City and attracted people of higher socioeconomic status to move into Taipei City and Metropolitan Taipei. Thus, we witness the spread of residents with higher educational attainment (Map 5.2) and the growth of technicians and professionals in Taipei City in the past two decades (Figure 5.3). However, the residents with higher educational attainment and professional jobs do not necessarily work in Taipei City; some of them reside in Taipei City and commute to other cities to work. Conversely, the residents with various levels of educational attainment and holding all kinds of jobs may live in other cities and commute to Taipei City to work. Those residents who can afford the high housing prices in Taipei City (and Xinbei City) are mostly upper- and upper-middle classes with higher (individual and household) incomes. Those who cannot afford to buy or rent a house in Taipei City, including the low-income households and those with higher educational attainment, either chose to live in the peripheral areas of Taipei City or moved to the neighboring cities and counties. The residential differentiation in Taipei City must be examined from a metropolitan perspective focusing on the interaction between Taipei City and its neighboring cities. More importantly, the data suggest that the relative homogeneity of higher educational attainment and concentration of higher income households in certain quarters of Taipei City have more to

132

Residential Segregation in Comparative Perspective

do with the inconsistent spatial planning, the commodification-oriented housing policy, and the symbiotic mechanism of real estate speculations, than the developmental state’s industrial policy and the development strategy of Taipei City Government. The absence of residential segregation in Taipei City is the path-dependent result of the past three decades, especially the incomplete spatial planning and housing policy of the Taiwanese developmental state characterized with predatory effects from the mid-1990s to late 2000s. This absence also implies that only the people with higher educational attainment and higher incomes have been capable of staying in or moving into Taipei City, and that those with lower educational attainment or inferior socioeconomic status have been squeezed to the peripheral areas of Taipei City to other cities in northern Taiwan. Finally, the absence of residential segregation based on ethnic/racial prejudice and discrimination and socioeconomic status does not mean there are no clusters of lower socioeconomic status and minorities, especially those of aborigines and new immigrant spouses. The government should pay attention to the phenomena, solve the problem of outrageous property speculations, and restructure the urban regime to coordinate the industrial policy and development strategy, on the one hand, with spatial planning mechanism and the housing policy, on the other. References Baum, S. 1997. Sydney, Australia: A global city? Testing the Social Polarization Thesis. Urban Studies, 34(11), 1881–901. Baum, S. 1999. “Social transformations in the global city: Singapore.” Urban Studies 36(7), 1095– 117. CEPD (Council for Economic Planning and Development) Executive Yuan. 2010. Strategic Plan for National Spatial Development [Online]. Available at: http://www.cepd.gov.tw/encontent/ dn.aspx?uid=8708. [accessed: 1 May 2011]. Chang, C.O., Chen, S.M. and Somerville, T. 2003. “Economic and social status in household decision-making: Evidence relating to extended family mobility.” Urban Studies 40(4), 733– 46. Chang, C.O., Kao, K.F. and Li, V.C.C. 2001. “Reasonable housing prices in Taipei: Demand side analysis” (in Chinese). Journal of Housing Studies 10(1), 51–66. Charles, C.Z. 2003. “The dynamics of racial residential segregation.” Annual Review of Sociology 29, 167–207. Chen, T.S. 1995. The City of Golden Power: A Sociological Analysis of Local Factions, Conglomerates, and Urban Development. Taipei: Grand Currents Publications (in Chinese). Chen, Y.L. 2011. “New perspectives for social rental housing in Taiwan: The role of housing affordability crises and the housing movement.” International Journal of Housing Policy 11(3), 305–18. Chen, Y.L. and Li, W.D. 2012. “Neoliberalism, the Developmental State, and Housing Policy in Taiwan.” In Locating Neoliberalism in East Asia: Neoliberalized Spaces in Asian Developmental States, eds B.-G. Park, R. Child Hill and A. Saito. Oxford, UK: Blackwell, 196–224. Chen, Y.S. and Huang, Y.Y. 2008. “Spring beyond Taipei City” (in Chinese). Commonwealth Magazine 403, 38–44. Chen, Y.S. and Xiong, Y.X. 2010. “Our land, their luxury house” (in Chinese). The Commonwealth Magazine, 445, 80–88, 90. Chiu, S.W.K. and Lui, T. 2004. “Testing the global city-social polarization thesis: Hong Kong since the 1990s.” Urban Studies, 41(10), 1863–88.

A Portrait of Residential Differentiation in Taipei City (1980–2010)

133

Chou, T.L. 2005. “The Transformation of Spatial Structure: From a Monocentric to a Polycentric City.” In Globalizing Taipei: The Political Economy of Spatial Development, ed. R.Y.-W. Kwok. New York/London: Routledge, 55–77. Control Yuan. 2011. Investigation on the Policy of Public Land by National Property Administration [Online: The Control Yuan]. Available at: http://www.cy.gov.tw/sp.asp?xdUrl=./ di/edoc/eDocForm_Read.asp&ctNode=1488&AP_Code=eDoc&Func_Code=t01&case_ id=100000002 [accessed: 1 May 2011] (in Chinese). Cox, R. and Watt, P. 2002. “Globalization, polarization and the informal sector: The case of paid domestic workers in London.” Area 34(1), 39–47. CPAMI (Construction and Planning Agency, Ministry of the Interior). 2011. “An Analysis of the Housing Market in Metropolitan Taipei: Based on the Database of Housing Information.” Available at: http://ehi.cpami.gov.tw/(S(yxek3y2lyiy3bi3varc3k545))/upload/sys/study/708. pdf. [accessed: 25 September 2011] (in Chinese). Department of Urban Development, Taipei City. 2009. Northern Taiwan: Leading the Way. Taipei City: Department of Urban Development, Taipei City (in Chinese). Doling, J. 1999. “Housing policies and the little tigers: How do they compare with other industrialised countries?” Housing Studies 14(2), 229–50. Friedmann, J. 1986. “The world city hypothesis.” Development and Change 17, 69–83. Friedmann, J. and Wolff, F. 1982. “World city formation: An agenda for research and action.” International Journal of Urban and Regional Research 6(3), 309–42. Galster, G.C. and Keeney, W.M. 1988. “Race residence, discrimination, and economic opportunity: Modeling the nexus of urban racial phenomena.” Urban Affairs Quarterly 24(1), 87–117. Hamnett, C. 1994. “Social polarization in global cities: Theory and evidence.” Urban Studies 31(3), 401–24. Hamnett, C. 1996. “Social polarization, economic restructuring and welfare state regimes.” Urban Studies 33(8), 1407–30. Hill, R.C. 2004. “Cities and nested hierarchies.” International Social Science Journal 56(181), 373–84. Hill, R.C. and Kim, J.W. 2000. “Global cities and developmental states: New York, Tokyo and Seoul.” Urban Studies 37(12), 2167–95. Hsu, C.I. and Guo, S.P. 2006. “CBD oriented commuters’ mode and residential location choices in an urban area with surface streets and tail transit lines.” Journal of Urban Planning and Development 132(4), 235–46. Iceland, J. and Scopilliti, M. 2008. “Immigrant residential segregation in U.S. metropolitan areas, 1990–2000.” Demography 45(1), 79–94. Jou, S.C. 2005. “Domestic Politics in Urban Image Creation: Xinyi as the ‘Manhattan of Taipei’.” In Globalizing Taipei: The Political Economy of Spatial Development, ed. R.Y.-W. Kwok. New York/London: Routledge, 120–40. Koll-Schretzenmayr, M., Ritterhoff, F. and Siebel, W. 2005. “How global is the world city hypothesis?” DISP 163, 50–73. La Grange, A., Chang, C.O. and Yip, N.M. 2006. “Commodification and urban development: A case study of Taiwan.” Housing Studies 21(1), 53–76. Lee, K.M., Hun, W. and Law, K. 2007. “Social polarization and poverty in the global city: The cases of Hong Kong.” China Report 43(1), 1–30. Lemanski, C. 2007. “Global cities in the South: Deepening social and spatial polarisation in Cape Town.” Cities 24(6), 448–61.

134

Residential Segregation in Comparative Perspective

Li, W.D.H. 2006. “Privatising social housing in Taiwan.” International Journal of Social Welfare 16(1), 12–17. Li, Z. and Wu, F. 2006. “Socioeconomic transformations in Shanghai (1990–2000): Policy impacts in global-national-local contexts.” Cities 23(4), 250–68. Maloutas, T. 2007. “Segregation, social polarization and immigration in Athens during the 1990s: Theoretical expectations and contextual difference.” International Journal of Urban and Regional Research 31(4), 733–58. May, J., Wills, J., Datta, K., et al. 2007. “Keeping London working: Global cities, the British state and London’s new migrant division of labour.” Transactions of the Institute of British Geographers 32(2), 151–67. Mii, F.K. 1988. “Public housing policy in Taiwan.” Taiwan: A Radical Quarterly in Social Studies 1–2/3, 97–148 (in Chinese). Nørgaard, H. 2003. “The global city thesis—social polarization and changes in the distribution of wages.” Geogratiska Annaler B 85(2), 103–19. Rhein, C. 1998. “Globalisation, social change and minorities in metropolitan Paris: The emergence of new class patterns.” Urban Studies 35(3), 429–77. Sassen, S. 2001. The Global City: New York, London, Tokyo. Princeton/Oxford: Princeton University Press. Tai, P.F. 2005. “Social Polarization and Income Inequality: Migration and Urban Labour markets.” In Globalizing Taipei: The Political Economy of Spatial Development, ed. R.Y.-W. Kwok. New York/London: Routledge, 141–66. Tai, P.F. 2006. “Social polarisation: Comparing Singapore, Hong Kong and Taipei.” Urban Studies 43(10), 1737–56. Tai, P.F. 2010. “Beyond “social polarization”? A test for Asian world cities in developmental states.” International Journal of Urban and Regional Research 34(4), 743–61. Taipei City Government. 2010. Taipei City Statistical Abstract. [Online: Department of Budget, Accounting & Statistics, Taipei City Government]. Available at: http://w2.dbas.taipei.gov.tw/ NEWS_WEEKLY/summeng/HPDF/100E.PDF [accessed: 2 April 2011]. Walks, R.A. 2001. “The social ecology of the post-Fordist/global city? Economic restructuring and socio-spatial polarization in the Toronto urban region.” Urban Studies 38(3), 407–47. Wang, C.H. 2003. “Taipei as a global city: A theoretical and empirical examination.” Urban Studies 40(2), 309–34. Wang, C.H. 2006. “Planning Taipei: Nodal status, strategic planning and mode of governance.” Town Planning Review 77(3), 283–309. Wang, C.H. 2011. “Moving Toward Neoliberalization? The Restructuring of the Developmental State and Spatial Planning in Taiwan.” In Locating Neoliberalism in East Asia: Neoliberalized Spaces in Asian Developmental States, eds B.-G. Park, R. Child Hill and A. Saito. Oxford, UK: Blackwell, 167–95. Wang, J.H. and Huang, S.W. 2009. “Contesting Taipei as a world city.” City 13(1), 103–09. Woodward, R. 1995. “Approaches towards the study of social polarization in the UK.” Progress in Human Geography 19(1), 75–89.

Chapter 6

Residential Segregation and Social Structure in São Paulo: Continuity and Change since the 1990s Eduardo Marques, Renata Bichir and Celi Scalon

This chapter discusses the patterns of residential segregation in the São Paulo Metropolitan Region,1 as well as their main transformations in the recent period. São Paulo is the largest and most important metropolitan region in Brazil and Latin America, concentrating significant proportions of the country’s wealth and poverty. Historically, rich social groups tended to be located in the central areas of the region, especially in the so-called Centro Expandido, which concentrates most of the region’s amenities and opportunities (high-quality jobs, public services, leisure opportunities, etc.), whereas poor groups tended to be located in the most distant rings of the metropolitan region, in places with poor access to those opportunities. In recent years, this model of segregation has been facing some changes, especially in peripheral areas, due to several dynamics: political changes in the country, mainly the return to democracy; changes in the labor market and economic structure of the city; increasing levels of access to services and public policies, even in the peripheries; improvement of social indicators, among others. However, despite increasing heterogeneity in peripheral spaces, the macro-segregation patterns are very stable. The aim of the chapter is to present these patterns in detail and discuss the main reasons for them. Firstly, we present the literature which analyses the patterns of segregation in São Paulo, as well as the most important social, political and economic changes in the metropolitan region during the last decades. In the second section, we analyze the pattern of segregation in 2000 in terms of the distribution of social classes in space. In the third section, we explore the evolution of social segregation during the 1990s, considering the recent transformations in the economy and in state provision. We use a definition of social classes taken from the EGP classification (Erikson, Goldthorpe and Portocarrero 1979), combined with the International Index of Socioeconomic Occupational Status (ISEI) calculated for the survey areas2 of the São Paulo Metropolitan Region for the years 1991 and 2000, as well as information on income and level of schooling. This data is based on information produced by the population censuses undertaken by the Brazilian Bureau of Statistics (IBGE). Given the lack of spatial data for 1991 comparable with those for 2000,3 we discuss segregation by social classes in 2000, but use data on per capita family income—gathered by the Subway Company for 1997 and 2007—to discuss changes over time.

1  The São Paulo Metropolitan Region is officially delimited and includes 39 municipalities. 2  The survey area corresponds to an intermediate area between census tracts and districts, as used by the Brazilian Population Census in 2000, including around 8,000 inhabitants, on average for the whole country. 3  The information provided by the sample of the 1991 Census does not allow the identification of the questionnaires by survey area, preventing the analysis of São Paulo’s social structure in spatial terms for the year in question.

136

Residential Segregation in Comparative Perspective

The Legacy of Inequality and Segregation of the Metropolitan Region São Paulo is the largest and most important metropolitan region in Brazil and Latin America and is roughly 70 km from north to south and 65 km from west to east, considering the official metropolitan region. It has around 11 million inhabitants in the municipality of São Paulo and 20 million in the metropolitan region in 2010, meaning that one out of 10 Brazilians lives in this region. Between 1970 and 2000, the metropolitan region doubled in population, rising from 8.1 to 17.9 million inhabitants, but since the 1990s growth rates have fallen sharply. The city of São Paulo is considered Latin America’s most important financial and corporative hub, responsible for 12.5 percent of the Brazilian GDP in 2005. As a vivid example of the country’s inequalities, São Paulo concentrates both a significant portion of its most modern productive activities associated with globalized businesses, and a large proportion of its poor population who tend to live in widely segregated spaces with poor access to services and policies. In spatial terms, the city has also been marked historically by strong segregation patterns and high level of inequality in the access to public goods and services. The metropolitan region was consolidated in the 1960s and 1970s in a period marked by intense intra-national migration, especially from the Northeast of the country, and fast economic growth based mainly on industrial production. The working class ended up living mainly in irregular settlements on the outskirts of the metropolitan region—the so-called peripheries in what has been called the “Brazilian metropolitan model” (Santos and Bronstein 1978). These areas were located in segregated regions produced by private developers with almost no state regulation (Bonduki and Rolnik 1982), no infrastructure and very poor access to public services (Kowarick 1979, Camargo 1976, Santos and Bronstein 1978). On the other hand, wealthy people tended to live in the most central areas where the majority of amenities—jobs, urban infrastructure, services and so on—was located (Villaça 2000). The resulting concentric segregation pattern was a combined product of infrastructure distribution, selective state regulation, and land market dynamics. At the center lies the so-called Centro Expandido, or Expanded Center, occupied almost exclusively by richer social groups and containing the large majority of jobs, amenities and activities. The outer regions correspond to extended and highly segregated peripheries. In the 1980s, when the metropolitan region was already home to 12.6 million people, the distance from the remotest regions to the center had already surpassed 40 kilometers, which would take as long as three hours to travel by bus. This urban feature of the exclusionary Brazilian development model became known as urban plunder (Kowarick 1979)—the combination of increasing poverty, high macro-segregation and the lack of urbanity for most of the population, in a context of rapid economic growth. Recent analyses have shown that several important transformations have taken place in the last few decades, adding new contours to this legacy of segregation. Firstly, the spatial distribution of the different social groups inside the metropolitan region has become increasingly heterogeneous and complex, at the same time as levels of access to public policies have improved substantially, even among the poorest sectors of the population (Marques and Torres 2005). Social indicators have improved, reducing the intensity of social deprivation (Torres, Pavez and Bichir 2006, Figueiredo, Torres and Bichir 2006). In fact, access to services among the poorest people rose considerably in relation to water (from 94.0 to 96.7 percent), garbage (91.8 to 92.7 percent) and sewage collection (61.2 to 75.0 percent) between 1991 and 2004 (Figueiredo, Torres and Bichir 2006). On the other hand, social policies—especially basic healthcare and elementary education—became almost universalized, although with lower quality for the poor. A substantial part of the improvement in social indicators in the 1990s and 2000s

Residential Segregation and Social Structure in São Paulo

137

is due to the combined effects of the reform of services and policies provided by the state since the 1980s, electoral politics, and the actions of social movements, policy communities and state agencies in the new democratic political environment (Marques and Bichir 2003). This improvement was reinforced by substantial demographic changes experienced by the São Paulo Metropolitan Region in the last decades. These included a very rapid decrease of demographic growth rates - which dropped from 5.5 percent per year between 1960 and 1970 to 1.2 percent per year between 2000 and 2010—due to decreases in both fertility rates and emigrational flows. The structure of the families also changed substantially with a rapid increase of the age structure— individuals with 60 years or older represented 8.2 percent in 2000 and jumped to 12.5 percent in 2010—and a transformation of family arrangements—couples with children dropped from 68.9 to 52.8 from 1986 to 2005, while all other arrangements increased, mainly men living alone (from 1.1 to 7.3 percent), women living alone with children (10.9 to 15.3 percent) and couples without children (11.8 to 14.3 percent) (Baeninger 2011). Politically, that period represented a moment of major political change in Brazil. After a long dictatorial period that started in 1964, the civilians returned to power in 1985, as a result of an intense political mobilization in the country. The labor movement, which had returned strongly to the political scene by the late 1970s, and local urban grassroots movements were very important actors in the regime transition. These movements were reinforced by the restoration of elections for governors and mayors, in 1982 and 1985, respectively, and pushed for more participation, government accountability and public policy reform, both nationally and locally. The 1988 Constitution consolidated many of these demands for democratization, including the redesign of several public policies. Its enactment represented a redefinition of the Brazilian federal system, through a slow and complex transfer of decision-making capacities, functions and resources from the federal government to states and municipalities. Since the 1990s, the majority of national social policies—health, education, social assistance, housing, sanitation—have been reformed due to intense pressure from below and from technical and professional communities, and have led to a substantially changed Brazilian social protection system, which historically was feeble and very unequal. The metropolitan region’s peripheries became increasingly diversified, not only in urban terms, but also in terms of the social groups they house, mainly due to the growing presence of lower middle-class inhabitants. A similar phenomenon has occurred in the favelas, making it more accurate to talk about peripheries and favelas in the plural (Valladares and Préteceille 2000, CEM/ SAS 2004, Marques and Torres 2005). In some regions, gated communities sprang up, creating wealthy enclaves (Caldeira 2000), responding in particular to the spread of urban violence and a pervasive feeling of insecurity. Nevertheless, despite the improvements in public policies and reduced social heterogeneity, the center of the metropolitan region remains highly exclusive, as does the spatial concentration of opportunities and amenities. Hence, the structure of macro-segregation—with the wealthiest living in central areas and the poorest in the peripheries—remains almost untouched and appears to be the most durable feature of social inequality in São Paulo (Marques and Torres 2005). Another enduring feature of the metropolitan region is the lack of urban planning, one of the factors leading to the stable macro-segregation pattern of São Paulo. If there is something that has not changed in recent years it is the selective character of the state regulation of land markets and the absence of territorial planning in almost all Brazilian cities. An active land market policy could create incentives to counterbalance the segregation patterns, leading to some degree of social mix. However, state regulation of economic activity, land use and construction patterns operates almost exclusively in the “formal city”—the richest areas with formal housing development. By contrast,

138

Residential Segregation in Comparative Perspective

the majority of the metropolitan territory remains unregulated and is ‘planned’ almost exclusively by private developers, most of them irregular or illegal. São Paulo has also changed intensely in economic term in recent decades. While in the 1970s the metropolitan region housed 33.6 percent of the country’s labor force working in manufacturing industry, by 1980 this number had fallen to 29.3 percent and, by 1990, to 26.5 percent (Campolina Diniz and Campolina 2007). This process only partly involved deindustrialization, and a significant part of the activities remained in the city or very close by, in a process termed “concentrated deconcentration” in the literature. By the end of this process, São Paulo was exerting new economic functions related to its centrality as a major city-region, within which the activities of command, finance, innovation and knowledge production are even more concentrated today than they were in the 1970s. These changes were matched by important shifts in salaries, job instability and unemployment. In this respect, the 1990s represented a moment of sharp decline in the metropolitan labor market with a huge increase in unemployment and a big expansion in informal labor. These trends continued practically unabated until around 2003, when they were reversed and unemployment began to fall, and the degree of labor formalization tended to increase (Comin 2011). Despite these improvements, all these characteristics remain worse now than they were in the second half of the 1980s. The changes were triggered by the recent period of economic growth, as well as the public policies implemented by federal and local governments (Soares et al. 2007). As we know, several large cities in the world have experienced similar processes since the 1980s (Sassen 1991, Knox and Taylor 1995). In Brazil, several studies adopted the global cities interpretation in a more or less critical form, explaining several processes as outcomes of global processes such as the rise of unemployment, the decline of tenured jobs, the growing number of people living below the poverty line, even though trends occurred in the 1980s (Ribeiro and Telles 2000), when the Brazilian economy remained mostly closed and the impact of global processes was still small (Marques and Torres 2000). According to the local literature, these processes were likely to lead to increases in segregation, the rise of illegal or precarious housing solutions and the dissemination of wealthy gated communities in Brazilian cities. In line with the international debate,4 our position is that although several of these processes did occur and had serious impacts on urban spaces and social conditions, their specific causes lay in complex associations between local and global processes, as well as various economic— and also political, institutional and cultural—dynamics. High and durable levels of inequality in Brazil preceded the economic transformations of the last decades (Scalon 1999 and Ribeiro 2007). Likewise, strong patterns of residential segregation preceded the most recent economic and political transformations (Kowarick 1979). So, even if the recent changes really triggered important dislocations in the city, they require a detailed analysis of the social structure and the residential segregation in the metropolitan region. Exploring these processes is the aim of this chapter. As we will see, despite the intense political and economic changes that took place in Brazil during the 1990s, only localized changes have occurred in social structure, as well as in residential segregation in São Paulo Metropolitan Region. 4  Several authors have argued that various national or local structures and political processes mediate global economic processes, resulting in scenarios with only some resemblance to those predicted by the global cities hypothesis. Among those mediating processes are welfare regimes, different urban histories, land markets and local policies and politics. The list of cities analyzed is very broad, but includes London, Paris, New York, Toronto, Singapore, Athens, Madrid, Oslo and Helsinki (Hamnett 1996a and 1996b, Préteceille 2006, Leal 2007, Baum 1999, Wessel 2000, Musterd and Murie 2002, Vaattovaara and Korteinen 2003, Maloutas 2007).

Residential Segregation and Social Structure in São Paulo

139

Residential Segregation in São Paulo—2000 In this section we analyze the structure of segregation in São Paulo using social classes, income and years of schooling, initially for the whole metropolitan region and later disaggregating the analysis for intra-metropolitan spaces. Studies of social structure employ different social classifications in order to describe the social attributes and divisions within a given population. In this chapter, we make use of two strategies: the classification proposed by Erikson, Goldthorpe and Portocarrero known as the EGP class scheme, and the ISEI—International Social-Economic Index of Occupational Status. The EGP classification starts from the framework of occupations—in our case, those described in the Population Census—and reclassifies them according to a combination of position in the labor market, level of qualification, and position within production relations. This classification has the advantage of having been used innumerable times internationally (Erikson and Goldthorpe 1993, Breen 2004) and in the Brazilian context (Scalon 1999, Ribeiro 2007). The Index of Status (ISEI), on the other hand, developed by Ganzeboom, De Graaf and Treiman (1992), is used to measure the educational attributes that convert the individual’s schooling into income. In order to construct the scale, the occupations are ranked in terms of the indirect influence of education on income, taking into account occupation, age and income. We have chosen this analytic strategy because it allows us to describe the social structure and simultaneously provides a continuous variable—the index—that helps measure the status variation of the classes over time. In general, the available information suggests the existence of a stable strongly segregated metropolitan space, with small changes toward more heterogeneity in the peripheries, largely due to a greater intermixture of middle income groups there. Firstly, we discuss the residential segregation of social groups for the whole metropolitan region. We use two indexes. As well as the Dissimilarity Index, we use the Global Moran Index (Anselin 1995)5 which measures the degree of association of a given indicator for a specific area with its neighboring areas. Both indexes indicate the degree of macro-segregation in a certain territory by investigating patterns of concentration, at a given scale of analysis. The Moran Index, however, also reveals the presence of localized spatial patterns, leading to a better understanding of the distribution of segregation in space. The Global Moran index varies between 0 and 1 and can be interpreted as an index of the correlation between the value of the indicator in an area and those of its neighbors. Values closer to 0 indicate low spatial correlation while values closer to 1 indicate high spatial correlation. The Moran Global Index represents the synthesis of the correlations between areas and their neighborhoods and although it does not allow for the discussion of each space, it represents an important tool for obtaining a broad description. As is widely known, the DI measures the share of a given social group that would have to be moved in order to create a completely even distribution of social groups in the city as a whole. In this case, we considered the sum of the absolute differences between the proportions of each EGP class in a given area and the total active population in that area. Table 6.1 presents both indexes for the São Paulo Metropolitan Region (SPMR). The units of analysis were the 814 survey areas of the Census, which had an average population of 24,500 inhabitants in 2000. As we can see, the more segregated groups are the professionals and the skilled manual workers—between 0.7 and 0.8. Both non-manual routine workers (high and low levels) and

5  As the neighborhood matrix we have used the immediate neighborhood (called “first order”) and direct contiguity (queen contiguity).

140

Table 6.1

Residential Segregation in Comparative Perspective

Global Moran Index and Dissimilarity Index, São Paulo Metropolitan Region, 2000

Classes I. High level professionals II. Low level professionals IIIa. High level non-manual routine workers IIIb. Low level non-manual routine workers IVa. Employers Ivb. Self-employed workers IVc. Farmers V. Technicians and supervisors of manual work VIIa. Skilled manual workers VIIb. Unskilled manual workers VIIc. Rural workers

Moran Index

Dissimilarity Index

0.79 0.73 0.66 0.59 0.59 0.23 0.25 0.35 0.78 0.62 0.51

0.43 0.21 0.14 0.09 0.37 0.13 0.88 0.13 0.15 0.16 0.59

Source: Scalon (2006), from the 2000 Population Census, IBGE.

unskilled manual workers present somewhat lower segregation indexes—namely, 0.59 and 0.66. In spite of this difference, however, the general level of segregation is high. Dissimilarity Indexes (DI) for the same social classes showed that the highest indexes are found among the farmers and the rural workers; however, as mentioned before, those tend to be residual categories. Among the larger classes, the highest DI indexes are found among the highlevel professionals (0.43) and the employers (0.37), followed by the low-level professionals (0.21). These highly positioned classes are the most segregated ones, being concentrated in the expanded center of the São Paulo Metropolitan Region. Another way of analyzing the global pattern of macro-segregation is provided by the Moran Index of the ISEI, calculated from the average ISEI of the sample’s questionnaires in the survey area. The Global Moran Index for the ISEI is quite high (0.78), which is further evidence of the existence of an intense pattern of segregation by status in the city. Indexes of 0.69 for Rio de Janeiro and 0.73 for São Paulo are cited in Marques, Scalon and Oliveira (2008), but without including some of the rural survey areas. The lowest indexes refer to classes with a lower numerical presence and should be treated with some caution, since they are more susceptible to sample fluctuation. But how do these general patterns unfold in space? The following maps present this information, showing the ISEI distribution for 2000. As we can see, the distribution of the index is highly segregated around the Southwest sector of the so-called Centro Expandido (the city’s expanded center, highlighted in a black circle). The patch of higher status values, however, spreads out toward the Northern zone, the industrial region of the ABC paulista (toward the Southeast) and the start of the Eastern zone, as well as toward some isolated survey areas situated to the west and east. The two large areas in the West correspond to concentrations of edge-city gated communities, in the sense defined by Caldeira (2000). In very general terms, we could say that the distribution of the index is radial and concentric, but at a detailed level there is fairly significant heterogeneity. This heterogeneity tends to be greater in the intermediate areas and at the start of the peripheral areas. Similar outcomes have been found for Rio de Janeiro, Belo Horizonte and Salvador, although sometimes using different methodologies and categories (Ribeiro and Préteceille 1999, , Carvalho et al. 2004, Préteceille and Cardoso 2008, Marques, Scalon and Oliveira 2008).

Residential Segregation and Social Structure in São Paulo

Map 6.1

141

Average International Index of Socioeconomic Occupational Status (ISEI) in the survey areas, São Paulo Metropolitan Region, 2000

Note: Classes constructed considering the standard deviation of the average ISEI. The black circle indicates the city’s expanded center. Source: 2000 Population Census, IBGE.

The distribution of classes in space leads to the same interpretation: it is generically radial and concentric, but with further specific local patterns. The maps that follow display the distribution of selected classes.6 Due to space limitations, we have included as illustrations only the maps of classes located at opposite ends of the social structure—high level professionals and the unskilled manual workers. 6  It is important to point out that the maps show class distributions and not predominance in space, and a significant portion of the classes (especially the higher strata) represent minorities even in the spaces where they are most prevalent. The following exercise, therefore, provides an analysis of classes rather than spaces, which tend to be fairly heterogeneous.

142

Map 6.2

Residential Segregation in Comparative Perspective

Distribution of the high level professional class, São Paulo Metropolitan Region, 2000

Note: Classes constructed considering equal size of intervals. The black circle indicates the city’s expanded center. Source: 2000 Population Census, IBGE.

As we can see in Map 6.2, the high level professionals tend to be concentrated in the Southwest region of the city, within the Expanded Center. This segregated pattern of the upper classes is characteristic of Brazilian metropolitan regions (Villaça 2000, Marques, Scalon and Oliveira 2008) and is substantially higher than the patternfound in metropolitan regions such as Paris, where different groups among the upper classes cluster in different places (Préteceille and Cardoso 2008). Small spatial pockets toward the West, East and Southeast fail to show this pattern, however. The first two pockets correspond respectively to areas occupied by large closed condominiums (mostly gated communities) and by the center of a peripheral municipality. The third pocket marks the center of the municipalities of the ABC paulista, the stronghold of the region’s industrial activities, especially Santo André. The distribution of low level professionals is very similar, while that of

Residential Segregation and Social Structure in São Paulo

Map 6.3

143

Distribution of unskilled manual workers, São Paulo Metropolitan Region, 2000

Note: Classes constructed considering equal size of intervals. The black circle indicates the city’s expanded center. Source: 2000 Population Census, IBGE.

skilled manual workers tends to be almost the reverse, with higher concentrations in the peripheries and in the most outlying areas of the ABC region. It is interesting to note that these were the classes with higher Global Moran indexes in Table 6.1. Map 6.3 shows the distribution of unskilled manual workers. The pattern is much less clear, although certain peripheries display high concentrations of these workers. The relative presence of those workers in the Expanded Centre tends to be quite small, although some areas in this region also show a high presence of unskilled manual workers, especially a sample area in the Southern part of the region, where a large favela is located. A similar pattern is found in the distribution of high and low level non-manual routine workers, not presented here due to the lack of space. Hence, while the location of some classes shows a generic radial and concentric distribution (mainly

144

Residential Segregation in Comparative Perspective

high and low level professionals and skilled manual workers), that of the other classes cannot be described in the same manner. Indeed, it is difficult to specify any single pattern of social class distribution in the city. The data therefore indicates the existence of high segregation by social class in the São Paulo Metropolitan Region in 2000. The spatial patterns suggest a generic radial and concentric distribution of the classes. Nonetheless, when observed in detail, a more complex distribution emerges. The structure of segregation reveals a tendency for the lower social classes to be excluded from the Southwest sector (the Expanded Center) of the metropolitan region where the richest groups are located, but a substantial heterogeneity is present across the rest of the metropolitan space. Recent Transformations in São Paulo As we have seen in the first section, São Paulo has experienced intense transformations during the last two decades as a result of both national and international dynamics. These processes involve political and economic changes, which may have substantially impacted on the composition of social groups and the structure of segregation in the metropolitan region. In this section we start by exploring the transformations in social structure for the metropolitan region as a whole, and later turn to changes in segregation inside the metropolitan region. The comparison is based on samples taken from the population censuses of 1991 and 2000.7 The year 2000 was the worst socioeconomic moment of the last decade when unemployment, informality and poverty reached their highest levels and salaries their lowest. Thus, although the information below relates to class structure, the measurement points represented rather unfavorable moments in social and economic terms. Table 6.2 presents the EGP classes in the São Paulo Metropolitan Region in 1991 and 2000. Table 6.2

EPG class distribution (%), São Paulo Metropolitan Region, 1991/2000

Classes

1991

2000

I. High level professionals II. Low level professionals IIIa. Non-manual routine workers, high level IIIb. Non-manual routine workers, low level IVa. Employers IVb. Self-employed workers IVc. Farmers V. Technicians and manual work supervisors

4.6 9.7 16.0 13.4 3.4 2.0 0.1 2.0

5.8 11.3 13.9 14.8 1.9 2.9 0.0 2.4

7  Unfortunately there are no data available to analyze changes in social structure for a longer period of time. It should be emphasized, however, that important recent debates in the international literature were based on this 10-year time frame: to analyze Singapore, Baum (1999) used data for the 1983–1996 period; comparing Singapore, Taipei and Hong Kong, Tai (2005) focused on the 1993–2003 period; Wessel (2000) analyzed Oslo focusing on the 1980–1990 period; Hamnett (1996b) analyzed changes in London over the 1981–1991 period; Maloutas (2007) studied Athens based on data for 1991 and 2001; and Préteceille (2006) investigated Paris over the periods 1982–1990 and 1990–1999 respectively.

Residential Segregation and Social Structure in São Paulo

145

Classes

1991

2000

VIIa. Skilled manual workers VIIb. Unskilled manual workers VIIc. Rural workers

26.3 21.6 1.0

25.5 21.0 0.5

Total

100

100

Source: Population census data.

As we can see, the data do not reveal any great changes in terms of social structure between 1991 and 2000. There were transformations, however, with fairly important sociological implications: 1. Professionals, both high level (4.6 to 5.8) and low level (9.7 to 11.3) increased their relative share. 2. The share of high level non-manual routine workers decreased from 1991 to 2000 (16.0 to 13.0) while that of low level workers from the same category increased (13.4 to 14.8). 3. Employers decreased (3.4 to 1.9) and self-employed workers increased their relative shares over the decade (2.0 to 2.9), although their presence is small; this makes the analysis more sensitive to sample variations, suggesting that we should treat these changes with a degree of caution. 4. There has been a reduction in the relative proportions of both skilled and unskilled manual workers (26.3 to 25.5 and 21.6 to 21.0, respectively). In general, these changes are consistent with some of the economic transformations already described. Firstly, in terms of economic sectors, the categories classified as professionals and routine non-manual workers tend to be numerically more pronounced in the service sector, while manual workers are more concentrated in industrial activities. The concentration of change in professional activities is also in line with the development of services associated with the restructuring of the business enterprise world, at the same time as the relative increase in low level non-manual workers is consistent with the development of mid-level service and commercial activities, especially retail. The size of the changes also matches the kind of transformation that usually happens in social structure during a relatively short period of 10 years. On the other hand, the decrease in manual workers is consistent with the decline in industrial activities (as well as with the increase in their productivity provoked by international competition). The small increase in the number of self-employed workers might be an indication of an expansion of informality, but the size of the group is too small in statistical terms for us to be certain about this trend. In any case, even if this increase was added to the decrease in unskilled manual workers, we would end up with a very small variation in those occupations (23.6 to 23.9). Unfortunately, the available data does not allow a more detailed breakdown of the classes. Since important changes may be hidden inside the classes, we analyze now average income, education and the ISEI within each class. We have already seen that average wages and income tended to decline over the decade. In fact, that period was marked by a nominal increase in average per capita family income, which rose from R$ 493.7 to R$ 775.8. This represented an increase of 57 percent, but inflation during the period reached 155.3 percent,8 resulting in a reduction of average real income. Since we are 8 

Based on the General Market Price Index (IGPM) of the Getúlio Vargas Foundation (FGV).

Residential Segregation in Comparative Perspective

146

Figure 6.1

Total income by social class, São Paulo Metropolitan Region, 1991/2000

Note: Farmers and rural workers omitted. Source: Population census data.

only interested in the relative differences between the classes over time, we eliminated this effect by dividing the average total income of each class by the average total income of the São Paulo Metropolitan Region as a whole.9 Figure 6.1 compares the information for 1991 and 2000. Farmers and rural workers were omitted due to their small representation within the sample. As we can see, the result points to a small increase in income inequality between classes. The average income of high and low level professionals, as well as that of employers, showed relative increases when compared with the metropolitan average. On the other hand, the average income of manual workers, both skilled and unskilled, tended to fall, especially among the latter group. Among the intermediate classes, relative income tended toward stability. The standard deviations remained relatively stable within each class. On the whole, the changes in relative income confirm an increase in income inequality during the decade, caused especially by the professional classes. However the size of the changes tends to be small and the stability of the standard deviations does not suggest an increase in social heterogeneity within each class. This interpretation is complemented by an analysis of education. Figure 6.2 shows the average years of schooling for each class for 1991 and 2000. Average schooling increased from 6.6 to 8.8 9  This procedure also solves the problems of inflation and changes in the relative prices of goods.

Residential Segregation and Social Structure in São Paulo

Figure 6.2

147

Years of schooling by social classes, São Paulo Metropolitan Region, 1991/2000

Note: Farmers and rural workers omitted. Source: Population census data.

years during the decade, due in particular to improved access to public education even among the poorest social groups (Figueiredo, Torres and Bichir 2006). Although the situation at the end of the decade was still a low level of schooling (with great inequalities between classes), Figure 6.2 confirms the overall improvement, showing increases in schooling levels for all social classes. It is important to add that the largest increases are found among the intermediate and low classes—high and low non-manual routine workers, and skilled and unskilled manual workers—who showed lower levels of schooling in 1991.10 Again farmers and rural workers were omitted and again the standard deviations tended to stability within each class. In conclusion, then, there are no signs of increases in the internal heterogeneity of the social classes. Finally we can turn to the average Index of Socioeconomic Status (ISEI) of each class. The situation of the metropolitan region is again one of relative stability, with the average index moving from 44 in 1991 to 45 in 2000. This suggests that, on average, the distribution of the educational attributes that convert the individuals’ education into income remained stable regardless of the changes in education (or precisely because they tended to be generalized, maintaining the same structure, but at a higher level). Figure 6.3 presents the ISEI averages by classes in 1991 and 2000, with the groups ranked in decreasing order according to the 1991 index. As we can see, on average there is no significant variation within the classes, except, perhaps, in relation to the technicians and supervisors of manual work, whose index rating improved (from 48.1 to 53.4). Low level 10  All with increase rates between 35 and 52 percent of the previous levels of schooling.

148

Figure 6.3

Residential Segregation in Comparative Perspective

International Index of Socioeconomic Occupational Status (ISEI) by social classes, São Paulo Metropolitan Region, 1991/2000

Source: Population census data.

non-manual routine workers and skilled manual workers also saw an increase, but of smaller magnitude. The indexes for employers and self-employed workers worsened slightly, though this fall was relatively small and may be due to sample variation. The other classes remained with ISEI values very close to those found at the beginning of the decade. The ISEI standard deviations by classes tended to be stable as well. Thus what is noticeable about this data as a whole is the general stability of the social classes, despite some increase in income inequality among classes and a general improvement in education, even for the intermediate and lower classes. This suggests that the occupations making up these strata tended to be of slightly better quality, although the changes are too small to establish any clear pattern. At the same time, the presence (and income) of the professional classes increased. Unfortunately, the absence of spatial information in the sample data of the 1991 Population Census prevents us from making an over-time comparative study of social classes in space. Instead, to give an idea of what has happened recently in terms of changes in segregation, we use data on per capita family income from a different source of data: the Subway Company Origin and Destination survey (OD research) for 1997 and 2007. The spatial units in this case are the 389 zones defined by the 1997 OD research. Each zone had, on average, 43,200 inhabitants and 12,200 households.11 In order to eliminate the effects of the different income levels (and inflation rates) for each of the dates, we divided all the data by the average for the metropolitan region for each year. 11  The São Paulo Subway Company runs the Origin and Destination survey (OD research) every 10 years since 1987. Further information may be obtained at www.metro.sp.gov.br/empresa/pesquisas/origem.

Residential Segregation and Social Structure in São Paulo

Figure 6.4

149

Distributions of the average income of the area/average metropolitan income in São Paulo, 1997/2007

Source: Origin and Destination survey, Subway Company.

Figure 6.4 shows the distributions of the resulting income indicators in 1997 and 2007 per decile. Each curve presents the level of the income indicator for each decile. As we can see, the 2007 curve presents higher income levels for the first deciles, indicating a relative improvement in the income conditions of the lowest social areas. From the 6th decile up, the curves are similar and at the top of the distribution the 2007 areas present slightly worse income levels. Thus, the income inequality among the areas became smaller. Since the income inequalities among classes have increased—as we saw earlier—the reduction of inequalities among areas points to greater social heterogeneity inside the poorest areas, leading to a reduction in segregation. This confirms the results of other studies relating to the 1990s which also pointed to stability, or to a small reduction in segregation, due especially to a more even distribution of middle class groups in the peripheries, although the center of the metropolitan region continues to be highly exclusive. Torres (2005) and Torres and Bichir (2009) identified stability in the Dissimilarity Index for education in São Paulo between 1991 and 2000 and a slight increase for income. Marques, Gonçalves and Saraiva (2005) also reported stability using factorial analysis of a wide range of indicators to study the social structure in the city. At the same time, they found signs of a small reduction in residential segregation by social indicators, possibly associated with the increase in the average level of public policy access rather than with any significant changes in social group location (and segregation).

150

Residential Segregation in Comparative Perspective

Final Considerations This chapter has showed that the spatial distribution of social classes in São Paulo Metropolitan Region tends to be highly segregated, with an almost exclusive center and more mixed peripheries. These latter areas show greater internal heterogeneity both due to the distribution of low middleclass groups in a mosaic pattern, and to the presence of wealthy gated communities within poor neighborhoods, pointing to a complex combination of spatial proximity with social distance. The evidence also pointed to the stability of this pattern during the last decade, with only localized changes. We analyzed a period of major transformations provoked by both national and international dynamics, some of them still in progress: great changes in the provision of public services and equipment even among the poorest sectors of the city’s population, large-scale sector reconfigurations, changes in the world of labor, unemployment and wages, alterations in poverty levels. The information presented here suggests that none of these dynamics have had a significant impact on the city’s rigid class structure, although we did observe a tendency toward a slight increase in income inequality among classes and a decrease in income inequalities among the areas of the metropolitan region, suggesting increased social heterogeneity, especially in the peripheries. References Anselin, L. 1995. “Local Indicator of Spatial Association—LISA.” Geographical Analysis 27, 91–115. Baeninger, R. 2011. “Crecimiento de la población en la Región Metropolitana de São Paulo: Deconstruyendo mitos del siglo XX.” In São Paulo: miradas cruzadas: sociedade, política y cultura, eds L. Kowarcik and E. Marques. Quito: Olacchi, 53–78. Baum, S. 1999. “Social Transformations in the Global City: Singapore.” Urban Studies 36(7), 1095–117. Bonduki, N. and Rolnik, R. 1982. “Periferia da Grande São Paulo: reprodução do espaço como expediente de reprodução da força de trabalho”. In A Produção Capitalista da Casa (e da cidade) do Brasil Industrial, ed. E. Maricato, São Paulo: Alfa-ômega, 117–54. Breen, R. 2004. Social Mobility in Europe. Oxford: Oxford University Press. Caldeira, T. 2000. City of Walls: Crime, Segregation and Citizenship in São Paulo. Berkeley: The University of California Press. Camargo, C. 1976. São Paulo, 1975 – Crescimento e pobreza. São Paulo: Ed. Loyola. Campolina Diniz, C. and Campolina, B. 2007. “A região metropolitana de São Paulo: Reestruturação, re-espacialização e novas funções.” Revista Eure XXXIII(98), 27–43. Carvalho, I., Souza, Â. and Pereira, G. 2004. “Polarização e Segregação Socioespacial em uma Metrópole Periférica.” Caderno CRH 17(41), 281–97. CEM/SAS. 2004. Mapa da Vulnerabilidade Social da População da Cidade de São Paulo. São Paulo: CEM/Cebrap, SAS/PMSP. Comin, A. 2011. “Cidades-Regiões ou hiper-concentração do desenvolvimento? O debate visto do Sul.” In São Paulo: Novos percursos e atores: Sociedade, cultura e política, 34th edn., eds L. Kowarick and E. Marques. São Paulo, 167–202. Erikson, R. and Goldthorpe, J.H. 1993. The Constant Flux. Oxford: Oxford University Press. Erikson, R., Goldthorpe, J.H. and Portocarrero, L. 1979. “Intergenerational Class Mobility in Three Western European Societies.” British Journal of Sociology 30, 415–41.

Residential Segregation and Social Structure in São Paulo

151

Figueiredo, A., Torres, H. and Bichir, R. 2006. Renda e votos: o democrático toma lá dá cá. Revista Insight Inteligência, XI (33), 2006, 40–8. Ganzeboom, H., De Graaf, P. and Treiman, D. 1992. “A Standard International Socio-Economic Index of Occupational Status.” Social Science Research 21, 1–56. Hamnett, C. 1996a. “Why Sassen is wrong: A response to Burgers.” Urban Studies 33(1), 107–10. Hamnett, C. 1996b. “Social polarization, economic restructuring and welfare state regimes.” Urban Studies 33(8), 1407–30. Knox, P. and Taylor, P. 1995. World Cities in a World-System. Cambridge: Cambridge University Press. Kowarick, L. 1979. A Espoliação Urbana. Rio de Janeiro: Paz e Terra. Leal Maldonado, J. 2007. “Desigualdad social, segregación y mercado de vivienda en Madrid.” In Vivienda y segregación en las grandes ciudades europeas, ed. J.L. Maldonado. Madrid: Universidad Computense de Madrid, 15–46. Maloutas, T. 2007. “Segregation, social polarization and immigration in Athens Turing the 1990’s: Theoretical expectations and contextual difference.” International Journal of Urban and Regional Research 31(4), 733–58. Marques, E. and Bichir, R. 2003. “Public policies, political cleavages and urban space: State infrastructure policies in São Paulo, Brazil – 1975–2000.” International Journal of Urban and Regional Research 27(4), 811–27. Marques, E. and Torres, H. 2000. “As transformações recentes em São Paulo e o debate das cidades globais.” Novos Estudos CEBRAP 56, 139–68. Marques, E. and Torres, H. 2005. São Paulo: Segregação, pobreza e desigualdades sociais. São Paulo: Editora SENAC. Marques, E., Gonçalves, R. and Saraiva, C. 2005. “As condições sociais na metrópole de São Paulo na década de 1990.” Novos Estudos CEBRAP 73, 89–110. Marques, E., Scalon, C. and Oliveira, M. 2008. “Comparando Estruturas Sociais no Rio de Janeiro e São Paulo.” Dados 55, 1. Musterd, S. and Murie, A. eds. 2002. The Spatial Dimensions of Urban Social Exclusion and Integration. Available at: http://ec.europa.eu/research/social-sciences/pdf/finalreport/ soe2ct983072-final-report.pdf [accessed 09/04/2012]. Préteceille, E. 2006. “La ségrégation sociale a-t-elle augmenté? La métropole parisiense entre polarization et mixité.” Sociétés Contemporaines 62, 69–93. Préteceille, E. and Cardoso, A. 2008. “Rio de Janeiro y São Paulo: Ciudades duales? Comparación con Paris.” Ciudad y Território XL(158), 617–40. Ribeiro, C. 2007. Estrutura de classe e mobilidade social no Brasil. São Paulo: Edusc/Anpocs. Ribeiro, L. and Préteceille, E. 1999. “Tendências da Segregação Social em Metrópoles Globais e Desiguais: Paris e Rio de Janeiro nos Anos 80.” Revista Brasileira de Ciencias Sociais 14(40), 143–62. Ribeiro, L. and Telles, E. 2000. “Rio de Janeiro: Emerging Dualization in a Historically Unequal City.” In Globalizing Cities: A New Spatial Order? eds P. Marcuse and R. Kempen. Oxford: Blackwell Pub. Santos, C. and Bronstein, O. 1978. “Meta-urbanização – o caso do Rio de Janeiro.” Revista de Administração Municipal 25(149), 28–41. Sassen, S. 1991. The Global City: New York, London, Tokyo. Princeton: Princeton University Press. Scalon, C. 1999. Mobilidade Social no Brasil: Padrões e tendências. Rio de Janeiro: Revan. Scalon, C. 2006. Relatório final de bolsa de pós-doutoramento. São Paulo: CEM/Fapesp.

152

Residential Segregation in Comparative Perspective

Soares, S., Osorio, R., Soares, F., et al. 2007. Conditional Cash Transfers in Brazil, Chile and Mexico: Impacts upon Inequality. Brasília: IPC (IPC Working Paper 35). Tai, P. 2005. “Social Polarisation: Comparing Singapore, Hong Kong and Taipei.” Urban Studies 43(10), 1737–56. Torres, H. 2005. “Medindo a segregação.” In São Paulo: Segregação, pobreza e desigualdades sociais, eds E. Marques and H. Torres. São Paulo: Ed. SENAC, 81–100. Torres, H. and Bichir, R. 2009. “Residential Segregation in São Paulo: Consequences for Urban Policies.” In Urban Segregation and Governance in the Americas, eds B. Roberts and R. Wilson. Austin, Texas: Palgrave Macmillian, 145–66. Torres, H., Pavez, T. and Bichir, R. 2006. “Uma pobreza diferente? Mudanças no padrão de consumo da população de baixa renda.” Novos Estudos CEBRAP 74, 16–23. Vaattovaara, M. and Kortteinen, M. 2003. “Beyond polarization verssus professionalization? A case study of the development of the Hensinki region, Finland.” Urban Studies 40(11), 2127– 45. Valladares, L. and Préteceille, E. 2000. “Favela, favelas: Unidade ou diversidade da favela carioca.” In O futuro das metrópoles: desigualdades e governabilidade, ed. L. Queiroz. Rio de Janeiro, Observatório: Ed. Revan. Villaça, F. 1999. “Uma contribuição para a história do planejamento urbano no Brasil.” In O processo de urbanização no Brasil, eds C. Deák and S. Schiffer. São Paulo: Edusp/Fupam, 169–243. Villaça, F. 2000. Espaço intra-urbano no Brasil. São Paulo: Ed. Nobel. Wessel, T. 2000. “Social polarization and socioeconomic segregation in a welfare state: The case of Oslo.” Urban Studies 37(11), 1947–67.

Chapter 7

Segregation, Social Mix and Public Policies in Paris Edmond Préteceille

Introduction Images of cars burning and young men throwing stones and Molotov cocktails at the police have been seen worldwide on TV in 2005. For many who saw them and heard them presented as ethnic riots, it must have convinced them that the French banlieues had really become similar to the violent ghettos of US cities. A US journalist even commented that Paris had become a more dangerous city than Baghdad. We have there two basic elements of the global media construction of representations: a fascination for violence and dramatization; a taste for global interpretations with a US bias. Probably no one had any interest in taking the time to comment later on that no one had been killed during the riots in spite of their length and intensity, and that the ethnic character was debatable. This may be an extreme case, but it highlights the fact that segregation is an issue for which there is a constant interaction between the media exposition of dramatic news, the discourse and actions of politicians who feel the need to make strong public statements, and the contributions of social scientists. Segregation in the Paris metropolis has been and is today a contentious issue. It has been an object of controversy between social scientists, it has been an issue for social movements and it has inspired various public policies. The purpose of this chapter is to provide an account of the main results on and interpretations of urban segregation in Paris produced by social scientists in recent decades in relation to the social and policy debates of the time. One underlying argument in this contribution is the need to consider the historical dimension of urban processes. The first section gives a rapid outline of contributions and debates since the 1950s. The second section presents in greater detail the main results regarding recent changes in the structure and intensity of socioeconomic segregation in the Paris metropolis. The third section complements it with the analysis of ethno-racial segregation. The fourth section discusses the main processes that have been identified as causes, in order to explain both the general trends and intensity of segregation, its relative and moderate character in the Paris case, its main spatial features. And the final section discusses more specifically the various public policies addressing segregation or having an impact upon it. Half a Century of Segregation Studies Segregation in the Paris metropolis began to be studied as such by social scientists after World War II. Before that, there were since the 19th century studies of living conditions of the poor, in the line of the hygienist movement, with a strong focus on slums (taudis), but there had been no systematic effort at mapping the social structure of the city of the kind developed for London by Booth (1902). Social movements had addressed the issue of segregation on various occasions,

154

Residential Segregation in Comparative Perspective

such as the protests against the Haussmann renewal policy pushing the working class out of certain neighborhoods or the uncontrolled growth of peripheral working class settlements devoid of transport and urban infrastructure (mouvement des mal-lotis, see Fourcaut 2000) but not in a systematic way. Chombart de Lauwe (Chombart de Lauwe et al. 1952) was the first to study and map the distribution of social classes and related urban inequalities, considering not only the central city of Paris, but the whole urban agglomeration. His work may be related to the perspectives opened by Park and Burgess, whom he explicitly mentions, for the spatial dimension of the analysis, but even more so to the social morphology of Halbwachs. The social class categorization that he used was a largely shared vision in those post-war years, when the legitimacy of the Résistance and Libération reforms went with a critical appraisal of a capitalist system that had led to the defeat and collaboration. The same Marxist inspiration of social class categories can be found in the official census occupation categories established by INSEE (Institut National des Statistiques et Études Économiques) in the late 1940s, the first version of CS, categories socioprofessionnelles (see Desrosières and Thévenot 1988). Segregation became a central category of urban sociology with the neo-marxist perspective that developed in the late 1960s, and the Paris case was definitely the major example of the critique of the social consequences of the new capitalist urban policies, since the new forms of capitalist interests in the city were particularly concentrated there. Castells (1972), Lojkine (1972) and others proposed slightly different readings of those capitalist interests and policies, but converged on a characterization of their consequences in terms of increasing class segregation, the working class being pushed out of the central city into suburban developments of poor urban quality. Their views converged with many urban social movements of the late 1960s in Paris against urban renewal, and the critique of the “rénovation-déportation” that was strongly expressed in May–June 1968, and was captured by Lefebvre in his claim for the “Right to the city” (1968). Lojkine focused on the increasing urban inequalities affecting the working class, essentially defined by the growing disjunction between trends in job location and in new working class housing, and by the unequal access to the transport system, both contributing to an increasing contrast between the center and the banlieue, and the west (more bourgeois, more jobs) and east (more working class, fewer jobs) of the agglomeration, previously discussed by Chombart de Lauwe and various geographers. Castells, who was the first to show a detailed knowledge of the developments of urban sociology in the US, used it not to “import” the US paradigm but on the contrary to argue for a more complex reading both of social dimensions of segregation and of urban inequalities (1977: 169–71). He did not however develop his arguments into a systematic analysis of segregation in Paris, but rather incorporated them in an analysis of processes contributing to segregation at both ends: the expulsion of the working class from central areas through urban renewal (Godard et al. 1973) and the concentration of working class and lower middle class households in suburban housing estates (Castells et al. 1978). A key issue in his approach was a more complex view of social structure than the binary class opposition, with a growing interest in the middle classes and their contribution to social movements, that he would subsequently develop much beyond the Paris case (Castells 1983). In the early 1970s French social scientists began to have access both to computers and to more detailed census data. Freyssenet, Regazzola and Retel (1971) used those new resources to produce the first systematic study of segregation in the Paris metropolis. They described the social profile of municipalities using five social categories—professionals and business owners; intermediate categories; white collar workers; skilled blue collar workers, unskilled blue collar workers—based on INSEE’s catégories socioprofessionnelles, for the 1954, 1962 and 1968 censuses. Their results give the first sketch of the relative complexity of the social structure and its changes in space,

Segregation, Social Mix and Public Policies in Paris

155

through the calculation of indexes of segregation and spatial association between categories, and through a cluster analysis producing eight types, several of which clearly have a socially mixed profile. Although they produced a strong empirical contribution, more systematic than Castells’ or Lojkine’s, it had little impact for various reasons that can be hypothesized. First of all, their work was only published as a research report in the Centre de Sociologie Urbaine series, with a limited circulation, and no transformation into academic journal articles, the authors moving away to other subjects after that. Secondly, the report included almost no theoretical discussion, whether in the initial statement of the problem they dealt with, or in a conclusion: the results are synthesized in 10 pages at the beginning, and it is essentially descriptive, not engaging in any debate with the various views expressed in the literature at the time, whether regarding the theoretical debate about social classes or the intensity and trends of change of segregation in Paris—this may have been a choice of the authors to neutralize their theoretical differences or disagreements. Thirdly, we may think that their approach and results were positioned too differently from the main debates of the times, although in many ways they corresponded to the more complex view of the socio-spatial structure that Castells explicitly mentioned in mostly theoretical terms. That research was the beginning, however, of a steady stream of research following the same path, i.e. using more detailed social categories to assess the similarities and differences in their urban situation, therefore trying to make out empirically which socioeconomic categories shared the same urban conditions and which other categories they were in contrast with, instead of predetermining homogeneous groups and their oppositions—as most of the “class” analyses of the time did. The main contributions in this perspective on the Paris case were those of Tabard, Rhein, and our own work. Tabard, investigating the relation between the work situation of persons and their urban consumption conditions (Clapier and Tabard 1981), developed a series of cluster analyses of the Paris metropolis with a categorization that used detailed catégories socioprofessionnelles combined with the economic sector of activity of employed persons, but for statistical reasons she limited her category to men only. She subsequently expanded her work to France as a whole (Chenu and Tabard 1993, Tabard 1993). Rhein, being interested in gender relations and the social composition of households, developed a categorization of households combining the occupations of the partners when there was more than one active person (Rhein 1986, 1994). Our own approach, after first using Tabard’s results (Pinçon-Charlot, Préteceille and Rendu 1986) then detailed CS of men and women (Préteceille 2000) was to combine the detailed CS with the type of job situation—stable, precarious, unemployed—in order to address the new questions about changing social structures derived from mass unemployment and the growing casualization of labor (Préteceille 2003). All three approaches also converged on the use of cluster analysis, stressing the importance of describing the various social configurations of municipalities or neighborhoods, of assessing the relative importance of the different situations within the whole range from completely segregated areas to completely mixed ones. This became a major stake and dividing line after the end of the 1970s, when new urban policies were progressively defined more and more exclusively as targeting “problem areas.” The media representation of urban problems became almost exclusively focused on such areas, because of a series of urban riots from the early 1980s onwards, and because public policy responding to the events defined the “urban question” as the question posed by those neighborhoods, “quartiers en difficulté,” today officially called zones urbaines sensibles (ZUS). To stress the point, the ministry in charge of that policy took the name of ministry of the city (ministère de la ville) and the largest part of the academic production on segregation and cities became almost exclusively focused on those neighborhoods. Some of these studies are undoubtedly monographs of high quality which produced a detailed knowledge of local situations and a deeper understanding of social processes and social relations in those neighborhoods with

156

Residential Segregation in Comparative Perspective

a high concentration of poverty, of unemployment, of immigrants, of young people experiencing difficulties in school and in access to the labor market, of drug dealing and violence (see for example Dubet and Lapeyronnie 1992, Lepoutre 1997, Kokoreff 2003, Masclet 2003, Lapeyronnie 2008). The main problem with this literature is that it leads to an uncontrolled and often implicit but systematic generalization from situations and problems of those areas to urban situations and problems in cities as a whole.1 It has validated the media representation of urban problems which always favors the most dramatic representation, and therefore has progressively imposed a vision of segregation which is essentially dualistic: the “normal city” against the poor areas. It was also a regression from the previous binary reading of the class division of the city, which, although too simplistic, at least stressed that you could not understand the working class part of the city without considering capitalist interests in the production of the city and the capitalist class part of the city. The new dualist reading disregarded the role of the dominant class interests, translating into urban terms the paradigm of “social exclusion” which became successful in the same period to describe the social question disconnecting it from class relations (the word exploitation almost disappearing from the vocabulary). The media push this image of a dual city to the caricature, stating it day after day as an opposition between the city (la ville) and the suburbs (la banlieue) where the banlieue has come to be synonymous with large public housing estates with a poor population, many immigrants, violent youth and drug dealing etc. This caricature is absurd: one only has to mention the fact that the municipality of Paris has two million inhabitants, thus the whole banlieue has 10 million, out of which about a million and a half live in one of such neighborhoods, the zones urbaines sensibles. But it is incredibly resistant, and has had an enduring effect on the conventional wisdom of the urban question in Paris and in France. During the last 30 years, the media representation of the problem areas in the suburbs has progressively shifted from a socioeconomic characterization—social exclusion, largely defined by unemployment, underemployment and poverty—to an ethno-racial characterization, leading to a vision of those areas as immigrant ghettos. Social scientists researching those areas in a monographic way have contributed to that image by studying mainly the idle young men controlling the public space of public housing estates, often school dropouts, sometimes connected to drug dealing, many of them children of immigrants from North Africa or Sub-Saharan Africa. There is less of a tradition of systematic study of the spatial distribution of immigrants, and except for a few contributions like that of Guillon (1992), most researchers studying immigrants in the city have focused on areas of high concentration: the shanty towns (bidonvilles) in the 1960s and early 1970s; the institutions built to accommodate single working men (foyers SONACOTRA etc.); and the various typical immigrant neighborhoods like Belleville, la Goutte d’Or, the Chinese area in the XIIIth arrondissement, the Turkish neighborhood, etc. Therefore the predominant image given by the social science literature is that immigrants live in immigrant neighborhoods, an image which contributes to validate the relevance of the US case as a point of comparison.

1  The most recent example is Lapeyronnie’s recent book (2008), based on a case study of a poor neighborhood in a medium sized city. The title, Ghetto urbain, expresses the thesis, and the subtitle, Segregation, violence and poverty in France today, summarizes the effort in the book to generalize the results.

Segregation, Social Mix and Public Policies in Paris

157

Class Segregation: Polarizing Trends in a Mixed Structure The systematic study of socioeconomic segregation in the Paris metropolis2 (Préteceille 2003, 2006) tells a different story, definitely not that of a dual city. This study, following Park and Chombart de Lauwe, has to consider the whole metropolis, not just the small central part which is the historical center and the tourist area; the metropolis as a whole is the socioeconomic system within which processes of unequal spatial distribution of residence take place. The use of catégories socioprofessionnelles (Table 7.1) provides a much more sociologically meaningful analysis, as discussed earlier, than the use of income or education categories which is the only way to study socioeconomic segregation in countries where occupational data are not accessible or of poor quality. Table 7.1

Paris metropolis—active persons by occupation (CS*)

CS (socioprofessional categories)

1982

1990

Farmers Craftsmen, shopkeepers, business owners Professionals Intermediate categories White collar workers Blue collar workers

18 301 726 998 1,606 1,233

0% 6% 15% 20% 33% 25%

Total

4,882 100%

13 313 1,019 1,185 1,587 1,184

1999 0% 6% 19% 22% 30% 22%

5,300 100%

8 285 1,157 1,351 1,636 984

2007 0% 5% 21% 25% 30% 18%

5,421 100%

7 264 1,517 1,539 1,661 909

0% 4% 26% 26% 28% 15%

5,896 100%

Note: * Simplest version of the CS—the one used in our analysis was the more detailed one. Source: INSEE, censuses—Absolute numbers in thousands.

First of all, the strongest segregation is not that of the poor but that of the rich. The upper categories (business owners, liberal professions, private sector professionals) are those with the highest segregation indexes. The results are clear although even the detailed statistical categories mix the real upper class with groups which are more middle class; specific studies of the power and business elite have shown their spatial concentration to be even higher than our figures (for Paris, see Pinçon and Pinçon-Charlot 1989). In our cluster analysis of the Paris metropolis neighborhoods, we found that more than half of the members of the upper categories lived in areas where they were overrepresented—we named them upper status areas. And in those upper status areas, we found that almost half of the population belonged to the upper categories and to business owners, and less than 30 percent to the white collar or blue collar worker categories (employés and ouvriers). These neighborhoods were spatially concentrated in 19993 in about two thirds of the inner city of Paris, the core of which were the traditional bourgeois quarters, and in the extension of these into the nearby suburbs west 2  We will synthesize here the results of two decades of research which have been published in detail in many articles and book chapters. We will select those results which we think are more meaningful in the context of this comparative book; the complete results, together with data and methodological considerations, can be easily found in the main publications we will refer to. 3  1999 was the year of the last census for which we have had access to detailed data. INSEE has published data for the recent 2007 census, but with aggregate variables only.

158

Residential Segregation in Comparative Perspective

of Paris, forming a compact triangle which includes municipalities like Neuilly-sur-Seine, SaintGermain en Laye, Saint-Cloud, Sceaux, Versailles …). The taste of upper categories for the western parts of Paris suburbs is a heritage of history, since those areas were traditional areas of castles and hunting for the aristocracy of Ancien Régime, and those uses have marked space in both symbolic and practical terms: urbanization has taken different forms, dividing the previous forests or large estates into bigger plots than elsewhere, producing a less dense type of urbanization with a greater presence of trees and green spaces. It is also interesting to note that there are very few upper status areas outside that center-west triangle, and that there are almost no areas of that status in the outer part of the metropolis—outer suburbs are not an option for upper categories, unlike US suburbs. This intense self-segregation of the upper categories has tended to increase in the last decades. In spite of their growth in absolute numbers, the segregation of private sector professionals has increased, as measured by their segregation index (Table 7.2). Table 7.2

Paris metropolis—segregation index for detailed occupations (CS*)

Categories

1990

1999

Variation

Craftsmen Shopkeepers Business owners Liberal professions (non salaried) Civil service professionals Professors, literary and scientific professions Professionals in media, arts and entertainment Professionals in manag. and commercial activ. Engineers and technical professionals in firms Unemployed professionals Precarious professionals School teachers and similar occup. Health and social work intermediate occup. Civil service intermediate professions Adm. and commercial interm. occup. in firms Technicians Foremen Unemployed intermediate occupations Precarious intermediate occupations Civil service white collar workers Office workers in firms Commerce workers Service workers Unemployed white collar workers Precarious white collar workers Skilled blue collar workers—manufacturing Skilled blue collar workers—craft and repair activ. Drivers Skilled workers—storage and transport

0,247 0,267 0,447 0,419 0,330 0,346 0,455 0,294 0,270 0,409 0,451 0,214 0,212 0,272 0,143 0,217 0,256 0,263 0,289 0,198 0,133 0,195 0,194 0,202 0,218 0,255 0,192 0,269 0,331

0,241 0,259 0,440 0,412 0,309 0,322 0,441 0,315 0,279 0,351 0,400 0,214 0,197 0,254 0,142 0,210 0,269 0,194 0,217 0,210 0,148 0,193 0,167 0,220 0,189 0,297 0,212 0,299 0,356

-0,006 -0,008 -0,007 -0,006 -0,022 -0,024 -0,015 0,022 0,008 -0,058 -0,050 0,000 -0,015 -0,018 -0,001 -0,007 0,013 -0,070 -0,072 0,011 0,014 -0,002 -0,027 0,018 -0,030 0,042 0,019 0,030 0,025

Segregation, Social Mix and Public Policies in Paris

159

Categories

1990

1999

Variation

Unskilled blue collar workers—manufacturing Unskilled blue collar workers—craft and repair activ. Apprentices Interim workers Precarious blue collar workers Unemployed blue collar workers Unemployed—never worked

0,330

0,334

0,004

0,280

0,308

0,028

0,368 0,290 0,313 0,314 0,390

0,305 0,282 0,319 0,322 0,366

-0,063 -0,008 0,007 0,008 -0,024

Note: * Detailed version of the CS combined with (un)employment and stability/precariousness. Source: INSEE, censuses—data by IRIS, neighborhoods with approximately 2000 inhabitants; no similar data for 2007 available yet.

This increase has taken two complementary spatial forms, first an increase of their concentration in the areas already belonging to the upper status clusters (Map 7.1)—and a decrease of all other categories—where they represented in 1999 a little less than half of the total population; and second a strong growth in adjacent areas, expanding the upper status areas into neighboring ones that were of middle-mixed status or even of working class status for a few. Upper categories are not only those more distant from the others in terms of dissimilarity but also those whose areas of concentration are the most clustered together (as can be measured by Moran’s index of spatial autocorrelation). Not all upper or upper-middle categories have experienced such an increase in segregation, not all live in those upper status areas. Public sector professionals, and researchers and professors, have seen a decrease of their segregation, and tend to be more spread out in the rest of Paris and in the various suburbs. Professionals in the media, arts and entertainment are highly segregated but have also seen their segregation decrease, and their areas of concentration, among the most central, tend to differ substantially from private sector professionals, with a movement toward not the west but the north-east of Paris and adjacent municipalities like Montreuil.4 Such results, briefly stated, show the importance of identifying precisely the various components of the category of professionals, and the growing social and spatial distances among them, which clearly contradict such theories as Florida’s (2002) creative class. The next most segregated categories are those of blue collar workers. Although this is a classic result, it is quite significant in a city which in France is the one where tertiary activities are the most predominant, and blue collar workers, with 15 percent of the active population in 2007, are now a smaller group than white collar ones. The segregation of blue collar workers has increased also, and since they are declining in numbers, it means that they tend to remain more concentrated in the neighborhoods where they were more present. Their distribution is the one most different from that of the upper categories, particularly private sector professionals, and this socio-spatial class opposition, typical of the industrial capitalist city, remains the predominant factor of contrast in the Paris metropolis of the global service economy. This is another aspect of segregation showing the importance of history: as for the upper class areas, the working class neighborhoods were historically those with the worst conditions, often close to heavy industry, noise and pollution. They have been permanently marked, in symbolic terms, as stigmatized areas, and have also maintained in time worse urban characteristics—higher density, less green space, less facilities etc.—although 4  For a more detailed analysis of that particular category, see Préteceille (2010).

160

Map 7.1

Residential Segregation in Comparative Perspective

Paris metropolis 1999—Cluster analysis* of the socioeconomic profile of neighborhoods (IRIS)

Note: * This is a simplified version in 3 classes of the detailed analysis in 18 classes presented in Préteceille 2003.

left-wing municipal policies have substantially improved living conditions in many of them (see Pinçon-Charlot, Préteceille and Rendu 1986). Although white collar workers are among the categories with the lowest levels of segregation, they have increased their contribution, together with blue collar workers, to the socio-spatial contrast between working class and upper categories, particularly the sub-categories of precarious and unemployed workers. Therefore there has been some shift from the traditional bi-polar industrial class contrast to an opposition between a wider range of working categories with a strong incidence of unemployment and precariousness, and upper categories. This could be interpreted as a partial validation of one aspect of the global city model (Sassen 1991) but only as an emerging trend, the blue collar working class—upper class contrast remaining the predominant structuring element. Another partial validation of the global city model may be found in the increasing distance between the extremes of the distribution: the upper class areas are becoming more exclusive still, whereas part of the working class areas are becoming more impoverished, with an increasing proportion of unemployed and precarious workers. However, this enhanced polarization, this increasing social distance between the extremes, does not mean dualization—the idea that the

Segregation, Social Mix and Public Policies in Paris

161

urban society would become split into two parts, socially divided between the global elite and the tertiary proletariat, spatially divided between affluent areas, old and new, and underprivileged ones and slums, as stated by Sassen in her analysis of the global city. The main evidence against dualization is that middle classes (upper-middle categories, but also middle-middle and lower middle ones), instead of disappearing, have increased in absolute numbers and in share of the total, and that middle-middle and lower-middle categories, together with white collar workers, are little segregated, and their segregation has even decreased for most subcategories. This means that they are present in all types of areas, from upper class ones to blue collar working class ones. But they are also present in intermediate areas, clusters of neighborhoods we call middle-mixed areas (Map 7.1), where those middle categories are slightly overrepresented but where some upper-middle categories are significantly present, although slightly underrepresented, and blue collar workers as well, in average or somewhat less than average proportions. Such middle-mixed spaces, far from being marginal in the urban social structure, are the largest group in our cluster analysis, accounting for about 45 percent of the population in 1999. The importance of those middle-mixed areas is a major counter-evidence against the global city dualization model. It also contradicts the discourse of the French politique de la ville in favor of recreating the supposedly lost urban social mix (mixité sociale). Together with the low and decreasing segregation of most middle categories, it goes against the two theoretical models which have been quite successful in the recent French debate. The first one is Donzelot’s idea of the “secession” of the middle classes which he considers to be the main cause of the claimed increasing segregation of French cities (Donzelot 1999), transposing in a way the US “white flight” analysis. While this may be part of the process of impoverishment of the most deprived areas, as we will discuss later, it is by no means a general feature of segregation trends as a whole; on the contrary, middle-middle classes are clearly part of the categories maintaining the existence of a large degree of social mix, in contrast with the increasingly self-segregating upper categories. The second model has been presented by Maurin (2004) who claims that urban segregation is due to a general taste of all social categories for living together with members of the same social groups, and avoiding mixing with the others—searching for entre-soi—that he sees as a “generalised social separatism.” His empirical demonstration is flawed (Préteceille 2006: 77) and our evidence suggests that if upper categories definitely have a strong taste for staying together and keeping the others out—as Pinçon and Pinçon-Charlot (1989) have analyzed in detail—this is not the case for many others, including middle categories. Ethno-racial Segregation: Stronger than Class but Still Moderate The issue of ethno-racial segregation has become the main issue in the public debate about urban segregation in France, as we have already mentioned. The generally accepted view is that the segregation of immigrants has increased considerably, overshadowing socioeconomic segregation and leading to the formation of ghettos. The evidence supporting this view is the situation in a number of neighborhoods, the worst cases of quartiers en difficulté, which have attracted most attention because of violent episodes, whether due to fights between youth gangs, gunfights between drug dealers, or urban riots like those of 2005. The media have built this equivalence between immigrants, poor neighborhoods becoming ghettos, and violence. Monographs on those neighborhoods have contributed to validate this image, as we argued earlier. Many researchers, foreign but also French, think that it is not possible to study ethno-racial segregation in France because of the absence of “ethnic statistics,” i.e. data registering the “race”

162

Residential Segregation in Comparative Perspective

or “colour” of persons in the census. Therefore only qualitative studies could tackle the problem. In fact, this is not the case, and the French statistical system has been producing ethnic statistics for many decades, in a soft version of the measure of ethnicity5 which records people’s place/ country of birth and nationality, including previous nationality in the case of those who have become French. It is true, however, that such data have not been easily accessible to researchers, and though accessibility has improved, it is still limited and controlled due to the rules established by CNIL (Commission Nationale Informatique et Libertés). We were able to use such data from the 1982, 1990 and 1999 censuses6 to study the spatial distribution of immigrants from various groups of origins in the Paris metropolis.7 As for socioeconomic segregation, we used the two complementary methods of calculation of indexes and of cluster analysis, in order to capture synthetically the main contrasts and trends between categories, and to describe the variety of local configurations of separation or mix between categories. Immigrants represented 14.7 percent of the total population of the Paris metropolis in 1999, and 17.1 percent in 2007. Regarding the second generation, only those living with their parents can be identified in the census data; they are mainly children, and mainly belong to the most recent immigration groups; the children of at least one immigrant parent living with their parents represented 9.7 percent of the total population in 1999. The first major result is that during the last decade of the twentieth century, ethno-racial segregation in the Paris metropolis, as measured by the dissimilarity index, has only increased a little, with a limited increase for some groups, stability or limited decrease for others: beyond all the statistical uncertainties, the results definitely invalidate the idea of an explosion of ethno-racial segregation (Table 7.3). Table 7.3

Paris metropolis—dissimilarity index8 between immigrants and native French born in mainland France

D

1982

1990

1999

Turkey USA, Canada, NZ, Australia East and South East Asia India, Pakistan Algeria Morocco Tunisia South-Sahara Africa

0,431 0,544 0,347 0,349 0,317 0,349 0,368 0,357

0,448 0,503 0,346 0,338 0,319 0,327 0,360 0,333

0,470 0,468 0,344 0,340 0,334 0,333 0,332 0,330

5  Héran (2010) has demonstrated that quite extensively in the final report of the COMEDD—an official committee nominated to give an advice about how to improve the measurement of discrimination and diversity. 6  As of this writing, we have not had access yet to the 2007 census detailed data to study the most recent trends. 7  We use the origin of immigrants as a proxy of imputed ethno-racial characteristics that might lead to ethno-racial discrimination; for a more detailed discussion of this theoretical point and for the detailed methods and results, see Préteceille (2011). 8  We calculated the dissimilarity index for immigrants from the different groups of origins in the table compared to French born in mainland France; children of immigrants living with their parents were included in the immigrant group—this is the only part of the “second generation” that the census can identify—and excluded from the French group.

Segregation, Social Mix and Public Policies in Paris

163

D

1982

1990

1999

Middle East Northern Europe French born DOM-TOM* Latin America Eastern Europe Spain Portugal Italy

0,439 0,309 0,284 0,403 0,295 0,246 0,197 0,186

0,370 0,313 0,287 0,315 0,281 0,231 0,187 0,193

0,312 0,308 0,295 0,261 0,259 0,202 0,197 0,186

Note: * French born, born in overseas French territories, like Martinique, Guadeloupe or Réunion. Source: INSEE, censuses—data by municipality, and quartiers for central Paris, average size 23000 inhabitants; no similar data for 2007 available yet.

A second complementary result is that the level of this segregation can be considered moderate when compared (see note 7 for the rationale of the comparison) to the standard reference of ethnoracial segregation in US cities: controlling for the size of urban areas, it was about 0.40 in 1999, whereas it was about 0.80 in New York or Chicago (Logan 2001–2004). The classic interpretation is that 40 percent of the group of immigrants considered would have to move in Paris to obtain the same distribution as the French group, compared to 80 percent of Blacks in New York: a significant minority can thus be considered as strongly segregated in Paris, as compared to a very large majority in NY; the order of magnitude is widely different, even taking into account the uncertainties resulting from quite different modes of categorization. A third result is that although moderate, this segregation of immigrants is significantly stronger than what would be expected from their economic situation only. At the same spatial scale, the modal intensity of immigrant segregation index is 20 to 50 percent higher than the segregation index of blue collar worker categories, who are the most segregated of the working class categories. A striking result underlines this difference, the case of French born in overseas areas (like Martinique, Guadeloupe etc.) who have a dissimilarity index just a little lower than non-European immigrants, but stronger than Portuguese immigrants, although their occupations are mainly in the white collar or intermediate categories, the two groups of less segregated occupations. Such results clearly support the hypothesis of a specific segregation of immigrants as a combination of socioeconomic segregation but also of a significant ethno-racial segregation, i.e. effects of ethnoracial discrimination. A fourth result is that the cluster analysis reveals that the majority of situations are situations of mixing between the various groups of origins. Unlike the US, we do not see in Paris a mosaic of neighborhoods each dominated by one particular ethno-racial group; areas of higher presence of immigrants are areas where most if not all groups of origins mix. And if some groups have a slightly more specific profile of spatial distribution, like the Portuguese or the East Asians, it never produces a complete specialization of the neighborhoods. This result clearly invalidates the claim made by a growing number of politicians and journalists that ethno-racial segregation is the result of increasing “communitarianism,” i.e. a voluntary choice to stay within one’s own group—a pernicious extension of the entre-soi theory discussed earlier. The fifth and last result that is important to mention here is that the majority of urban situations are also situations of residential mix between immigrants and French born (Map 7.2). At the scale of small areas of about 8,000 inhabitants, about 70 percent of all immigrants lived in neighborhoods

164

Map 7.2

Residential Segregation in Comparative Perspective

Paris metropolis 1999—share of immigrants* from Maghreb and South-Sahara Africa in the total population of neighborhoods (TRIRIS)

Note: * Immigrants plus children of immigrants living with their parents. Source: INSEE, Recensement de la Population de 1999.

where they represented less than 30 percent of the local population. And less than 20 percent of immigrants (including the second generation we could identify) live in neighborhoods where they represent more than half of the local population. When we looked at which neighborhoods were in that particular situation, we found, not surprisingly, the list of those which are systematically put forward as emblems of the poor-unemployed-immigrant-violent banlieues (Clichy-sous-bois, La Courneuve, Grigny, Mantes-la-Jolie, Gennevilliers, etc.). Our results thus confirm that these neighborhoods are definitely “problem areas” since they are outstanding both in terms of their high levels of socioeconomic segregation and ethno-racial segregation, and because in some cases the trend has been toward a further increase of unemployment and precariousness, for a population with a growing share of immigrants, leading to dramatic social urban problems. But for the same reason these results also invalidate the widespread implicit claim that such neighborhoods are representative of the banlieue, and that their situation is the typical situation of immigrants in the Paris metropolis. On the contrary, we can see that they are exceptional cases, concerning a minority only of immigrants.

Segregation, Social Mix and Public Policies in Paris

165

Processes Explaining Segregation The situations of absolute segregation are explained by relatively simple causes, mainly political. The original Jewish Ghetto (Wirth 1928), the apartheid townships in South Africa, the Black ghetto of Jim Crow America, are products of laws, enforced by repressive measures, imposing on specific dominated groups a separate residence from the category in power. There is complexity in the explanation of the historical establishment and enforcement of such policies, but not so much in the processes they organize. In the Paris case where segregation is relative, not absolute, the explanation cannot be as simple, and the analysis has to be able to explain not only the forces tending to separate various groups, but also the actual intensity of this separation (particularly in comparison with other cases similar in principle but different in intensity) and the degree of mix, as well as the specific spatial forms that this relative segregation/mix takes in the city. Not only are there different dimensions of segregation, as Massey and Denton (1988) have argued, but there are quite different local situations in one particular city for members of the same group: the relative segregation in a city like Paris is a complex socio-spatial structure, experiencing complex changes, and this complexity is what we have to explain through the analysis of processes influencing the distribution of social groups in urban space. This is an ambitious research program in itself, of which we have as yet completed some elements only, and the presentation of which would greatly exceed the space available here. We will thus mention only some aspects which we feel are particularly relevant to a comparative discussion. The major process producing segregation, common to all capitalist cities, is economic: it is the interaction between the economic hierarchization of places through the capitalist process of production and circulation of housing and the income inequalities between households of different socioeconomic groups. We found a correlation of 0.906 between the socioeconomic status of neighborhoods in the Paris metropolis (as measured by the value of the first factor of a correspondence analysis of the distribution of socioeconomic categories) and the median household income per consumption unit in 2001. Housing prices are the filter allocating different social groups in the economic hierarchy of places. Upper class groups choose the best places for them, which become the most expensive, the next group in terms of income choose the next best places they can afford, and so on until the poorest groups are left with only the worst places or no housing at all. This classic scheme of the capitalist city has worked more intensely than ever in Paris in the last decades for two complementary reasons. The first is that land and housing prices at the top end have gone up tremendously because of the concentration of income in firms and households involved in the financial domination of the economy which is a main feature of the present stage of capitalism, particularly in the global cities which are the key nodes of this domination, as analyzed by Sassen (1991). All the more so because firms and very high income groups (French but also rich foreigners) are competing for the same spaces, firms looking for locations which have the symbolic quality derived from the presence of upper class residents, as Pinçon and Pinçon-Charlot have shown for Paris (1992). The second is that the increase in financial sector profits implies a reduced share of the social product for most other groups, producing an increase in income inequalities affecting most social categories including middle class ones; with a general process of precarization of labor and growth of unemployment which also tends to affect all categories; and a contraction of welfare support under the pressure of neoliberal policies, affecting particularly the poorest and less skilled groups—this was less true in France during the 1990s due to left-wing social policies but has become more intense in recent years.

166

Residential Segregation in Comparative Perspective

Increasing housing prices pulled up by the top end of the market and impoverished and fragilized households of working class groups but also middle class ones end up in a trend of more intense spatial polarization. This is particularly clear in the city of Paris itself, the core of the metropolis, where upper class and private sector upper middle class groups have progressively expanded their traditional areas of concentration, pushing out all other social categories. Similar polarizing forces linked to financial globalization have been identified in other global cities like London (Hamnett 2003). They take a specifically centripetal-centrifugal form in Paris due to the fact that central areas have been the favorite places of the upper class for a long time—the center being also the concentration of symbolic values, monuments, places of the powers etc—whereas they do not valorize the outer suburbs for residence, unlike London or New York. Evidently, such processes will also be more or less marked according to the general intensity of inequalities in the country and city, which depends upon labor market conditions and fiscal policies. France was among developed countries a country of moderate intensity of inequalities, somewhere between the more equal Nordic countries and the more unequal USA, but Paris has been a more unequal place in France, mainly due to the concentration of the higher income groups. The global elite has substantially augmented its share of the domestic product in the last decade, reducing the share of all other groups, so that there is a serious risk that socioeconomic segregation will have increased in the most recent period. If these economic processes were the only ones organizing the spatial distribution of social groups in the city, however, one would expect a higher level of segregation, with a more complete hierarchization and much less social mix. For example, about one third of inner Paris, in the northeast and some of the south-east, is quite mixed with a significant presence still of working class categories. And there is a strong presence of working class categories in neighborhoods of the first ring of suburbs, which are near the center, have good urban infrastructure and facilities and excellent transport connections. There are various explanations for this. The first has to do with the historical dimension of urban processes. The city, as built environment, as socio-spatial structure, as a system of symbolic values, is not the result of present economic processes only, but the accumulation of layers of economic processes of decades if not centuries, each one transforming to a degree the results of previous ones. If working class categories are still significantly present in areas where the economic logic of the day should have expelled them entirely, it is because the type of housing there is either social housing (we will discuss later the contribution of public policies) or low quality housing which is not very attractive for gentrification; because the social image of those areas does not make them possible places of residence for many upper middle class categories, due to the inertia of the social profile, due to the actions of institutions like municipalities that have tried to maintain it, due to the persistence of industrial images of neighborhoods even when the industry has disappeared, etc. These processes explain the apparent paradox of a segregation pattern being primarily structured by the opposition between blue collar workers and upper categories in a city where manufacturing industry has been for several decades only a minority part of the economic structure. The second explanation has to do with the fact that in most of those historical phases of socioeconomic processes of production of the city, there has been production of a noticeable degree of de facto social mix. Not as a goal, but as an outcome of a fragmented production of space unable to homogenize urban space in terms of type of housing, quality of infrastructure and services, etc. beyond the scale of a building or development of a few dozen houses. Even the Haussmann renovation of Paris, the first strong wave of capitalist production of the city for a market which was mainly upper and middle class, created an urban space less homogenous than its reputation from the façades on the main streets, since the back part of the land plots was in

Segregation, Social Mix and Public Policies in Paris

167

most cases left for low quality working class housing. It is only since the 1960s that developers have become capable of creating new homogenous developments at the scale of neighborhoods of several hundred dwellings—whether for social housing estates, or upper class ones (cf. Cousin 2008), or middle class ones—leading in Paris to the debate about deliberate avoidance or deliberate production of social mix in new development or urban redevelopment programs of a large scale, that will be discussed later. The debate about the causes of segregation has traditionally opposed structural causes— economic or political, like Massey and Denton’s thesis (1993) for US cities—to individual choice processes, for which the most famous theoretical model is probably Schelling’s (1978). The analysis of the Paris case shows that we have to combine the two kinds of explanations in various ways. First of all, we have presented the self-segregation of upper categories as one of the major elements triggering the economic process of segregation. But this self-segregation itself is not only an economic process, though the economic aspect is important in the residential choice of upper categories, to protect their real estate investment; other important elements of the choice have to do with the symbolic value of the address, with the proximity to members of the same upper social circles, with the protection from the proximity of “undesirable others,” etc. (Pinçon and PinçonCharlot 1989). Secondly, all social categories, except perhaps the poorest, have, to a varying degree, some degree of choice within their economic constraints. Some elements of these choices contribute to the general hierarchization of space. One such factor of increasing importance is the growing concern of parents about the quality of schools that their children attend, related to the intensifying competition on the labor market. Since the quality of schools partly reflects the cultural resources transmitted by the families, those who can afford it –middle-middle class and upper-middle class parents mainly—will search for upper status residential locations, contributing to increased segregation, residential as well as in schools (Oberti 2007, van Zanten 2009). But there are also non-economic elements of choice, well documented by demographers (for Paris, see for example the contributions of Le Bras and Chesnais 1976, Bonvalet and Merlin 1988) that do not necessarily contribute to greater segregation, such as the proximity of family members, or the household life cycle. One classic result in this perspective, reported in many cities since the Chicago School first highlighted it in the 1920s, is the concentration of new immigrants in areas where people of the same origin are already present; there are factors of structural constraint in this, since recent immigrants being poorer might only find housing in the worst areas abandoned by previous ones who have experienced some upward mobility; but there are also factors of social relations, relatives or people of the same locality of origin living in the area who are the first contacts and can host them for some time, help them get access to the labor market, to institutions etc. Such processes definitely play a role in the more intense concentration of immigrants, particularly recent ones, in a number of neighborhoods. Pan Ké Shon (2009) has studied the changing profiles of ZUS (zones urbaines sensibles) and shown that socioeconomic segregation in those neighborhoods decreased, but that segregation of immigrants increased; he also showed that those neighborhoods represent in a large proportion the first place of residence for newcomers—as did the traditional immigrant slums of the Chicago School, or the traditional immigrant quartiers of Paris such as Belleville or La Goutte d’Or (Pan Ké Shon and Scodellaro 2011). Can this explanation be generalized to all social groups, following Maurin’s thesis (2004) of a “generalised social separatism”? If this were the case, and since it would work in the same direction as economic processes, we should observe a more intense segregation, whether social or ethnoracial, and much fewer situations of social mix. Trends in segregation indexes as well as survey and qualitative research have validated such a valorization of homogeneity for upper class categories

168

Residential Segregation in Comparative Perspective

((Pinçon and Pinçon-Charlot 1989) and for private sector executives and professionals (Cousin 2008); but segregation indexes have shown opposite trends for a large proportion of middle class categories (Oberti and Préteceille 2004) highlighting a positive evaluation of living in socially mixed neighborhoods for many of them, particularly for middle-middle and lower-middle class members who come from a working class background and maintain social, cultural and solidarity ties with their class of origin (Oberti and Préteceille 2011). Similarly the same research shows a generally positive appreciation of ethno-racial mix in mixed areas. Likewise, survey research (Brouard and Tiberj 2005) has shown that when asked whether they would prefer a mixed residential environment or a homogenous one with people of the same origin, a mixed school or a homogenous one, etc., immigrants largely respond in favor of the mixed option. An unequivocal answer which contradicts the claim by more and more politicians that ethno-racial segregation is the result of a “communitarian” attitude of immigrants who prefer to stay within their own group and reject integration. And in the poor neighborhoods of ZUS, immigrants who achieve some upward mobility tend to move out, as do their children when they attain a better status (Pan Ké Shon and Scodellaro 2011). However, we have seen that ethno-racial segregation in Paris, although moderate relative to the US case, is substantially higher than socioeconomic segregation of the most segregated socioeconomic categories. For most immigrant groups, socioeconomic category may explain a significant part of their residential segregation, but definitely not all of it. Even considering that immigrants may be more recent and less skilled workers within their category cannot explain the difference, as we highlighted earlier in the case of French citizens born in overseas French départements or territoires (DOM-TOM). This supports the hypothesis that specific ethno-racial discrimination processes are in operation, strengthening the role of economic processes. First of all, there may be processes of discrimination at the workplace or on the labor market leading to weaker economic resources for immigrants compared to non-immigrants of the same occupation category. There are now many case studies giving evidence of such discrimination, and several wider statistical studies (see for example Silberman et al. 2007); but there is yet no overall evaluation of its extension and effects, neither for France in general nor for the Paris metropolis. And the Paris case is not necessarily comparable to the whole country in that respect, having a much longer term tradition of a cosmopolitan city. Secondly, there is also substantial evidence of discrimination processes in the private rental market, as a number of testing studies have shown, revealing discrimination on behalf of real estate agents and/or private landlords. But again, it is impossible as of now to evaluate their overall effects. There may be also more difficulties in obtaining loans due to discrimination against immigrants by the banks but this is not clearly documented. More recently, discrimination in access to housing has been established by the TEO (Trajectoires et Origines) survey by INED (Institut National d’Études Démographiques) (Pan Ké Shon and Scodellaro 2011). Thirdly, a significant proportion of the segregation of immigrants is related to the spatial distribution of social housing. There are more contradictory claims here about the type of ethnoracial discrimination at play. Some NGOs fighting discrimination have sued social housing institutions—in some cases successfully—claiming that they discriminated against immigrants based on illegal ethno-racial profiling, and subsequently reduced their opportunities of access to social housing—that argument goes against a segregation of immigrants due to social housing. Other NGOs or researchers have claimed that illegal ethno-racial profiling has been used to disperse immigrants in the housing stock—which also argues against segregation due to social housing. Still others, though, argue that immigrants are being selected for and concentrated in the worst estates or buildings to avoid or limit their presence in others—an argument that works in

Segregation, Social Mix and Public Policies in Paris

169

favor of ethno-racial discrimination through social housing (Bourgeois 1996, Masclet 2003). Since case studies reveal evidence of all three different types of discrimination in different social housing institutions—a highly fragmented world in France—and in different municipalities, it is difficult to make out what the overall outcome may be; at best one can explain the situation and trends of concentration of immigrants in specific neighborhoods where such studies have provided enough evidence. Any stronger claim about the relation between discrimination in social housing and ethno-racial segregation would at this point go beyond what social science research has proved. There is one type of neighborhood however where the concentration of specific poor working class categories with a strong proportion of immigrants is related to social housing,9 that of public housing estates considered “problem areas,” quartiers en difficultés, or, to use the administrative term, zones urbaines sensibles (ZUS) (see Tissot 2007). As we established, they represent only a minority of the social housing stock, and only a minority of the working class and of immigrants live there. But these situations contribute to the overall levels of segregation, and some of them also to the increase in polarization between the two ends of the socio-urban distribution. They may be considered as the worst urban areas, because of the cumulative effects of poverty and social fragility, of stigmatization, of low quality schools, and sometimes also poor accessibility and services. The overrepresentation in social housing estates of poor households of working class categories is first of all a mechanical effect of the rules of access to social housing. Their stronger overrepresentation in some ZUS neighborhoods results partly from discriminatory practices of concentrating the poorest tenants and immigrants in some of the estates that we already mentioned. But it results also from non-institutional factors. In some “better” social housing estates, tenants may stay even when their social status has improved beyond the norm of access10 because they like the neighborhood, because they have family there, because the location is convenient etc. In the “worst” estates, they will have left; many households within the norm will move out when they can afford it; only poorer households who have the least housing opportunities will apply for or accept an apartment in those estates; and only the poorest households will stay because they cannot move out.11 Some have interpreted this process as a kind of “white flight” to avoid places of concentration of immigrants. This may be a factor to some extent, since qualitative research has shown that these working class or low middle class households who value social and ethno-racial mix when it takes place in mixed areas, as we mentioned earlier, are more critical of it when it refers to situations where it means their possibly becoming a minority among a larger group of poor immigrants, difficult school problems etc. But it is difficult to disentangle the social from the ethno-racial in that reaction, and immigrants or children of immigrants who experience some upward social mobility seem to move out of those areas in quite the same way.

9  Although not exclusively: there are cases where the concentration of poverty takes place in private property buildings, and the situation may be worse there because it is more difficult for local authorities to intervene to help improve the situation, even just physically, in terms of the maintenance of buildings. 10  They have to pay in that case an additional rent. 11  Nivière, in the 2005 ONZUS (Observatoire National des Zones Urbaines Sensibles) report, shows that the spatial mobility in ZUS is higher than generally thought, but that it is quite selective in terms of who moves in and out. The results are for the whole of France but we have reason to believe that they apply similarly to the Paris case, if not more strongly because of the higher cost of housing on the private market.

170

Residential Segregation in Comparative Perspective

Public Policies Addressing Segregation There is today an apparent general political consensus regarding the idea that urban segregation is negative, contrary to France’s republican principles of equality, and therefore has to be reduced. Before that consensus emerged in the 1990s, segregation was not considered an issue by the mainly conservative governments, and the urban policies of the most active government on the urban scene, that of De Gaulle in the 1960s, were implicitly, and sometimes explicitly, promoting social segregation. This was done, particularly in Paris, through a process of improvement and modernization of the city center destroying low quality working class housing through urban renewal or rehabilitation, replacing it with better housing for the middle or upper middle class, and relocating working class inhabitants in new social housing estates outside the central city, long before the words state-led gentrification came to be used by social scientists. This relocation could take place in Paris because the central government could impose such new social housing developments to suburban municipalities, and the city of Paris itself built social housing on other municipalities’ territory where it owned the land or where the central state could provide it, like in La Courneuve. The central state and right-wing municipalities were not the only initiators of public policies working in favor of stronger segregation. Many left-wing municipalities, mainly communist-led in the Paris metropolis until the late 1970s, did contribute to segregation too, although in a different way. Their policy was to promote on their territory social housing for the working class almost exclusively, and to block as much as they could housing projects for other social categories. The rationale for that explicit policy was that they saw it as their responsibility to provide housing (plus many other services, for which they were very active) for the working class, which many other municipalities did not. They also favored public provision against market provision, and opposed any project they perceived as led by the search for speculative profit. They also wanted to strengthen local working class communities culturally and politically—not to mention strengthening their own electoral base. Such a policy, implemented for many decades (some municipalities in the Paris metropolis have had a communist mayor since the creation of the Communist Party in 1920) has resulted in local situations where more than half of the housing stock, sometimes 70% or 80%, is social housing. This has unequivocally contributed to increasing social segregation of the working class. However, in a significant number of cases that increased segregation did not mean increased urban inequality, due to the high level of local public services that a significant number of municipalities of the “red belt” of Paris could provide to their working class population, because of their fairly high resources: being municipalities with lots of manufacturing industries on their territories, from which they received high local taxes which they could use to provide services for the population; a form of local municipal redistribution which was the base of what has been called “municipal socialism.” Not all working class municipalities, however, were able to undertake such redistribution, because many of them were more “dormitory” areas than industrial ones, therefore had little fiscal resources coming from firms; the difference between fiscally poor and rich municipalities had major social consequences, and even more so after the decentralisation laws of the early 1980s (Préteceille 1991). Paradoxically, what first contributed to increased class segregation contributed in later decades to its limitation and to maintain social mixing in some areas. Typically, working class households— and among them many immigrant households—would have disappeared from central Paris if it were not for the existence of social housing. That social housing is mainly concentrated in the northeast quarter, 18th, 19th and 20th arrondissements, and at the periphery of the municipal territory, between the boulevards des maréchaux and the ring highway of the boulevard périphérique. When

Segregation, Social Mix and Public Policies in Paris

171

it was built—during the 1930s and then during the 1950s and 1960s—these were still working class areas, and new social housing just reinforced that characteristic. But later on the numerical and territorial expansion of the upper and upper middle classes stimulated gentrification throughout Paris—although in different forms (Préteceille 2007)—and social housing became a factor of social mix after having been a factor of segregation. And the same is progressively happening to many municipalities adjacent to Paris which, being well located, well connected and well serviced, are now a land of expansion of private sector real estate development, both for housing and office space. Left-wing municipalities adjacent to Paris like Ivry, Montreuil, Saint-Denis, Saint-Ouen, or even Nanterre, on the other side of the new business district of La Défense west of Paris, have changed their policy as we will discuss now, and no longer refuse private developer projects but rather negotiate with them to obtain a proportion of social housing in all new projects. Today, France is one of the relatively few European countries to have public policies whose explicit goal is to reduce segregation. The oldest ones, beginning in the late 1970s, are the policies focused on the “problem areas” of zones urbaines sensibles. They aim at improving the urban conditions in those neighborhoods—housing conditions, public services, security, etc.—and promote greater social mix there. Although the concentration of immigrants is seen by most actors and the media as a large part of the problem, this is not acknowledged officially (Kirszbaum 1999)— the share of immigrants in the local population is not among the indicators used to characterize the target areas; the reports of ONZUS, monitoring the results of the policy, are quite shy on the matter and hardly discuss the issues regarding immigrants (except the very latest one, 2011). That policy really took off after the riots in several social housing estates in the 1980s, starting with Les Minguettes in the suburbs of Lyon in 1981. Twenty-five years later, the riots that spread out in many cities throughout France in the fall of 2005 (Lagrange and Oberti 2006) seemed to demonstrate the failure of the policy, and several politicians have argued that it should be abandoned. The reports by ONZUS monitoring the implementation and outcomes tend to show an improvement on some variables, but a degradation on key ones like unemployment, precarization and poverty; youth unemployment is particularly high. However, these data are for the whole of France, and different ZUS seem to experience diverging trends of change, so that the average may be misleading. For the Paris metropolis, various studies by IAU (Institut d’Aménagement et d’Urbanisme) seem to confirm such divergences. Regarding socioeconomic segregation (there are other dimensions for the evaluation of the policy), the main path to decreasing it in relation to ZUS has been the diversification of the type of housing offered after urban renewal of neighborhoods funded by ANRU (Agence Nationale de Rénovation Urbaine). Municipalities have followed quite different policies in that respect (IAU 2010), from those trying to improve social housing but maintain it, to those trying to decrease social housing and substitute it with middle-class housing— this has been the case particularly for some newly elected right-wing municipalities wanting to change the structure of housing in order to change that of the local population, to expand their electoral base and please their supporters. Another area-focused policy which should be mentioned and has close links to the policy of ZUS regards schools. Starting in 1981 with the designation of zones d’éducation prioritaires (ZEP), a new policy moved away from the republican principle of formal equality inherited from the times of Jules Ferry by deciding that in order to combat school difficulties and failures in poor neighborhoods, more resources would be given to schools in selected poor neighborhoods. This policy of territorially-defined positive discrimination has continued since then. Like the ZUS policy, its results are far from having solved the problem, and every new government reinvents the policy under a new name.

172

Residential Segregation in Comparative Perspective

If the ZUS policy has been designed to limit segregation by promoting more social mix in a definite number of poor neighborhoods (751 in 2010), it has been complemented by a more ambitious policy aimed at limiting segregation across urban space generally. The first attempt was the LOV (loi d’orientation pour la ville, 1991), voted by the Left but not implemented because of the 1993 election that brought the Right back into government. The second, under the left-wing government of Jospin in 2000, was the law SRU (solidarité et renouvellement urbain). The basic diagnosis on which it is based is that segregation results mainly from the concentration of poor working class categories in large social housing estates, themselves mainly concentrated in working class municipalities. Such a segregation is considered bad because of the separation between groups and because of the negative neighborhood effects of such concentrations. The solution proposed is to dilute the problem, to maintain social housing but avoid too much concentration in already poor areas, by obliging all municipalities with more than 3500 inhabitants to reach a minimum of 20% social housing in their housing stock. Those not complying have to pay a penalty. Although the Right opposed the law fiercely in Parliament, subsequent right-wing governments have preferred not to cancel it, thus consolidating apparently a larger non-partisan consensus against segregation. But in fact the law has not been enacted very energetically—many upper class municipalities have said from the beginning that they would prefer to pay the penalty than build social housing; and the préfets, representatives of the central government in charge of enforcing the law, have not been very zealous in doing so. In the Paris metropolis, 83 municipalities out of 181 concerned had not made the effort they were supposed to during the period 2005–2007, and were threatened with having to pay the penalty. Furthermore, as we have already mentioned, some right-wing municipalities manage the allocation of social housing in such ways as to favor political clients who are not necessarily working class, and to avoid undesirable tenants—immigrants, large poor families, etc. The municipality of Paris under Chirac and his successor Tibéri was famous for providing quite a lot of its well-located public housing apartments to upper-class residents; and the municipality of Neuilly has very few public housing apartments but provides them to very select tenants only! (Pinçon and Pinçon-Charlot 2008). During the first years of the Sarkozy presidency, beginning in 2007, the law was in many ways emptied of its efficiency because the government followed an explicit neoliberal agenda of cutting back social housing and promoting access to homeownership. Such a policy had to be revised after the disaster of the subprime mortgage market in the US. The model was a failure. The level reached by real estate prices, which did not drop very much in France and even less in Paris, on the one hand, and the squeeze on household resources due to the economic crisis, on the other, have brought the issue of difficult access to housing back onto the political agenda, and social housing has been reconsidered even by the Right as a necessary solution—but funding is being squeezed in the national budget as part of the national austerity policy. Left-wing municipalities try to build more social housing, but they depend on the resources allocated by the central government and on their ability to find and buy land. This last point is particularly crucial in the central parts of the metropolis, where land prices have gone up, and where most of the territory is already built up. The left-wing municipality of Paris has recognized that in order to maintain a working class population in the city center more social housing has to be built, but it has difficulty finding land for it, and faces fierce opposition from upper class residents when it wants to build social housing in upperstatus areas like the XVIth arrondissement. In spite of its efforts, only a few hundred additional social housing units were built during 1999–2007, and the share of social housing fell slightly (16.7 percent to 16.3 percent) due to the faster progression of homeownership. French anti-segregation policies are ambitious in their principles, but their implementation is inconsistent, unable to maintain a steady effort over the years in the use of the ZUS and ZEP

Segregation, Social Mix and Public Policies in Paris

173

policies, or unable to confront the resistance of upper and upper-middle class municipalities in the case of the SRU law. They have a structural weakness which is that they address what is more a symptom than a cause, and do not address the main cause which is the hierarchization of urban social space from the top by the self-segregation of the upper class and by the speculative investments of luxury office building. They are also counterbalanced by other policies that have a de facto segregative effect through the urban and social inequalities they produce. Researchers have shown, for example, that although the ZEP policy supposedly gives a bonus to secondary schools in deprived areas, when one takes into consideration all the public funding devoted to schools, those in upper status areas receive substantially more resources without any bonus than those in ZEP, due to higher salaries of more experienced professors, additional resources for various kinds of options for languages, music, arts, etc. (Davezies and Tréguer 1996). When a highway like A86 is built through the upper status areas west of Paris, it is planned from the start as a tunnel for most of the way to protect the green spaces, the landscape and avoid noise and pollution, whereas for the same A86 going through working class municipalities north-east of Paris, like in Bobigny, it took many years of local social movements and protest by the municipalities to obtain similar protection in some spots only where the negative impacts were particularly strong. Does this mean that those policies have no effect at all? Probably not, probably things would be much worse if they had not existed. Trying to explain the reasons why the 2005 riots took place in some neighborhoods and not in others of similar social profile, Lagrange and Oberti (2006) argue that the impact of policies had a role, particularly when they were able to maintain a steady support for local NGOs contributing to maintaining social ties. If market forces ruled the production of the city without regulation, inequality and segregation would be much stronger. The continued existence of quite a high degree of social mix in parts of the city is a combination of the complex effects of the history of the city, and of present day policies in which local policies of left-wing authorities are a major component, producing day by day old and new forms of social mix through land use regulation, provision of services and control of urban development projects. Conclusion Paris is definitely a segregated city, but its segregation is relative, not absolute; segregated groups are more or less mixed residentially, not totally separated. To characterize segregation in Paris, we need to move away from the dramatic images of the ghetto, the dual or divided city, in which areas or neighborhoods are totally dominated by one group or the other, and consider the complexity of the urban socio-spatial structure. Such a complexity calls for the simultaneous use of different methods to characterize it, measuring the intensity of segregation along different dimensions. Although the traditional debate, in politics as well as in social sciences, has been about socioeconomic segregation, and although socioeconomic segregation is the major dimension of differences between neighborhoods in the Paris metropolis, the segregation of immigrants from Maghreb or Sub-Saharan Africa exceeds that and reveals the effects of underestimated ethno-racial discrimination processes. But it is still a moderate level of ethno-racial segregation compared to US cities like New York or Chicago. Though it has been increasing a little, it has nothing to do with the media image of all immigrants living in ghettos, and it is probably lower today than it was in the late 1960s, when many immigrants from Algeria lived in slums (bidonvilles). Characterizing the complexity of segregation means characterizing the complexity of the spatial structure and the distribution of social groups. This is important in order to assess one type of effect of segregation, which is urban inequalities. Analyzing the consequences of segregation would have

174

Residential Segregation in Comparative Perspective

exceeded the limits of this chapter, and we can only mention here that the characterization of urban inequalities related to segregation is also a complex task. The banlieues of Paris are socially quite diverse, and they are also very diverse in terms of transport facilities, public facilities and services, quality of the environment, etc. The second major type of effect of segregation that has to be discussed is its consequences in terms of relations between social groups, from the level of the local neighborhood to that of the urban society, the metropolitan system as a whole. So far, local social relations have only been studied for a limited number of cases, like old working class areas, upper class areas, gentrifying neighborhoods, ZUS/quartiers en difficultés, but the different categories of middle-mixed areas, although they are the most frequent modality in our cluster analysis, are sociologically terra incognita. Sociologists have to resist the media taste for the extreme dramatization—dual city, ghetto, make striking titles, unlike relative segregation and complexity of relatively mixed spaces—and the fascination for extreme situations. The magnificent new upper class developments, or the opposite areas of worst concentration of poverty and exclusion, do exist and say something about some features of our societies and cities, but looking at them exclusively one misses most of the picture, of the complex social processes that organize our cities and make them change. Sociologists also have to resist the taste for the explanation by the latest “global emerging trends.” The dominance of finance capital, the growth of cultural industries, of new technologies of information and communication, etc., are important factors that have to be considered. But their impact on the city, on segregation and social structures and relations, cannot be understood without considering the thick accumulation, through decades if not centuries, of physical and social structures, of institutions and cultures, which make the city, which resist to, are changed by, reinterpreted by, interact with, shape and transform in return, those “ emerging global trends.” References Bonvalet, C. and Merlin, P. eds 1988. Les transformations de la famille et de l’habitat. Paris: La Documentation Française. Booth, C. 1902. Life and Labour of the People in London. London: Macmillan. Bourgeois, C. 1996. L’attribution des logements sociaux. Politiques publiques et jeux des acteurs locaux. Paris: L’Harmattan. Brouard, S. and Tiberj, V. 2005. Français comme les autres? Enquête sur les citoyens d’origine maghrébine, africaine et turque. Paris: Les Presses de Sciences Po. Castells, M. 1972. The Urban Question. London: Edward Arnold. Castells, M. 1983. The City and the Grassroots. A Cross-Cultural Theory of Urban Social Movements. London: Edward Arnold. Castells, M., Cherki, E., Godard, F., et al. 1978. Crise du logement et mouvements sociaux urbains. Enquête sur la région parisienne. Paris: Mouton. Chenu, A. and Tabard, N. 1993. “Les transformations socioprofessionnelles du territoire français, 1982–1990.” Population 6, 1735–70. Chombart de Lauwe, P.-H., Antoine, S., Couvreur, L., et al. 1952. Paris et l’agglomération parisienne. Paris: Presses Universitaires de France. Clapier, P. and Tabard, N. 1981. “Transformation de la morphologie sociale des communes et variation des consommations.” Consommation, Revue de Socio-Economie 2, 3–40.

Segregation, Social Mix and Public Policies in Paris

175

Cousin, B. 2008. Cadres d’entreprise et quartiers de refondation à Paris et à Milan. Thèse pour le doctorat en sociologie. Paris: Sciences Po. Davezies, L. and Tréguer, C. 1996. Les politiques publiques favorisent-elles les quartiers pauvres? Le cas de l’Éducation Nationale. Rapport de recherché pour le PIR-Villes CNRS. Desrosières, A. and Thévenot, L. 1988. Les catégories socio-professionnelles. Paris: Editions La Découverte. Donzelot, J. 1999. “La nouvelle question urbaine.” Esprit 258, 87–114. Dubet, F. and Lapeyronnie, D. 1992. Les quartiers d’exil. Paris: Editions du Seuil. Florida, R. 2002. The Rise of the Creative Class. New York: Basic Books. Fourcaut, A. 2000. La banlieue en morceaux. La crise des lotissements défectueux en France dans l’entre-deux guerres. Paris: Créaphis. Freyssenet, M., Regazzola, T. and Retel, J. 1971. Ségrégation spatiale et déplacements sociaux dans l’agglomération parisienne de 1954 à 1968. Paris: Centre de Sociologie Urbaine. Godard, F., Castells, M., Delayre, H., et al. 1973. La rénovation urbaine à Paris: Structure urbaine et logique de classe. Paris: Mouton. Guillon, M. 1992. Étrangers et immigrés en Ile-de-France. Thèse pour le doctorat d’État en géographie. Paris: Université de Paris I. Hamnett, C. 2003. Unequal City. London in the Global Arena. London: Routledge. Héran, F. et al. 2010. Inégalités et discriminations Pour un usage critique et responsable de l’outil statistique. Paris: Comité pour la mesure et l’évaluation de la diversité et des discrilinations. IAU. 2010. Le logement. Un champ d’étude et d’action pour Paris Métropole—1 État des lieux. Paris: Institut d’Aménagement et d’Urbanisme. Kirszbaum, T. 1999. “Les immigrés dans les politiques de l’habitat. Variations locales sur le thème de la diversité.” Sociétés Contemporaines 33–34, 87–110. Kokoreff, M. 2003. La force des quartiers. Paris: Payot. Lagrange, H. and Oberti, M. eds 2006. Émeutes urbaines et protestations. Une singularité française. Paris: Les Presses de Sciences Po. Lapeyronnie, D. 2008. Ghetto urbain. Ségrégation, violence, pauvreté en France aujourd’hui. Paris: Robert Laffont. Le Bras, H. and Chesnais, J.-C. 1976. “Cycle de l’habitat et âge des habitants.” Population 2, 269–98. Lefebvre, H. 1968. Le droit à la ville. Paris: Editions Anthropos. Lepoutre, D. 1997. Cœur de banlieue. Codes, rites et langages. Paris: Odile Jacob. Logan, J. 2001–2004. Director. American Communities Project, presented jointly by the initiative of Spatial structures in the social sciences, Brown University, and the Lewis Mumford Center, State University of New York at Albany: http://www.s4.brown.edu/cen2000/data.html [accessed 20 April 2012]. Lojkine, J. 1972. La politique urbaine dans la région parisienne. 1945–1972. Paris: Mouton. Masclet, O. 2003. La gauche et les cités: Enquête sur un rendez-vous manqué. Paris: La Dispute. Massey, D. and Denton, N. 1988. “The Dimensions of Residential Segregation.” Social Forces 67, 281–315. Massey, D. and Denton, N. 1993. American Apartheid. Segregation and the Making of the Underclass. Cambridge, MA: Harvard University Press. Maurin, É. 2004. Le ghetto français. Enquête sur le séparatisme social. Paris: Editions du Seuil. Nivière, D. 2005. “La mobilité résidentielle des habitants des zones urbaines sensibles entre 1990 et 1999.” In Rapport 2005 de l’Observatoire National des ZUS. Paris: DIV, 123–71.

176

Residential Segregation in Comparative Perspective

Oberti, M. 2007. L’école dans la ville. Ségrégation—mixité—carte scolaire. Paris: Presses de Sciences Po. Oberti, M. and Préteceille, E. 2004. “Les classes moyennes et la ségrégation.” Education et Sociétés 14, 135–53. Oberti, M. and Préteceille, E. 2011. “Cadres supérieurs et professions intermédiaires dans l’espace urbain, entre séparatisme et mixité sous contrôle.” In Cadres, classes moyennes: Vers l’éclatement? eds P. Bouffartigue, C. Gadéa and S. Pochic. Paris: Armand Colin, 202–12. Pan Ké Shon, J.-L. 2009. “Ségrégation ethnique et ségrégation sociale en quartiers sensibles. L’apport des mobilités résidentielles.” Revue Francaise de Sociologie 53, 451–87. Pan Ké Shon, J.-L. and Scodellaro, C. 2011. Discrimination au logement et ségrégation ethnoraciale en France, Document de Travail. Paris: INED. Pinçon, M. and Pinçon-Charlot, M. 1989. Dans les beaux quartiers. Paris: Editions du Seuil. Pinçon, M. and Pinçon-Charlot, M. 1992. Quartiers bourgeois, quartiers d’affaires. Paris: Payot. Pinçon, M. and Pinçon-Charlot, M. 2008. “Qui habite les HLM de Neuilly?” Le Monde 2, 26/01/2008. Pinçon-Charlot, M., Préteceille, E. and Rendu, P. 1986. Ségrégation urbaine. Classes sociales et équipements collectifs en région parisienne. Paris: Editions Anthropos. Préteceille, E. 1991. “From Centralization to Decentralization. Social Restructuring and French Local Government.” In State Restructuring and Local Power. A Comparative Perspective, eds C. Pickvance and E. Préteceille. London: Pinter Publishers, 122–48. Préteceille, E. 2000. Division sociale et services urbains, Vol. I. Inégalités et contrastes sociaux en Ile-de-France. Paris: Cultures et Sociétés Urbaines. Préteceille, E. 2003. La division sociale de l’espace francilien. Typologie socioprofessionnelle 1999 et transformations de l’espace résidentiel 1990–99. Paris: Observatoire Sociologique du Changement FNSP-CNRS. Préteceille, E. 2006. “La ségrégation sociale a-t-elle augmenté? La métropole parisienne entre polarisation et mixité.” Sociétés Contemporaines 62, 69–93. Préteceille, E. 2007. “Is Gentrification a Useful Paradigm to Analyse Social Changes in the Paris Metropolis?”. Environment & Planning A 39, 10–31. Préteceille, E. 2010. “The Fragile Urban Situation of Cultural Producers in Paris. City”. Cultura e Scuola 1, 21–6. Préteceille, E. 2011. “Has Ethno-Racial Segregation Increased in the Greater Paris Metropolitan Area? Revue Française de Sociologie”. An Annual English Selection 52, 31–62. Rhein, C. 1986. “Extension de l’agglomération parisienne et transformation socio-démographique de la Seine-banlieue (1954–1975).” Villes en Parallèle 10, 41–66. Rhein, C. 1994. “La division sociale de l’espace parisien et son evolution.” In La ségrégation dans la ville, eds J. Brun and C. Rhein. Paris: L’Harmattan, 229–57. Sassen, S. 1991. The Global City. New York, London, Tokyo. Princeton: Princeton University Press. Schelling, T. 1978. Micro-Motives and Macro-Behavior. New York: W.W. Norton. Silberman, R., Alba, R. and Fournier, I. 2007. “Segmented Assimilation in France? Discrimination in the Labour Market against the Second Generation.” Ethnic and Racial Studies 30, 1–27. Tabard, N. 1993. “Des quartiers pauvres aux banlieues aisées: une représentation sociale du territoire”. Economie & Statistique 270, 5–22. Tissot, S. 2007. L’État et les quartiers: Genèse d’une catégorie de l’action publique. Paris: Seuil. van Zanten, A. 2009. Choisir son école. Paris: PUF. Wirth, L. 1928. The Ghetto. Chicago: The University of Chicago Press.

Chapter 8

The Solidity of Urban Socio-spatial Structures in Copenhagen Hans Thor Andersen

The Social Geography of Copenhagen It is a striking fact that the social geography of Copenhagen has remained relatively stable despite major social changes through most of the last century despite a huge spatial expansion and a major change in industrial structure. The overall socio-spatial structure, which emerged in the period of industrialization, has been maintained since then; the north-eastern parts of Copenhagen have remained the preferred residential location for the upper class, while the working class has located in the west and south-west. The massive growth of suburban districts from the 1930s did not transform this pattern, but extended it more than 20 kilometers from existing, built-up areas. This pattern has only recently been disturbed as young families and couples of middle class origin tend to remain in the central parts of the city. If this change in the social composition of the inner city population represents gentrification, the contrast in form of deprived neighborhoods represents “ghettoization.” The social geography of today’s Copenhagen can only be understood by a careful investigation of its origin, i.e. in a historical perspective. The term social geography is here used as an expression for the socio-spatial pattern, which appears as the final outcome of a myriad of decisions taken by organizations, authorities, private firms as well as the millions of people who have lived in or near the city during centuries. Among these are social differentiation due to income, employment and education, promotion of various forms of tenure such as renting on market or non-profit conditions, ownership etc., construction and pricing of various forms of transport, household forms, migration and much else. Copenhagen is an urbanized region consisting of 1.9 million inhabitants, who are distributed among 35 municipalities and two regions (from 1 January 2007). The South Swedish city of Malmö (0.3 million inhabitants) is related to Copenhagen in terms of labor market, education, shopping and leisure activities; however, it will not be included in the analysis here. The core of the region, the city of Copenhagen, has half a million inhabitants and a strong concentration of state functions as well as universities, government departments, national organizations and business headquarters, leisure and so on. The suburban areas are first of all dominated by residential functions, although manufacturing, warehousing and retail do play a role. As municipalities are depending on local tax base, the ability to maintain a high-income level and avoid rising social costs, the local governments have strived to develop the preconditions for a balanced economy via taxation, housing and planning policies and service level. Some local governments have promoted detached housing (ownership), others non-profit rental housing in order to serve their citizens. Such local oriented policies have created clear differences in terms of socio-economic conditions from area to area in the region.

178

Residential Segregation in Comparative Perspective

The population of Copenhagen grew predominantly in the central districts up to 1940, when Greater Copenhagen reached 1 million inhabitants. From the 1940s a rapid expansion of housing and manufacturing industry transformed hitherto rural land near the central boroughs into suburban districts guided by the “Fingerplan” since 1947. However, by 1970 Greater Copenhagen stopped growing and both central boroughs as well as the inner suburbs had begun a decline in terms of population; only the outer suburban areas maintained their population. The City of Copenhagen reduced its population from 770,000 in 1950 to 460,000 in late 1980s; a trend, which first changed by the resurgence of the city in the 1990s. The City of Copenhagen has now 530,000 inhabitants and expects to reach nearly 600,000 in 2025. This will strengthen the city versus the rest of the region, although most parts of the metropolitan region now have a growing population. The social urban structure can be traced back to the pre-industrial period; the upper classes to the North around the king’s palaces, to the North West the middle classes and skilled workers and to the West and South West a population composition of newly arrived immigrants, unskilled workers and lower classes. This pattern partly reflected the quality of housing, too. Throughout the following 150 years, this social structure was built into the fabric of the expanding city: During the second half of the nineteenth century, the densely built inner cities were constructed around the historical core and again with a clear pattern of social division. Unfortunately, a good deal of the housing from this period became the slums of the twentieth century. The continuous urbanization included the outer districts of the city in the first half of the twentieth century and the first separation of residential from industrial functions. Increasing wealth allowed the general housing standard to improve over the years. The outcome was low-density residential areas with easy access to green space for recreation and modern infrastructure much in line with the ideas of functionalism. The overall objective became the wish to create a comfortable city with decent living conditions for the ordinary citizens. This intention was largely implemented in the newly developed parts of the city. After World War II, suburbanization took pace fuelled by major infrastructure projects like the commuter train system along the development axes (“fingers,” cf. Map 8.1). The existing social geography was stretched into the new suburbs: the western and south western parts of the urban region became modern, well planned working class and factory districts, while the northern suburbs developed accordingly to the demands of the upper middle classes in the attractive landscape. From the start, Greater Copenhagen was divided according to socio-economic factors such as income, education, living conditions, life expectancy and so on. But also according to political relations— the Social Democratic Party dominated in the center and to the west, while conservatives and liberals dominated in the northern suburbs (Andersen 1991). The 1970s is a period of major changes with strong effects on social relations in the city; the sudden explosion in oil prices triggered a round of economic restructuring, which eventually led to deindustrialization of the region and the collapse of manufacturing industry in central parts of Copenhagen. The result was a drastic rise in unemployment and a long period of stagnation in the city. At the same time, the anti-authoritarian movement peaked and dissolved many social institutions. The breadwinner model began its collapse and led to women’s liberation and to a growing number of divorces and hence to demands for more housing units. The governmental system of the country changed, too. Local governments merged to form stronger and more competent units, which could match the rising expectations from citizens. The period from late 1960s to early 1990s was a particularly harsh period for the City of Copenhagen, which became an island of social problems, unemployment, derelict housing and lack of development. The national as well as the regional policy supported a decentralization of public investments to the rest of the country and the newly urbanized spaces in the periphery of the city. In the meantime large parts of the existing housing stock were out of fashion and suffered

The Solidity of Urban Socio-spatial Structures in Copenhagen

Map 8.1

179

Copenhagen’s urban structure (simplified)

Note: 1: Medieval city; 2: predominantly housing zone from early industrialization (1860–1910); 3: mixed industrial and working class zone; 4: public housing zone; 5: single family zone; 6: post-WW2 suburban zone (“Fingerplan”); 7: new industrial areas. The border of the municipality of Copenhagen runs through the third (and partly the forth) zone, implying that affluent groups have settled outside the jurisdiction of the City. Source: Sketch by Matthiessen, C.W. (University of Copenhagen).

from lack of modern amenities. Several major programs and changes in housing policy, first of all by implementing a scheme for parceling out existing multi-storey buildings into owner occupied flats, changed the preconditions for social composition of the population. The national government decided in the late 1980s to relocate the navy from its historical location in the center of Copenhagen harbour; this meant a loss of many thousands of jobs and a fear that yet another institution would be led to continuous decline. The parliament then had a principle debate about the situation in Copenhagen (Andersen 2008) on 20 March 1990, where both the government and the leading opposition party committed themselves to formulate and implement a round of new initiatives to strengthen the city: a new city annex (Ørestad) connected to the existing City by a new metro line (extensions are still going on), a motorway and rail line

180

Residential Segregation in Comparative Perspective

connecting Copenhagen central with Malmø through a new, 20 km long, bridge and tunnel, linking Copenhagen airport as well. The airport, which also was refurbished and expanded, was later connected to the city via a second metro line in 2007. Finally, the government invested in cultural institutions, research and higher education. The outcome was a sharp change from stagnation to boom town and large parts of inner harbour spaces, which were unused and in disrepair in the 1980s, suddenly developed into expensive housing and prestigious business domiciles (Desfor and Jørgensen 2004). Moreover, a new regime was formed at the Town Hall; a regime, which began to implement a strategic plan for the City of Copenhagen where housing had a major position (Lund Hansen et al. 2001). The overall social geography of Copenhagen is closely related to the historical development of the city and to housing characteristics such as tenure, quality, relative location and price. These characteristics produced a social pattern, which is quite similar to the expectations from the Chicago school (Burgess 1925): old two room flats out of fashion in the inner cities and modern, family sized housing in green surroundings in suburban districts. As much of the office employment, together with higher education, culture, commercial amusement and so on, was concentrated in the City, the classic life cycle pattern with families with children in suburbs and singles and youngsters in central parts of the city prevailed to the late 1990s. The overall segregation level in Copenhagen has for a long period, i.e. at least since the 1960s, been in decline: First of all due to a more equal distribution of age and household composition at local level since the peak of suburbanization (cf. Andersen et. al 2000, Andersen 2005). Overall Segregation Pattern in Metropolitan Copenhagen Segregation is a complicated result, which combines numerous individual and general social processes; these processes manifest themselves, among other ways, in residential patterns. It is highly questionable whether residential location is the most important connection between individuals and space; access to education, amusement and work may be more relevant for modern people in particular if interpersonal relations and influences, such as socialization and local communities, are in question. However, residential segregation has been the traditional focus for such studies for a long time and moreover, the available empirical data is tied to people’s addresses. Finally, there has always been a strong belief in the importance of space and place in relation to human behavior. Thus, this chapter will continue in the tradition and look at the spatial distribution of socio-economic factors: employment, income, education, age groups and ethnic minorities. These factors, which in themselves are important facts of living conditions, do of course also serve as highly correlated indicators of a broader sense of living quality. Map 8.2 shows the relative geographical concentration of high earning individuals; the use of location quotients provides a simple, but easily comprehensible illustration: A location quotient at 1.0 is the average for Metropolitan Copenhagen; thus a quotient less than 1 indicates under representation of the variable and quotients above 1 over representation. The high earning individuals are those who belong to the two top income deciles; they are concentrated first of all to the north and north-west of the city center as well as the historical core of the city. In contrast, they are quite absent in the suburbs west of the City of Copenhagen as well as the western part of the city itself. The income distribution in Denmark is one of the most equal in Europe with a Gini-index of dissimilarity at 0.26 (Det Økonomiske Råd 2008); yet, incomes seems to be a most import factor in terms of segregating the population: Again, the classic pattern proposed by Hoyt (1939) fits quite

The Solidity of Urban Socio-spatial Structures in Copenhagen

Map 8.2

Location of high-income persons in Metropolitan Copenhagen (9th and 10th deciles) 2007, location quotient

Source: Samson, J. (2009).

181

182

Residential Segregation in Comparative Perspective

well to Copenhagen (Andersen, Andersen and Ærø 2000). This segregation pattern can easily be described as that of the classic industrial or modern city; a class based social differentiation has shaped an urban social landscape of more or less attractive localities, where the upper classes occupy the best located districts with beautiful surroundings and few disturbing functions like factories or heavy transport. The working classes find their residence next to large manufacturing plants, harbour industries, transport nodes and other concentrations of employment. Decent dwellings are important although the improvement of housing standard is a slow process. The middle classes try to imitate the upper classes, although they cannot afford the expensive location and have to accept second-class districts; in Copenhagen this is often to the North West around some big lakes, between forests and open fields. The life cycle pattern is embedded in this socioeconomic landscape; the younger families settle in the outskirts of the built up area unless they are particularly affluent, while youngsters search for central location as they leave home and look for education, job or fun in the central city. However, there are a few signs that this segregation pattern may be shifting. The spatial distribution of income reflects largely basic social elements such as employment and education, cf. Map 8.2: The traditional upper class districts along the coast north of the City of Copenhagen are marked by many high income earners just as in the historical center. However, the poorest districts are also found in the inner city, often neighboring quite affluent ones. The districts to the south west have a dominance of low income groups; many of these are immigrants and immigrant descendants. The spatial distribution of income is similar to education, labor market participation and housing tenure; the overall segregation pattern reflects the class structure of the industrial city, cf. also Figure 8.3, which reveals the occurrence of unemployment in 2007. The overall unemployment level for Metropolitan Copenhagen was 4.2 percent, but several times higher in parts of the central city and western suburbs. In contrast, unemployment was a rare phenomenon in the northern suburbs, in general less than one percent. Unemployment is nearly synonymous to reception of social benefits in terms of spatial distribution in the city; it is first of all located in parts of the inner cities—the northern and southern parts as well as western suburbs: All spaces, which have been localities for manufacturing industry and working class housing in the past and have suffered from industrial change. The unemployment, reception of social benefits and lack of education above minimum required level are features clearly interrelated; the co-location of these features are similarly high although other relations may influence residential choices. As a contrast, the location of the highly educated reflects quite a different situation: Strongly concentrated to the inner cities and northern suburbs and the university town “Roskilde” (30 km west of Copenhagen), they are nearly absent in western and south western suburbs. However, they are overrepresented in most parts of the inner city, following a residence established during their studies or attracted by multi-ethnic neighborhoods (Maps 8.4 and 8.5). However, the overall segregation level for the highly educated is at the same level as immigrants and their descendants. All together, the non-Danish population makes up 22 percent of the City’s population; however, half of the immigrants have a Nordic, a West-European or North American origin (cf. next pages). This is surprising since the highly educated in absolute numbers have increased fast during the last few decades, which in turn must force some to live in different districts from those they traditionally settled. Ethnic minorities have become a key issue in urban politics and in the wider public debate; the number of immigrants and descendants has grown from 150,000 to more than 450,000 from 1980 to 2007 in Denmark. Despite their tripled numbers, they remain relatively segregated; in 1986 immigrants had an index of segregation at 0.39 that 20 years later (2007) remained practically

The Solidity of Urban Socio-spatial Structures in Copenhagen

Map 8.3

Relative distribution of unemployed persons in Metropolitan Copenhagen, 2007, location quotient

Source: Samson, J. (2009).

183

Residential Segregation in Comparative Perspective

184

Map 8.4

Spatial distribution of persons with academic degrees (master level) in Metropolitan Copenhagen, 2007, location quotient

Source: Samson, J. (2009).

The Solidity of Urban Socio-spatial Structures in Copenhagen

Map 8.5

185

Distribution of ethnic minorities in Metropolitan Copenhagen, location quotient

Source: Samson, J. (2009).

186

Residential Segregation in Comparative Perspective

unchanged (0,38); this indicates a relatively concentrated settlement to particular areas, where minorities make up a substantial share of the population. The ethnic minorities and their descendants include about one hundred different nationalities; the Europeans make up 40 percent of non-Danish citizens in Copenhagen. The non-Europeans may be subdivided to three groups in terms of the cause of their arrival to Copenhagen. The Turks, Pakistanis and Moroccans and to some degree the Yugoslavs arrived as migrant labor. The second wave of immigrants came during the last three decades as war refugees. Iraqis, Iranians, Afghanis and Lebanese. The third wave of immigrants originates from countries such as the Philippines and Thailand; they function as domestic help in particularly wealthy areas. The characteristic location pattern for ethnic minorities consists of a few districts to the north of central Copenhagen, a number of districts to the west of the city and selected neighborhoods in inner city Copenhagen. The pattern has remained similar for the last three decades and reflects to a large degree the dominant tenure form. The affluent parts of the region, i.e. north of the City along the coast, have very few inhabitants of non-Danish origin, although the number of domestic service workers is rising. The only exception is the neighborhood “Nivå” about 30 km north of the City, where the middle sized neighborhood “Islandhøjparken” disturbs the overall picture. To the west of the City, several suburbs contain quite high shares of immigrants and their descendants; one municipality (Ishøj) has about 30 per cent non-Danish citizens. In some neighborhoods, often non-profit housing estates such as “Vejlegårdsparken,” “Brøndby Strand,” “Tåstrupgård” in the suburbs and inner city estates such as “Lundtoftegade,” “Akacieparken” and “Mjølnerparken,” the share is much higher—more than 90 percent. Causes and Consequences: Segregation in Copenhagen Although segregation and its possible effects on general social behavior have been studied intensively over the last century, it is still difficult to point out exactly the effect of segregation. There exist many assumptions both among the public and academics, but little evidence for the assumed effects. One reason for this disappointing situation is the difficulty to isolate the various causes in relation to people and their daily life; rather few people will remain in their residential area for long periods and attitudes, norms, ideology, behavior can thus be influenced by many other localities such as work, education, shopping areas, leisure facilities and so on. So, where does the influence come from? Electronic media may be much more influential than any impact from neighbors or workmates. A second reason is the lack of specific knowledge on how attitudes and norms are formed and transferred between people. Is, for example, a regular contact necessary? Or is it primary relations, which are decisive for forming norms and meanings? In this context, the decisive factors are considered to be, first of all, access to work and decent income linked to occupation and educational level. Second, the welfare state system (Esping-Andersen 1990) has a major role for the equalization of life chances as health service, social transfers and care for elderly, children etc. Third, the housing markets, categories of tenure, privileges and duties as well as accessibility for young people without means play an important role. Fourth, government strategies and priorities; in particular local governments have been able to impact on local development considerably (Andersen 1991) by favoring selected forms of tenure in order to prevent inflow of marginal groups.

The Solidity of Urban Socio-spatial Structures in Copenhagen

187

Labor Market Denmark is characterized by a high activity rate, which includes all age groups; the Lisbon target of labor market participation rate at 70 percent for males and 60 percent for women was reached before the formulation of the targets. Combined with a labor market, which implies relatively high minimum wages, this means that poverty is a rare phenomenon for employed people. Special programs have managed to include younger people (below 25 years) into the labor market or provide educational training for them; as a result Denmark has one of the lowest unemployment rates for young people together with countries like Bulgaria, The Czech Republic, Austria, The Netherlands and Germany (about 10 percent). Poland, Sweden, Hungary and Portugal have a youth unemployment rate around 20 percent, while Spain reaches 25 percent (2007). For adults (over 25 years) the same variation between countries exists although at a lower level: Spain and Slovakia are at the top with 10 percent unemployment compared to Denmark and The Netherlands with 2 percent. Among the unemployed, only 1 percent was recorded as long term unemployed in Denmark against 8–10 percent in Greece, Lithuania, Poland and Slovakia. Especially women are hit by long term unemployment. The high labor market activity rate has ensured that almost all adults have an income; furthermore, the labor market agreements have produced a relatively high minimum salary. The latter prevents workers and others with reduced capacity on the labor market to get employed. The high activity rate, on the other hand, has a major, but less visible effect; households with double income can afford substantially more expensive housing than singles. Moreover, the geographic distribution of housing suitable to families is quite dissimilar to that of smaller and less expensive dwellings: High activity rate is mostly found in the suburbs, while unemployment of people with reduced capacity is located, first of all, in the inner city. The successful inclusion of young people and marginal groups into education and employment (known as “active labor market policy”) substantially reduces poverty. In turn, this allows vulnerable groups to get access to decent housing, including homeownership. In particular, people with immigrant background have benefited from this. Employment in Metropolitan Copenhagen has grown since the end of the recession in the mid 1990s from around 860,000 to 916,000 jobs in 2006, two years before the current recession. The general process of industrial change has only induced relatively small changes in the overall industrial structure calculated by employment categories; deindustrialization is an ongoing process and the city is still losing jobs in manufacturing industry. The public sector, which has been a main employment sector for decades, has only increased employment at minor scale during the last three decades. Instead, private services have expanded and now employ every third employed person against 25 percent in the mid 1980s. And, in particular, business services, banking, insurance and real estate management have grown by around 60 percent from 1984 to 2008. Occupation categories have been relatively stable as well for the period 1981 to 2009; due to a major change from occupational category based on education to a new system based on employment position, data from before 1997 cannot be compared with data after 1997. They simply measure different phenomena; yet, it is possible to identify some important changes: First of all, the number of workers is decreasing, including both skilled and unskilled workers. From 1981 to 1996 there is a steady decline in the number of workers and since 1997 a reduction in the share of employees at basic level, too. This alteration reflects the general decline of manufacturing industry and does not surprise. The second change is the growth of employees at upper and medium level after 1997, which may reflect the parallel transformation of industrial sectors. Although there are good reasons to assume that a similar transformation took place before 1997, the statistical categories cannot

188

Figure 8.1

Residential Segregation in Comparative Perspective

Occupational categories in jobs located in City of Copenhagen: a) 1981–1996 and b) 1997–2009

Note: Notice that the two periods are not comparable due to changes in categorization of employment. Source: Statistikbanken, Danmarks Statistik (www.statistikbanken.dk), author’s calculations.

reveal it. It is possible that categories like “employees, not specified” included the rising number of upper and medium positioned employees. Again, there are only signs of a moderate change in terms of occupational categories for the City of Copenhagen. The trend points to a general upgrading of employment categories and a reduction of basic level jobs.

The Solidity of Urban Socio-spatial Structures in Copenhagen

Table 8.1

189

Occupation categories, index of segregation 2009

Self employed

Top managers

Employees: upper level

Employees: medium level

Employees: basic level

Other employees

8.5

15.4

14.3

3.3

5.8

9.0

Note: Calculated at municipal level. Source: Statistikbanken.

Segregation indices for the main occupational categories are very low, but this is most likely the consequence of using data on municipal level (they vary in size from 12.000 to 530.000 inhabitants) even though we should not expect spectacularly different index values if the calculation is performed on a more disaggregated level. However, it is worth to notice that top managers and upper level employees, who reside in the northern part of Copenhagen, present the highest segregation levels (assisting spouses is a quite small category and should not be included). This pattern is confirmed by indices of dissimilarity between the main occupational categories (Table 8.2); the top managers clearly have the most different residential pattern compared to other occupational groups followed by upper level employees. Table 8.2

Self employed Top managers Employees: upper level Employees: medium level Employees: basic level Other employees

Occupation categories, index of dissimilarity 2009 Self employed

Top managers

Employees: Employees: Employees: upper level medium level basic level

Other employees

*

0.24

0.30

0.18

0.26

0.32

*

0.36

0.27

0.41

0.48

*

0.31

0.40

0.44

*

0.22

0.09

* *

Note: Calculated at municipal level. Source: Statistikbanken.

Welfare Arrangements The special type of labor market system existing in Scandinavia will of course only provide decent conditions for those working, not for children, elderly, unemployed or long term ill people. This is a task for the welfare state and the Danish welfare state has got universal scope and in practice, the welfare state has been able to include the vast majority of people unable to support themselves. The

Residential Segregation in Comparative Perspective

190

Figure 8.2

European countries at risk of poverty rate before and after social transfers, 2007

Source: Eurostat (2010).

combined effect has been the ability of preventing poverty to become a serious risk for individuals. Calculated as income level below 50 percent of the median income and excluding students and so on, the poverty level is about 3 percent while long term poverty has been kept as low as below 1 percent of the population. Including students and migrants, Denmark has a poverty level of approximately 6 percent of the population below poverty limit, while the USA and the UK have 17 percent and 12 percent respectively. This in itself does not guarantee an equal distribution of social groups within the city, but it does ensure access to decent housing for most individuals. A major explanation of this low poverty level is the income redistribution and general welfare schemes, which aim at reducing substantially the “at risk of poverty rate”: While social benefits and other transfers hardly reduce poverty risk in southern Europe or Bulgaria, the welfare systems in Scandinavia have a marked influence on the incidence of poverty: Without social transfers poverty would be at a clearly higher level in Denmark, Norway, Sweden than in Greece. Similar are the poverty rates for children growing up with one parent: about 10 percent in Denmark and Sweden contrasting with 50 percent for the USA, above 30 percent for the UK and Spain and close to 50 percent for Ireland (Eurostat 2010). These figures are only indicators of an important precondition for segregation; the income variation is relatively limited (Gini-coefficient is 0.26) and traditionally vulnerable groups are covered by various welfare programs. This has, to a large degree, reduced poverty to a minor problem and had an important impact on segregation. The most important effect of the Scandinavian welfare state is the inclusion of the weakest members of society by ensuring a reasonable living standard to all. However, as it will be presented in the next section, the provision of decent, non-profit housing to guarantee an acceptable standard of living involves a problematic spatial dimension: that the non-profit housing tends to concentrate in particular locations and thus promote segregation.

The Solidity of Urban Socio-spatial Structures in Copenhagen

191

Housing Market Housing has a key position in relation to segregation; it is the medium through which numerous social processes sort out more or less marked concentrations of various groups. The geography of housing relates first of all to quality and tenure; obsolete dwellings lacking modern facilities have been reduced considerably in numbers thanks to costly renewal programs and private investments. Often such efforts trigger a series of social transformations as the increased housing standard causes rents to raise and a different type of residents populate the neighborhoods. The older parts of the city do still contain a large share of private renting, which together with cooperative housing recruit a different type of inhabitants than the suburban districts with their high shares of detached ownership (to the north) and non-profit housing to the west. Of particular importance was the calculated attempt to change the housing tenure in the City of Copenhagen in favor of home ownership, which would induce a lasting improvement of the fiscal balance for the City budget. The general expansion of inner city Copenhagen and redevelopment of inner harbour spaces can be seen in relation to this strategy for becoming economically sustainable. A case in point is the inner city district of Vesterbro; for more than a century, the Vesterbro district received thousands of immigrants annually searching for jobs and housing. As the poorest and most dense district Vesterbro became a low status district, comprising a large part of the city’s marginal population. A huge improvement program from the early 1990s has turned the district into the most attractive and expensive area of the city—but also exchanged the hitherto population with yuppies and middle class families. This process was further fuelled by transforming renting into cooperative housing or homeownership (Lund Hansen and Gutzon Larsen 2008). Housing tenure has become a factor of major importance behind the observed socio-spatial changes in the city; while the overall housing market has been of almost identical size during the last few decades, the form of tenure has changed considerably (Andersen 2002). Several changes have induced deep changes in the housing stock of the city; due to its fiscal collapse in the early 1990s, the city was forced to sell its rented housing stock. Most of it was converted into cooperative housing, some though into ownership flats. These changes led to the drastic reduction of publicly owned dwellings. A similar conversion has been going on in the privately owned stock of rented dwellings: since 1981, where this type of tenure was the dominant with 40 percent of all dwellings, its share has fallen to 14 percent (Figure 8.3). The combined effect of still fewer rented dwellings –public or private owned– and a rising share of condominiums and cooperative housing has had direct effect on socio-economic relations among residents. Rapidly rising prices from the mid 1990s to the property crisis in 2008 have efficiently prevented other than middle and upper middle class people from settling in the inner city. As a result average income has started to grow at a similar rate as the rest of the metropolitan region; the inner cities are not lagging behind as previously. The transformation of tenure for the housing stock in Metropolitan Copenhagen has reduced income segregation, but at the same time made access to housing much more difficult for young people, singles and others. Moreover, it has sharpened a polarization between homeowners and renters. In particular, residents in non-profit housing (‘social housing’) are marked by low income, labor market inactivity and sometime with health problems. Figure 8.4 shows how the share of low incomers is rising despite the fact that the majority of the residents belong to this group. This is a clear indication of the continued marginalization of the inhabitants of these estates; a growing share of the tenants is marginalized, i.e. characterized by early retirement, long term unemployment, low income, divorce, health problems and so on. The opposite trend characterizes the income profile of

192

Residential Segregation in Comparative Perspective

detached housing owners (Figure 8.5). These processes have become a very important issue both in local and national politics and prove that segregation is far from just an academic discussion.

Figure 8.3

Type of ownership in multi-storey housing, City of Copenhagen, 1981–2010 (percent)

Source: Statistikbanken, Danmarks Statistik (www.statistikbanken.dk), author’s calculations.

Figure 8.4

Number of residents in social housing sector in Copenhagen distributed by income deciles, 1986–2001

Source: Statistikbanken, Danmarks Statistik (www.statistikbanken.dk), author’s calculations.

The Solidity of Urban Socio-spatial Structures in Copenhagen

Figure 8.5

193

Number of residents in detached ownership housing in Copenhagen distributed at income deciles, 1986–2001

Source: Statistikbanken, Danmarks Statistik (www.statistikbanken.dk), author’s calculations.

During the period considered (1986 to 2007), the overall change at the housing market is a continuous social decline of the non-profit sector: From being a preferred and desired form of tenure for working class and lower middle class, it has developed into the residence for marginalized people: Early retired, long term unemployed, long term unhealthy people and immigrants from the third world. This change among the tenants has in itself pushed the middle class groups out of non-profit housing and the term “ghetto” is now applied to a considerable number of such estates, primarily suburban. Finally, the local government’s housing policies have obviously had a substantial impact on segregation between, as well as within, the municipalities; while some municipalities include as much as 40–60 percent non-profit housing and similarly low shares of ownership housing, the opposite is the case in other municipalities. This difference in itself produces important differences between municipalities, which in turn are embedded in places with different historical background and symbolic values. Recent conversions of the rental stock into cooperative or ownership housing do further add to the processes of segregation in Copenhagen and has lead so far to the emergence of gentrified districts close to deprived neighborhoods. New lifestyles and trends have made urban life more attractive for young couples during the last two decades and thus have changed traditional life cycle movements. Whether this will last is a highly relevant issue for the City of Copenhagen and for the way segregation will evolve in the wider metropolitan area. Conclusion European cities differ to some degree from their American counterparts; the inner cities are lively and have been able to develop into attractive neighborhoods even for the upper middle classes. A main reason for this is most likely the massive public efforts to maintain the historical cores as

194

Residential Segregation in Comparative Perspective

acceptable residential districts for middle class families. Thus, massive investments have been induced over the last half century to ensure the resurgence of old, big European cities. The policy has been successful as most city centers have been refurbished and thus been able to maintain their centrality for a number of functions: commercial activity, politics, culture, higher education, leisure and recreation as well as residence for the better-off. A key to this success is the property market, i.e. the investments into the built environment, and the stabilization effect it has on social development and change. Both the physical structures such as buildings, infrastructure, urban facilities and the institutional setting enable and prevent at the same time specific situations. Homeownership in attractive environments excludes a large group of people with more moderate incomes living in rented housing; moreover, these environments are spatially separated and the city is thus soon broken up in various social-tenure segments at different locations. As the market immediately adds a price structure to different tenures and locations, supply and demand reflects the social differentiation of society in the pricing of various neighborhoods and forms of tenure. Segregation in this chapter has been analyzed at three different, but equally important, social arenas: Housing, employment and welfare. Housing tenure and standard are lasting qualities of the built environment as it is difficult, or at least costly, to change and almost impossible to move. Once constructed, the built environment tends to remain in place for long periods and function as a media of sorting different social groups in the urban landscape. Working class and labor market inactive groups have tended to concentrate in western parts of Copenhagen, often within the nonprofit sector. In contrast, the upper middle class families have located to the north of the city center in primarily ownership housing. This has always produced a marked difference in living standards, income, labor market participation, health, and education and so on between the north and the west of Copenhagen. The housing tenure pattern has furthermore been reflected into a clear separation of different occupations, education levels and income categories; as these three factors are strongly interconnected, the neighborhoods with high incomes do also have a high share of top managers and highly educated people, as observed in the northern suburbs of Copenhagen. In contrast, many rented housing estates west and south west of the city contain a high share of unemployed, low income, poorly educated people with difficult access to the labor market. Moreover, these districts show a high dependence on public benefits and other social programs. The welfare system has obviously a most important function for immigrants, early retired, singles and other vulnerable groups; however, they are often found in rented housing rather than home ownership. Although the welfare system is identical across Copenhagen, the underlying differences on which it is imposed produce dissimilar outcomes; some districts suffer from the concentration of too many elderly without additional means than their basic pension; other districts have a high share of households with private pension schemes. The former districts do also have a larger share of youngsters without education beyond basic schooling; despite the fact that secondary school is tuition free, there is a clear class difference. In upper middle class districts about 95 percent of all children get a secondary school degree, while only a third of the working class children manage to do the same. Inequality in terms of recruiting to university education is even higher. Segregation in Copenhagen has followed a classic Chicago School model with clear socioeconomic sectors mixed with life-cycle shaped concentric circles around the central business center. While this socio-spatial order has remained rather stable for decades, minor changes have appeared: First of all, immigration has introduced a growing number of non-Danish citizens, who due to lack of integration at the labor and housing markets form concentrations both in the inner city as well as in suburbia. At the same time, parts of the inner city have passed through public

The Solidity of Urban Socio-spatial Structures in Copenhagen

195

led refurbishment, that has opened up its neighborhoods for conversions into different forms of ownership and thus induced a change in their social composition (“gentrification”). A key element in social-democratic policy has been ample investments in education in order to raise social mobility and general social equality. A main principle is, and has always been, provision of high quality education free of charge. In order to counteract negative social inheritance a generous system of benefits for students and young people was introduced in the 1960s in order to avoid that lack of income prevents anyone from getting an education. It worked in the 1960s and 1970s, but since then the effects have been minimal in terms of social mobility. The Danish school system, despite being of equal standard across the city, has a relatively low ability to provide social mobility compared with other Nordic and West European countries. It seems that parents’ own education and social position is more important for a child’s opportunities than the educational system—despite being provided free at all levels. This may be rooted in traditions, socialization and impact from relatives, friends and neighbors together with the settings of educational institutions. All of this produces significant barriers to individuals unfamiliar with middle class values and norms. However, the general social geography of Copenhagen and many other north European cities is by and large constant. There appears to be a remarkable stability considering the above mentioned major social changes during the last few decades and the fact that the then existing conditions— social, economic and political—have been replaced by others long ago. Moreover, many of the people who lived and maintained these districts have now disappeared. Yet, the very same districts, which 60 years ago or more were considered to be among the most attractive, have kept that position and the same applies to the less desirable districts. Although the overall pattern of segregation in Copenhagen fits quite well with the “classic” pattern described by the Chicago School, two important modifications, at least, can be identified during the last two decades: the central boroughs have started to grow in population and this has to some degree broken the traditional trajectory linked to life cycle in Copenhagen. Young families seem to remain in the city after finishing education instead of returning to suburbia like their parents before them. The second major change is the rise of clearly distinct ethnic neighborhoods; some of these have been labeled ghettoes in the media. The outcome is a more complicated social geography than previously: the traditional class segregation model has been supplemented with “ethnic ghettoes” both in the inner cities and suburban edges and with wealthy enclaves in former working class districts in the inner city. It is noteworthy that income inequality, despite the unchallenged social-democratic welfare regime, has not been substantially reduced; this reflects an interesting bifurcation of the Danish society. On the one hand, labor market and businesses have been regulated only at an absolute minimum degree in order to benefit from free market potential and, on the other hand, government intervention has modified, but never challenged neither market institutions nor their economic consequences. To sum up, income distribution in Denmark (Figure 8.2) does not belong to the most equal in Europe: The low Gini-index is primarily an effect of government regulation. A second issue worth noticing is the paradox impact of welfarism on the social structure of cities; the most generous welfare systems have at the same time supported marginal groups in society as well as institutionalized segregation in the form of modern housing units organized as non-profit tenure located as isolated enclaves. Thus, the institutions of the welfare state managed to relieve the harsh consequences of market forces, but did this via the construction of new and separate housing estates isolated from other parts of the city and organized with different tenure. This has caused severe social, economic and political challenges since the 1980s.

196

Residential Segregation in Comparative Perspective

References Andersen, H.S., Andersen, H.T. and Ærø, T. 2000. “Social Polarisation in a Segmented Housing Market: Social Segregation in Greater Copenhagen.” Geografisk Tidskrift 100, 71–83. Andersen, H.T. 1991. “The Political Urbanization. Fringe Development in Copenhagen.” Espace, Populations, Sociétés 1991–2, 367–79. Andersen, H.T. 2002. “Globalisation, Spatial Polarization and the Housing Market.” Geografisk Tidskrift 102, 93–102. Andersen, H.T. 2005. Storbyens ændrede socialgeografi. Storkøbenhavn i et nordvesteuropæisk perspektiv. Kulturgeografiske skrifter. København: Reitzel. Andersen, H.T. 2008. “The Emerging Danish Government Reform—Centralised Decentralisation.” Urban Research & Practice 1(1), 3–17. Burgess, E.W. 1925 [1961]. “The Growth of the City: An Introduction to a Research Project.” In Studies in Human Ecology, ed. G.A. Theodorson. New York: Harper & Row, 37–45. Desfor, G. and Jørgensen, J. 2004. “Flexible Urban Governance. The Case of Copenhagens’s Recent Waterfront Development.” European Planning Studies 12(4), 479–96. Det Økonomiske Råd 2008. Dansk Økonomi Efterår 2008. KonjunkturvurderingPrincipper i skattepolitikken. København. Esping-Andersen, G. 1990. The Three Worlds of Welfare Capitalism. Cambridge: Polity Press. Eurostat 2010. Combating Poverty and Social Exclusion. A Statistical Portrait of the European Union. http://epp.eurostat.ec.europa.eu/cache/ITY_OFFPUB/KS-EP-09-001/EN/KS-EP-09001-EN.PDF. Hoyt, H. 1939. “The Pattern of Movement of Residential Rental Neigbourhoods.” In Studies in Human Ecology, ed. G.A. Theodorson. New York: Harper & Row, 499–509. Lund Hansen, A., Andersen, H.T. and Clark, E. 2001. “Creative Copenhagen: Globalization, Urban Governance and Social Change.” European Planning Studies 9(7), 851–69. Lund Hansen, A. and Gutzon Larsen, H. 2008. “Gentrification-gentle or Traumatic? Urban Renewal Policies and Socioeconomic Transformations in Copenhagen.” Urban Studies 45(12), 2429–48. Samson, J. 2009. Social segregation i danske byer. En kvantitativ undersøgelse af segregationsmønstre i Hovedstadsregionen, Århus. Odense, Aalborg: Institut for Geografi og Geologi, Københavns Universitet.

Chapter 9

Residential Segregation in Budapest before and after Transition Zoltán Kovács

Introduction Modern urbanization came late to Eastern and Central Europe and followed different pathways from the West in the twentieth century (Enyedi 1990). Eastern and Central Europe was underdeveloped compared with Western Europe in many respects. After World War II with the implementation of the communist system huge social transformations took place in the region. Industrialization was considered to be the main tool to catch up with the West and brought about a robust development of heavy industry and collectivization of agriculture. Towns with industrial functions, as symbols of “modernity” overdeveloped at the expense of the countryside. The high level of mobility (both in social and geographical terms), full employment, strong social security and the growing equality represented the cornerstones of the communist welfare system. From the 1960s onwards, largescale public housing programs were launched and implemented, especially in urban areas, to meet the working classes’ need for housing and, simultaneously, to lower residential segregation. We must also note here that the practice of redistribution of resources (e.g. housing) was not equal during state-socialism either. As Szelényi (1987) pointed out the nomenclature (party officials, managers, members of the armed forces etc.) gained an unequal share of resources. After 1989 the principles of state-socialist redistribution of income and goods were replaced by the rules of the market which inevitably set off profound changes. Communist type inequalities were replaced by capitalist ones. The socio-economic transformations altered-the spatial organization of cities. The growing income differentiation led to social inequality and new dimensions of segregation. In this respect I can truly say that like 1945, the collapse of communism represented once again the beginning of a new era in the urban development of the region. Following the long decades of central planning, the cities of Eastern and Central Europe became subjects of market conditions. Urban researchers repeatedly questioned whether these cities would follow a pathway similar to the Western European cities or whether they retain certain specific features in their development. This question was examined in several thematic volumes (see for example Andrusz, Harloe and Szelényi 1996; Enyedi 1998), and by individual papers focusing on certain countries and cities (Hirt 2006, Standl and Krupickaite 2004, Ruoppilla and Kährik 2003, Sýkora 1999, 2005, Weclawowicz 1998, 2002). Budapest as one of the leading metropolizes in Eastern and Central Europe also became the focus of academic research among urban sociologists and geographers (Dingsdale 1999, Földi 2006, Kovács 1994, 1998, 2009, Kovács and Wiessner 2004, Ladányi 1997, Pickvance 1996). After 1989 in line with the post-socialist transformations an increasing social differentiation took place in the city, resulting in new types of social inequality. The mushrooming of scientific publications clearly reflects the growing academic interest in the new socio-spatial pattern of Budapest.

198

Residential Segregation in Comparative Perspective

Research on class-based residential segregation in Budapest was scarce prior to 1989. Among the few analyses, the study of Probáld (1974), who examined the residential segregation of manual and non-manual workers and the theoretical problem-raising paper of Szelényi (1974) have, first of all, to be mentioned. The most comprehensive analysis, however, was carried out by Csanádi and Ladányi (1985) in the mid-1980s. They investigated the city’s residential segregation over the period between 1930 and 1980. Yet, these earlier studies make it clear how difficult it is to study residential segregation over a long period of time and based on systematic statistical analyses. The main difficulty lies in changing methods of data collection and social categories which changed from time to time. Nevertheless, the importance of the earlier studies must be emphasized here. They were the first publications which used the concept of segregation in Hungary and made it possible to draw a clear picture about what social and residential differences looked like in Budapest under state socialism. In this chapter I will investigate the long-term transformation and pattern of residential segregation in Budapest. As I believe that the city’s segregation pattern today can be explained by the historical pathway the city followed over the last century, I begin with an analysis of the urban development boom in the second half of the nineteenth century. Then I present the pattern of residential segregation during state socialism. Finally, I focus on the post-socialist period in the last two decades, providing an overview on the contemporary pattern of residential segregation in the city. I would like to focus on the following questions: • How has the pattern of residential segregation changed in Budapest over the past century? • How far state socialism could change the historically evolved pattern of social segregation in the city? • What was the impact of post-socialist transformation on the socio-economic structure of the city? • What are the new destinations of residential mobility in the post-communist era and how do they affect the inherited patterns of segregation? Patterns of Residential Segregation in Budapest before State Socialism The historically-evolved ecological structure of Budapest coincides very much with the physical geographical features of the city. The traditional high status areas of the city inhabited by better educated and better off people can be found close to the city center on either side of the Danube, on the hilly Buda side in the West, and in the center of Pest in the East. The pattern of social segregation was very similar to other Central European cities: moving from the city-center toward the urban periphery the social status of residents traditionally declined. Although the enlargement of the territory of Budapest through the annexation of the suburbs in 1950 and mass housing construction in the 1960s and 1970s somewhat altered this picture, the traditional center-periphery dichotomy has remained strong until now (Map 9.1). Budapest with a total population of 1.7 million is subdivided into 23 districts. Each district has elected local assemblies and there is a body of representatives for the whole city of Budapest called the Budapest City Government. Budapest, as one of the youngest capital cities in Europe, was officially established only in 1873, through the unification of three independent and geographically more or less separated towns: Pest, Buda and Óbuda. The city owes a large part of its building stock and virtually its entire urban structure to the late nineteenth century urban boom, when the speed of urban development

Residential Segregation in Budapest before and after Transition

Map 9.1

199

Structure of Budapest

Source: Author’s design.

resembled very much that of North American cities (Bender and Schorske 1994). The last three decades of the nineteenth century was the peak of urban development, due to extensive industrial growth and the subsequent mass immigration of labor from the countryside (Lukacs 1989). However, due to strict planning control the housing stock erected in this period was generally of a high standard with a strong social mix. The typical form of housing was the 3–4 storey tenement building with an inner courtyard containing dwellings of very different size. High-status families normally occupied the street-facing, larger apartments, whereas low-status families rented the smaller back apartments facing the inner courtyards. This architectural pattern allowed a relatively strong social mix at the level of buildings, streets and neighborhoods. However, at the turn of the century housing construction became gradually dominated by speculative builders. This was induced by the mass influx of labor, seeking shelter in the young metropolis, as well as the world economic crisis and the subsequent narrowing of opportunities for capital investment. At the fringe of the compact city, overcrowded shanty towns were expanding with low quality tenement blocks, which were in a sharp contrast with the inner part of the city both in terms of the quality of the housing stock and the social status of the residents (Enyedi and Szirmai 1992). Writing on the pre-1945 ecological structure of Budapest, Beynon (1943) pointed out that ”the traditional seat of the Hungarian aristocracy in the capital naturally was the Vár (Castle District)

200

Residential Segregation in Comparative Perspective

on the Buda side, whereas...the large vice area in Budapest is in the Sixth and Seventh Wards (i.e. districts) between the Eastern and Western Railway Stations.” According to Beynon, the location of slum areas is related to the modern economy of Budapest, “fashionable residences avoided the area because of the blighting effect of the railway stations, with their smoke and aesthetic hideousness.” He also referred to parts of District VII adjacent to the Eastern railway Station as the worst slum in Budapest referred to by the public as “Chicago” (Beynon 1943). During the inter-war period the development of Budapest slowed down. This was partly connected with the altered geopolitical position and isolation of the country and the economic stagnation typical for the period. According to Ladányi (1989) the segregation index of manual and nonmanual workers at the level of districts slightly increased in Budapest between 1930 and 1939. This can be explained basically by two factors: the social consequences of the world economic crisis 1929–1932, and the gradual recovery of the housing market. The world crisis brought about massive decline in factory employment and the impoverishment of wide sections of the society. Changes on the housing market exacerbated the growing social inequalities. Housing rents in Budapest were frozen during World War I in order to protect the families of soldiers from eviction. This decree remained in force after the war and the housing market of the city was liberalized only as late as 1926. New housing projects of the 1930s like Újlipótváros or Lágymányos also contributed to growing segregational indices as high social status groups started to move to these new fashionable neighborhoods. Through these new housing projects smaller or larger high-status ’islands’ developed in different parts of the city in addition to the Buda hills and city-center. In the inter-war period the rate of population growth was also much lower than in the pre-war period, and the direction of migration shifted gradually to the suburbs. This represented the first major phase of suburban growth around Budapest. However, the growth of suburbs could not be attributed to a proper “suburbanization” process (i.e. the mass-movement of middle class families from the city-center toward the periphery) but rather to the so-called “rural urbanization,” in which landless proletarians from the provinces were heading toward the periphery of the city. This flow of labor from the countryside toward Budapest led to the rapid expansion of the “red outskirts” (e.g. Csepel, Ujpest) which became strongholds of the Communist Party. Soon after the war these working class suburbs were amalgamated into Budapest as part of the communist administrative reform, in order to control the development of urban periphery and partly to counterbalance the ’right wing’ bourgeois city. As a consequence of World War II the population of Budapest dropped by more than half a million and the pre-war figure was reached again only in 1972. Residential Segregation during State Socialism There is almost universal agreement in the literature that the degree of social segregation and inequality under communism was much less than under capitalism (Sýkora 2009). After 1945, in line with the political and economic changes, a huge social transformation took place in Budapest. The wealthiest sections of the former landlord and capitalist strata (7.9 percent of the population according to the 1941 Census) had left the country by the end of the 1940s. Several thousand families, mainly members of the former upper class were deported from Budapest in 1951–52, partly for political reasons. At the end of the 1940s a new communist constitution was implemented in Hungary, land and property was nationalized and nearly all commercial functions were prohibited or severely controlled (Enyedi and Szirmai 1992). Capitalist extremes of social inequality were attacked and to a large extent eliminated via different channels (e.g. social housing provision, the new rent and wage system). The rapid expansion of heavy industry in the 1950s and

Residential Segregation in Budapest before and after Transition

201

the consequent high rate of immigration considerably increased the size of the working class, as did the annexation of the 23 formerly independent suburbs in 1950. Perhaps the biggest loser of the communist transformation was the urban intelligentsia. According to the statistics, before World War II wage differences between white and blue collar employees in Budapest stood at 2.5–3.0 to 1.0, in favor of white collar professionals. Due to the new wage system this difference was virtually eliminated by the early 1960s. As part of the state-socialist welfare system people were provided for by the state with highly subsidized food, housing, transport and basic necessities, and free education and healthcare services. Universal social insurance funds such as those for pension, health care and family allowance were part of the state budget. Esping-Andersen summarized the main features of the communist welfare system; “The old communist regime was characterized by three pillars: full employment and quasiobligatory employment: broad and universalistic social insurance: and a highly developed typically company-based, system of services and fringe benefits” (Esping-Andersen 1996, 9). The capitalist form of housing production and distribution which was blamed for previous inequalities and segregation was abolished and replaced by a communist type housing system. Housing was proclaimed a universal right that meant that every family was entitled to its own dwelling at low cost. A radical nationalization policy abolished the private rental sector dominant in the inter-war period and replaced it with the growing public rental sector. The state took responsibility for the maintenance and renovation of housing and kept rents artificially low, for political reasons. The low rents did not even cover the operational costs of the buildings, nor allow proper building maintenance. As a consequence, the old housing stock in the city core started to deteriorate severely. Larger dwelling units were subdivided for and reallocated to the working class. The state also began mass state house-building programs from the beginning of the 1960s to decrease the level of segregation. The intense development of large-scale housing estates in peripheral urban locations adopted standardized dwellings and was intended to make housing more homogeneous (Compton 1979). Social differences started to diminish very quickly under communist welfare and housing policies. This change is confirmed by Csanádi and Ladányi (1985) who found significant decreases in residential segregation over the period between the second half of the 1940s and first half of the 1950s. The difference compared to the previous period was so marked that the authors presumed that different sources of the data might have had neglected influences on the result. They also pointed out that the decreasing trend slowed down noticeably in the 1950s and gradually began to increase residential segregation after that. The main reason why residential segregation began to increase was that Hungary moved away from the Stalinist model of redistributive economy and egalitarianism in the 1960s, and started to liberalize its economy in a set of reforms known as the “New Economic Mechanism” (1968). The reforms recognized the multi-sector nature of the economy, with state, co-operative and private, small-scale economic activities enjoying equal rights within a socialist economy. At the individual level it meant more opportunities for profit-making, and more freedom in cultural and private life. This was basically the so-called “goulash communism”—a mixture of economic policies and personal opportunities. The role of private initiatives increased. The artificially-repressed social inequalities reappeared and income differences started to grow again. Better housing, westernmade consumer goods, cars and second homes became available for the better-off strata. The city also entered a new phase of economic development from the 1970s. Structural economic changes brought a massive decline in factory employment and a rapid growth of services, especially trade and tourism services. As a result, the ratio of active earners employed in manufacturing industry decreased from 54 to 36 percent between 1970 and 1990. This process accompanied the

202

Residential Segregation in Comparative Perspective

rapid expansion of the second economy and further enhanced social inequalities during the 1980s. The term “second economy” was used in opposition to the first (i.e. state sector) economy during communism (Hann 1990). It included full-time entrepreneurs and artisans in the private sector, and many kinds of ‘informal’ work performed by persons who also had ‘formal’ employment in some branch of the state sector. This diverse informal work ranged from the continuity of work actively encouraged by the state (e.g. in agriculture) to criminal activity. Estimates of the second economy varied considerably, but economists calculated that about 20 percent of total national income was produced by the informal economy sector in the late 1980s. With this figure Hungary—together with Poland—was clearly the front-runner for the second economy among the communist countries. Housing policy also began to intensify inequalities and residential segregation after the end of 1960s (Szelényi 1987). New state measures introduced in 1971 and 1983 were to abolish the previous deeply subsidized nature of the housing system. These reforms could be seen as an admission on the part of the state that it was not capable of satisfying the housing demand of all families, and that it should have targeted its resources on the lower strata. The large scale withdrawal

Map 9.2

Percentage of university and college graduates in Budapest urban planning districts (1990)

Source: National Census 1990. Hungarian Central Statistical Office, Budapest, author’s calculation and design.

Residential Segregation in Budapest before and after Transition

203

of the state and state builders from the housing market represented a green light for private forms of housing provision. Then expansion of the private sector led to increasing inequalities in the housing market. The growing share of co-operative and private housing construction allowed those who had substantial resources and mostly earned in the second-economy to leave the state rental tenure and provided them with an opportunity to make profits. Segregation trends were further intensified with the economic difficulties of the 1980s. Budapest reached a relatively high level of segregation by the eve of political changes, compared with other East European communist cities (Kovács 1990). In order to demonstrate the dimensions of residential segregation on the eve of political changes and after the first big wave of transition I use census data on university and college graduates in 1990 (Map 9.2). For the sake of analysis I use the system of urban planning units which are functionally and morphologically more or less homogeneous areas between the levels of districts (23) and census tracts (over 10,000). There are 522 urban planning units in Budapest. The map already shows a very strong east-west polarization in the social structure of the city in 1990. The highest social status areas of Budapest could be found on the Buda hills in a compact area. The segregational pattern between high-status and lowstatus strata shows distinct differences. Writing on the segregational pattern of Budapest in the 1980s, Ladányi (1989) pointed out that social segregation in the city historically did not follow the classical ecological models because whereas low status groups were more likely to be concentrated in micro-segments (blocks, streets etc.), the upper and upper-middle strata lived more or less in one homogeneous zone in the western (Buda) side of the city. The increasing social and residential segregation also appeared in the suburbs. Suburbanization in the Western sense did not emerge in Budapest until the late 1980s. The continuous growth of suburbs throughout the 1970s and 1980s was due to the influx of migrants from the countryside. It was only after the mid-1980s that the outflow of people from the city let the suburban ring grow faster than the urban core. In the suburban belt white-collar strata gravitated toward the environmentally more attractive hilly areas to the north and west of Budapest, while manual workers were overrepresented on the eastern and southern sides of Pest. Factors of Socio-spatial Differentiation after 1989 Gradual processes of urban change are sometimes interrupted by radical urban restructuring influenced by turbulent developments in the wider socio-economic systems. A recent example of such changes is the urban reshuffle in post-socialist countries conditioned by the transition from single party totalitarian to democratic political regimes and from centrally planned to market economies (Sýkora 2009). Post-socialist transition consists of multiple transformation processes that include both the political application of normative concepts as well as spontaneous societal changes following the re-established market environment. Out of the great number of transformations that took place in East Central Europe after 1990 I focus on social and economic transformation and housing market change. Transformation of the Economy Soon after the change of political system the old structure of the Hungarian economy collapsed and it underwent profound changes in the first half of the 1990s. In 1990 the private sector accounted for only 10 percent of GDP, whereas by 1994, 65 percent of the national GDP was already produced by foreign and privately owned domestic firms. On the other hand, the restructuring process was

Residential Segregation in Comparative Perspective

204

protracted as GDP fell by 20 percent, agricultural production declined by 40 percent and industrial production by 33 percent during the same period. Large state companies were privatized and/or disintegrated. The decline of industry was marked in Budapest, lowering the share of industrial workers from 36 to almost 20 percent of the active population between 1990 and 2001. The impact of the economic breakdown was tremendous but was still less dramatic in Budapest than in the rest of the country. Also the economic recovery was much faster in the capital city, just like its integration into the European and global networks which commenced quickly in the first half of the 1990s. After 1990 Budapest and its metropolitan region became the magnet for capital investment. Due to the aforementioned conditions causing its high competitiveness, Budapest became a major target of FDI basically due to its size, favorable geopolitical position and good accessibility from Western Europe. In the first decade of transition after 1990, 57 percent of the FDI arriving in Hungary was concentrated in Budapest. The special position of Budapest could also be attributed to the high concentration of human and financial capital and the highly developed infrastructure and means of production. Typical of the weight of Budapest in the Hungarian economy was that in 1996 35 percent of the national GDP was produced there, and that the per capita GDP was 181 percent of the national average. Since then the weight of the city in the national economy has not decreased at all. In 2008 the city produced 38 percent of the Hungarian GDP, and the per capita GDP was 221 percent of the national average. In certain sectors of the economy (e.g. cultural industries, R&D, business services, ICT) the role of Budapest is extremely dominant (Kovács et al. 2007). Societal Transformations The economic change mediated through labor markets led to huge social transformations after 1990. Economic restructuring was expressed in growing wage and income inequalities and a social polarization within the society. As classic indices of income inequality such as the Robin Hood index, the q10 index or the Gini coefficient all show, income disparities first gradually decreased in Hungary until the middle of the 1980s, then started to increase drastically (Table 9.1).1 Table 9.1

Changing income inequalities in Hungary (1972–2004)

Robin Hood q10 Gini coefficient-Index

1972

1977

1982

1987

1995

2004

16.6 4.93 0.232

15.0 4.13 0.212

14.9 3.80 0.206

17.0 4.71 0.236

21.0 7.55 0.296

21.4 7.55 0.312

Source: Éltető and Havasi (2009).

In line with economic transformation the welfare system was also transformed and social policy gradually inclined to neoliberalism. As opposed to communism unemployment became 1  The Gini coefficient was already introduced in Chapter 4. Robin Hood index, which is also known as Hoover index, is a measure of income inequality ranging from 0 (complete equality) to 100 (complete inequality). The value of the index approximates the share of total income that needs to be transferred from households above the mean to those below the mean to achieve equality in the distribution of incomes. Similarly, q10 index refers to the quotient of the averages (shares) of the upper and bottom deciles.

Residential Segregation in Budapest before and after Transition

205

widespread, welfare benefits started to deteriorate, the universal benefits and services such as family allowance and child care leave with grant were converted into means-tested ones. The replacement of the centrally-planned state socialist economy and welfare system by the market economy and neoliberal welfare system intensified social differentiation and launched a substantial polarization process, with a growing distinction between the lowest and the highest segments of the society. This process also became widespread in other post-communist cities (Weclawowicz 1998). According to Ladányi and Szelényi (1998) the major beneficiary of the transformation was the late communist technocracy and the managerial elite. While the majority of the former communist nomenclature lost its positions a new economically privileged stratum emerged, which was composed of the technocratic-managerial elite, their opinion-forming intellectual allies and private entrepreneurs. Later, this group was expanded by the new wave of young, highly educated technocrats working in financial and business services. On the other hand, the widening urban poor came to be formed mainly of the elderly, the unemployed and other socially disadvantaged groups. The urban poor could be considered commonly as the losers of post-communist transformation (Smith and Timár 2010). The closure of many state enterprises meant a loss of jobs and a sharp increase of unemployment, a phenomenon unknown in the communist period. Unemployment and impoverishment affected especially the less-educated groups like the Roma. In the mid-1990s about a third of the total Hungarian population lived below the poverty level, however, among the Roma the figure was 61 percent. Even though the scale of impoverishment was less striking in Budapest and its agglomeration zone than in the poorest regions of the country, income inequalities reached their peak values in the capital as indicated by the Gini coefficient (Table 9.2). Table 9.2

Gini coefficient of income distribution in Hungary by settlement categories (1987–2004)

Budapest Other cities Villages Hungary

1987

1995

2004

0.252 0.234 0.225 0.236

0.336 0.291 0.257 0.296

0.362 0.355 0.348 0.312

Source: Éltető and Havasi (2009).

The growing income inequality also generated new forms of spatial mobility and left its mark on post-communist patterns of spatial segregation. Marked segregation patterns prevalent already in the 1980s took extreme forms after 1990. The affluent population established its own new enclaves and segregated itself from other districts (Sýkora 2009). Core areas of the city and well-situated suburbs became desirable to the educated and wealthy groups, while physically deteriorated urban quarters, slums and cheap housing estates attracted the urban underclass. The consequences of socio-economic restructuring expressed in territorial disparities were nowhere so marked and distinguished as in Budapest and its urban region.

206

Residential Segregation in Comparative Perspective

Transformation of the Housing Market A basic precondition for new forms of residential mobility and the development of new patterns of segregation was the liberalization of the housing market. Prior to 1989 residential mobility was restricted in Budapest, due to administrative restrictions (e.g. only those who had spent at least five years at a workplace or studied in Budapest had the right to settle in the city), the overall shortage of housing and the high level of state ownership. In the transformation of the housing market, the privatization of public dwellings played a very important role. Before the political changes state housing made up 51 percent of the total dwelling stock in Budapest. In most of the centrally located districts housing was owned predominantly (95–97 percent) by the state. Privatization of state housing in Budapest meant a pure ‘give away’ type of privatization to sitting tenants, at a very low price. Most state dwellings were sold for between 15 and 40 percent of the estimated market value depending on the physical conditions of the dwelling. A further 40 percent discount was offered to those who paid in cash, which meant that the great majority of the public dwellings was sold for 9 percent of the market value. In the light of these circumstances it is easy to understand that the best quality (i.e. most valuable) segment of the public housing stock was the first to be privatized. The logic of privatization favored the better-off families, since tenants (now “buyers”) of best quality dwellings with desirable locations saw the largest growth in property prices. These changes in the housing market had far-reaching ecological implications. First of all, households had more opportunity to realize their relocation desires as they became owners, secondly, new conditions allowed a greater plurality of values and promotion of self-interest. Factors in housing preferences of households like security or access to green spaces gained great importance. The outcome was a rapid differentiation of the housing market. While public housing became residualized and served as a shelter for the urban poor, new residential areas for the better off were developed. Geographically, public housing became increasingly concentrated in traditionally low-status areas (Józsefváros, Erzsébetváros). In a survey conducted in summer 1995 we found that households who remained in the public housing sector were on average less-educated, with lower incomes and were much more likely to be unemployed than other housing classes (Kovács 1998). All these changes on the housing market left their mark on the pattern of residential segregation in Budapest. New Patterns of Residential Segregation after 1990 Post-socialist economic transition and the transformation of the housing market led to increasing differentiation of both housing classes and neighborhoods in Budapest. As I pointed out earlier, socio-spatial differences existed even in the heyday of communism. However, the advent of the market economy and the transformation of the housing market have exacerbated these differences. Map 9.3 shows the city’s residential segregation by the degree of education in 2001. Firstly, there is some continuity between 1990 and 2001 as the stronghold of intelligentsia is located on the Buda side of the town. On the other hand, it can also be seen on the map, that the boundary of high status areas has expanded outwardly into neighboring districts further into the north and south on the Buda side and spread to smaller clusters of the elite on the Pest side. These imply that the compactness of higher-status areas described by Ladányi (1989) for the pre-1990 Budapest gradually diminished after the political changes and a patchwork type spatial pattern evolved. Similarly, a more dispersed pattern of segregation has been found in Warsaw and other Polish cities (Weclawowicz 2002) and Prague (Sýkora 1999).

Residential Segregation in Budapest before and after Transition

Map 9.3

207

Percentage of university and college graduates in Budapest urban planning districts (2001)

Source: National Census 2001. Hungarian Central Statistical Office, Budapest, author’s calculation and design.

What are the reasons for the patchwork type development of segregation and how the different sections of the city have been affected since 1990? To answer this question I divided the city into different functional zones following the traditions of classical human ecology (Figure 9.1). Each zone is characterized by distinct socio-economic, functional and architectural features that are the outcome of city growth in the last 130 years, when the expansion of Budapest developed concentrically from the city center outwards. The zone of agglomeration comprises 80 municipalities around the city. I indicate the directions of changes after 1990 in the model. After 1990 perhaps the most radical changes in the ecology of Budapest took place in the inner city and the suburban zone. In the inner-city neighborhoods we could observe both upward and downward processes in the post-socialist period. Upgrading became first apparent in the very narrow strip of the city center (district 5) where, due to rapid expansion of services and business investment, substantial changes had already commenced in the early 1990s. The influx of global capital in the form of company headquarters generated high demand for high quality office space. Consequently, old traditional buildings were renovated and vacant lots were rebuilt with offices

Residential Segregation in Comparative Perspective

208

Figure 9.1

Transformation of Budapest Metropolitan Area after 1990

Source: Author’s design.

that reflected the rapid integration of the city into global capitalism (Smith 1996). The inflow of investment was also enhanced by mass tourism resulting in a boom in hotel construction, and the mushrooming of tourist facilities. At the same time city functions started to expand in the adjacent neighborhoods, especially those lying close to the Danube (Földi and Van Weesep 2007). The pattern of modernization in the rest of the inner city, the so-called inner residential zone, has been more mosaic-like. Upgrading areas can be classified into two groups: neighborhoods affected by spontaneous (market-led) renewal, and those affected by rehabilitation programs (Kauko 2006). The former contains mainly areas close to the CBD where rent gap is highest. These quarters are typically affected by isolated renovations where the main actors are often foreign investors (e.g. along Andrássy út). According to real estate experts the emergence of a considerable demand from foreign citizens dates from Hungary’s accession to the European Union. As a result of the unlimited right of EU citizens to obtain property, the Spanish, British and Irish were the first to massively invest in housing in these inner-city quarters. The group of foreigners buying apartments in the historical neighborhoods has also been expanded by professionals working in company headquarters and bank offices in the city center and preferring to live close to their place of work.

Residential Segregation in Budapest before and after Transition

209

More extensive upgrading could be observed beyond the arc of the Grand Boulevard, in certain parts of Middle-Ferencváros and Middle-Józsefváros, where local rehabilitation programs have been implemented in recent years. Both quarters share the same history as inner-city lower-class neighborhoods, with high densities of population, and the late nineteenth-century tenement blocks with inner courtyards. After the first 15 years of rehabilitation the social composition of these neighborhoods changed considerably, due to the practice of rehabilitation and the housing policy of the local (district) governments. When a building is selected for demolition, tenants of the building are either placed in apartments owned by the municipality or they receive a maximum of 90 percent of the market value of the flats in cash and leave. Since most of the flats are in bad condition with low levels of comfort, the market value of these dwellings is also relatively low. The process is somewhat different if the building is subject to renovation. In this case the local government decides who can come back after renovation and who has to leave. To those who are selected to leave (problem cases with rent arrears, etc.) three possible rentals are offered by the municipality and they must choose one (just like tenants of buildings designated for demolition). Those who are selected to return will be moved to a temporary shelter during the reconstruction. Through this mechanism the large majority of original residents of the neighborhood are moved out and replaced by younger, better-off families. In these neighborhoods displacement of the original population is occurring with the active participation of the local governments and the process might be labeled as “organized gentrification.” Next to upgrading areas we can also find neighborhoods in the inner city of Budapest where the outcome of transformation was not revitalization but further physical and social decline. These are typically old, working-class neighborhoods with multi-storey tenement buildings in the eastern periphery of the inner city, e.g. Józsefváros (Magdolna quarter), Erzsébetváros (next the Eastern railway station). The reasons for the physical decay of these residential quarters are manifold. One of them is the long-lasting neglect of maintenance and infrastructural development during communism. The lack of investment was not alleviated by the political changes and subsequent transfer of ownership (i.e. privatization), either. Most of the new owners (former tenants) had neither capital nor expertise and entrepreneurial spirit to renovate their housing, many of them being old and poor. The deterioration of the building stock of the old working-class neighborhoods was accompanied by the erosion of the local community. Residential mobility played an important role in this process, as younger and more affluent households gradually left these neighborhoods. Later they were followed by the blue-collar workers of active age, and finally the mobile part of the elderly, e.g. those with second homes or with children or relatives in the countryside. As a result of this filtering down process the population of these old residential quarters has become more and more marginalized. Typically Roma families are overrepresented among the residents of such neighborhoods (Ladányi 2002). The increasing concentration of urban poor in Budapest can be explained mostly by the functioning of the housing and labor markets. In the run-down parts of the inner city the residual state housing sector is dominant, providing low quality, cheap housing for poor families. Roma families attracted by the formal and informal job opportunities of Budapest can find accommodation only in these run-down neighborhoods, which stand out due to their lack of minimal sanitation and comfort, and their very high crime rates. The growing concentration of poverty and higher level of residential segregation can be considered to be the outcome of the development of a free, market based housing system accompanied by a highly competitive capitalist labor market (Ladányi 1993). Another, but equally important, factor in urban change was the migration of better-off households to the suburbs after 1990. Since the beginning of the transition period, the population

210

Residential Segregation in Comparative Perspective

of Budapest has decreased by over 300,000 residents—from a little over 2 million in 1990 to about 1.7 million in 2009. This sharp population loss is a result of a combination of natural decrease (accounting for two-thirds of the population decline) and an accelerated migration of Budapest residents to the suburban periphery. As a consequence of suburbanization the balance of population between the core city and the agglomeration has shifted. In 1990 only 17 percent of the functional urban region resided in the zone of agglomeration whereas by 2009 the figure was 35 percent. Before 1990, the communist state made hardly any investments in communal infrastructure or services in the suburban areas, which consequently were not considered desirable at all by the better-off households. From the mid-1980s a new migration tendency started in the Budapest urban region, which could be labeled as initial, “low intensity” suburbanization, marked by the change of the balance of migration of Budapest against the surrounding Pest county into negative. The loss was not significant for several years and was still counterbalanced by a positive migration balance to Budapest from other parts of Hungary. From the early 1990s, however, the period of intense suburbanization started. Due to suburbanization the city’s population decreased rapidly, almost by 18,000 persons a year. The outflow of people from Budapest, however, gradually reversed after 2000. The social consequences of suburbanization are well-known in our cities as it was usually the young, middle class families with children who tended to leave the core city. The selective residential deconcentration of the population resulted in further changes in the social and demographic characteristics of the Budapest metropolitan region. The social status of the suburbs has generally increased and at the same time the population has become younger. This process is confirmed by statistics on educational level, the proportion of people with university or college degree (within the age group 15+) increased from 3.2 to 12.7 percent in the zone of agglomeration between the last two censuses (1990 and 2001). To some extent this can be perceived as lowering the level of segregation of the highly educated between the city and its agglomeration (Figure 9.2). However, suburbanization was fairly uneven in space, creating new dimensions of socio-spatial disparities. Generally, municipalities lying north and northwest to Budapest were very much affected by suburbanization. In contrast, municipalities in the eastern and southern parts of the metropolitan region, characterized by less scenic natural conditions, were not much preferred by those who left Budapest. During communism, these settlements attracted the majority of the lowskilled labor force migrating from rural areas to the capital city and subsequently from the late 1980s and 1990s, a distinct pattern of social polarization has taken shape in the outer periphery of Budapest. The northern and northwestern regions have become the stronghold of upper-middle class residents, whereas the southern and eastern parts of the metropolitan area were inhabited primarily by lower-income households. Neighborhoods lying between the inner city and the suburban zone have shown controversial development over the last two decades. A downward spiral is typical for large housing estates erected during communism. The popularity of housing estates in Budapest was at its peak in the 1970s, when large pre-fab estates were developed in peripheral locations by the state (Egedy 2000). Dwellings located in housing estates were very popular at that time, especially among young families with children, due to their relatively high level of comfort. On the eve of the political changes, 36 percent of the population already lived on housing estates in Budapest, which comprised altogether 268,000 dwelling units (typically two-roomed flats). After 1990 no new housing estate project was completed in Budapest. On the other hand, existing housing estates lost their popularity rapidly because of their architectural monotony, lack of green spaces, decreasing security and relatively high costs of amenities (especially heating). This was directly reflected in the gradual escape of younger and better off people from these estates. According to census statistics more than 110,000

Residential Segregation in Budapest before and after Transition

Figure 9.2

211

Ratio of university and college graduates in different urban zones of Budapest (1990, 2001)

Source: National Census 1990 and 2001. Hungarian Central Statistical Office, Budapest, author’s calculation.

people left the housing estates between 1990 and 2001 and their total population decreased by 15.2 percent. Due to the highly selective out-migration the social composition of housing estates has also been changing, the ratio of elderly is increasing, as is that of the socially disadvantaged families. Due to the decreasing demand for such dwellings, the housing estates of the communist period, especially the high-rise pre-fab housing estates, have become increasingly isolated on the housing market. With growing social and income differentiation, and concomitant increasing degrees of segregation, there is a real potential for such estates to become ghettos in the future. In contrast, the outer residential zone of Budapest comprised of traditional villages and garden cities attached to Budapest in 1950 has been through a gradual rejuvenation. The most important advantages of this formerly socially-mixed and rather insufficiently developed zone are: its low rise character with lots of green spaces, relatively good accessibility and an unlimited quantity of plots for new development. These factors made the zone attractive both for individuals and some major real estate developers who built larger housing projects in this zone. As a consequence, nearly half of the 40,000 new dwelling units that were built in Budapest between 1990 and 2001 were constructed in this zone. A new phenomenon in Hungary is the mushrooming of the so-called residential parks (“lakópark”). These residential parks are very similar to the North-American “gated communities” as they are completely cut-off from the surrounding areas and provide lots of additional services for the residents (e.g. guard and security system, parking facilities, playgrounds). According to Hegedűs (2009) there were 183 such complexes within the boundary of Budapest in 2009 comprising approximately 31,000 dwellings. The boom of residential parks started in the 1990s and it was supported by several factors. Due to communist housing policy the upmarket segment was very narrow in Budapest before 1990. After the change of regime there was a high demand on the housing market arose for larger housing units located in attractive (green) environment,

212

Residential Segregation in Comparative Perspective

equipped with extra services like security system, fencing, gardening, etc. International and domestic developers recognized this demand and started to satisfy it. Loose planning control and a boom in the credit and banking system also helped them to realize these projects and satisfy the dreams of the wealthiest (Tosics 2005). Since these dwellings are affordable only for better-off Hungarian households and foreigners, the construction of residential parks resulted in the rising status of the outer residential zone. The proportion of inhabitants with higher education rose from 10.5 to 16.6 percent between 1990 and 2001, and at the same time the population of this belt became also younger. The process of rejuvenation and upgrading of the outer residential belt can be considered a kind of “internal suburbanization” in Budapest. Conclusion This chapter has shown that residential segregation in Budapest has a complex history. Before World War II the dimensions of segregation and the driving forces behind them were very similar to other Central European cities (Vienna, Berlin, for example). Segregation was basically the outcome of class differences typical for the then capitalist societies and a market-based housing system where private rental sector prevailed. However, central planning along with state ownership of land introduced by the communist system in the late 1940s changed the organically-evolved pattern of segregation and resulted in new (‘socialist’) forms of urban development. Social inequalities were attacked and artificially repressed in the early phase of communist regime, but after a while new inequalities were produced by the state. Better-off people and members of the nomenclature segregated themselves in the villa quarters of the Buda hills, where previously aristocrats and members of the bourgeoisie had been concentrated. At the same time the densely built historical neighborhoods were seriously neglected and started to deteriorate both physically and socially. Later on, the construction by the state of good quality, large housing estates provided the younger and better-off part of society with a place from which other groups were excluded. After the collapse of the communist system in 1989, Budapest went through far-reaching changes in socio-economic conditions. Due to the re-establishment of market mechanisms, the privatization of housing, the liberalization of the property market, and the growing presence of global capital, the social structure of the city has started to change abruptly. Extreme forms of residential segregation—including ghettoization on the one hand, and the mushrooming of gated communities on the other—started to symbolize free-market political and economic conditions and the weakness of the welfare state. Budapest implies that there is not a single model for post-socialist urban transformation, demonstrating that path dependency and local transformation policy and politics really matter. In this respect Budapest represents a radical and shock-therapy model of urban restructuring which has accompanied unexpectedly strong socio-economic disparities. This chapter has also revealed that urban processes (e.g. gentrification, suburbanization, ghetto formation) previously known from Western Europe quickly evolved in Budapest as the post-socialist transformation commenced. Does this mean that the city has finally joined the European path of urbanization? It remains to be seen.

Residential Segregation in Budapest before and after Transition

213

References Andrusz, G., Harloe, M. and Szelényi, I. eds. 1996. Cities after Socialism. Oxford: Blackwell. Bender, T. and Schorske, C.E. eds. 1994. Budapest and New York: Studies in Metropolitan Transformation, 1870–1930. New York: Russell Sage. Beynon, E.D. 1943. “Budapest: An Ecological Study.” Geographical Review 33(3), 256–75. Compton, P. 1979. “Planning and Spatial Change in Budapest.” In The Socialist City, eds R.A. French and F.E.I. Hamilton. New York: John Wiley, 461–91. Csanádi, G. and Ladányi, J. 1985. Budapest – a városszerkezet történetének és akülönböző társadalmi csoportok városzerkezeti elhelyezkedésének vizsgálata [Budapest – a study of the history of urban structure and the spatial distribution of social groups]. Budapest: University of Economics. Dingsdale, A. 1999. “Budapest’s Built Environment in Transition.” GeoJournal 49(1), 63–78. Egedy, T. 2000. “The Situation of High-rise Housing Estates in Hungary.” In Hungary Towards the 21st Century. The Human Geography of Transition. Studies in Geography in Hungary 31, ed. Z. Kovács. Budapest: Geographical Research Institute, Hungarian Academy of Sciences, 169–85. Éltető, Ö. and Havasi, É. 2009. “A hazai jövedelemegyenlőség főbb jellemzői az elmúlt fél évszázad jövedelmi felvételei alapján” [Main characteristics of income inequalities in Hungary on the basis of income records in the last half a century]. Területi Statisztika 87(1), 5–40. Enyedi, G. 1990. “Specific Urbanization in East-Central Europe.” Geoforum 21(2), 163–72. Enyedi, G. and Szirmai, V. 1992. Budapest – A Central European Capital. London: Belhaven Press. Enyedi, G., ed. 1998. Social Change and Urban Restructuring in Central Europe. Budapest: Akadémiai Kiadó. Esping-Andersen, G. 1996. “After the Golden Age? Welfare State Dilemma in a Global Economy.” In Welfare States in Transition, ed. G. Esping-Andersen. London: Sage, 1–31. Földi, Z. 2006. Neighbourhood Dynamics in Inner-Budapest – A Realist Approach. Netherlands Geographical Studies 350. Utrecht: Faculteit Rumtelijke Wetenschappen Universiteit Utrecht. Földi, Z. and Van Weesep, J. 2007. “Impacts of Globalisation at the Neighbourhood Level in Budapest.” Journal of Housing and the Built Environment 22(1), 33–50. Hann, C.M. 1990. “Second Economy and Civil Society.” In Market Economy and Civil Society in Hungary, ed. C.M. Hann. London: Frank Cass, 21–44. Hegedűs, G. 2009. “A Review of Gated Communities in Some Hungarian Cities.” Geographica Pannonica 13(3), 85–96. Hirt, S. 2006. “Post-socialist Urban Forms: Notes from Sofia.” Urban Geography 27(5), 464–88. Kauko, T. 2006. Urban Housing Patterns in a Tide of Change. – Spatial Structure and Residential Property Values in Budapest in a Comparative Perspective. Amsterdam: IOS Press. Kovács, Z. 1990. “Rich and Poor in the Budapest Housing Market.” In Market Economy and Civil Society in Hungary, ed. C.M. Hann. London: Frank Cass, 110–24. Kovács, Z. 1994. “A City at the Crossroads: Social and Economic Transformation in Budapest.” Urban Studies 31(7), 1081–96. Kovács, Z. 1998. “Ghettoization or Gentrification? Post-socialist Scenarios for Budapest.” Netherlands Journal of Housing and the Built Environment 13(1), 63–81. Kovács, Z. 2009. “Social and Economic Transformation of Historical Neighbourhoods in Budapest.” Tijdschrift voor Economische en Sociale Geografie 100(4), 399–416.

214

Residential Segregation in Comparative Perspective

Kovács, Z., Egedy, T., Földi, Zs., et al. 2007. Budapest. From State Socialism to Global Capitalism. Pathways to Creative and Knowledge-based Regions. ACRE Report 2.4. Amsterdam: University of Amsterdam. Kovács, Z. and Wiessner, R. 2004. “Budapest – Restructuring a European Metropolis.” Europa Regional 12(4), 22–31. Ladányi, J. 1989. “Changing Patterns of Residential Segregation in Budapest.” International Journal of Urban and Regional Research 13(4), 555–72. Ladányi, J. 1993. “Patterns of Residential Segregation and the Gypsy Minority in Budapest.” International Journal of Urban and Regional Research 17(4), 30–41. Ladányi, J. 1997. “Social and Ethnic Residential Segregation in Budapest.” In Prozesse und Perspektiven der Stadtentwicklung in Ostmitteleuropa, ed. Z. Kovács and R. Wiessner. Passau: L.I.S. Verlag, 83–95. Ladányi, J. 2002. “Residential Segregation among Social and Ethnic Groups in Budapest during the Post-communist Transition.” In Of States and Cities, eds P. Marcuse and R. van Kempen. Oxford: Oxford University Press, 170–82. Ladányi, J. and Szelényi, I. 1998. “Class, Ethnicity and Urban Restructuring in Post-communist Hungary.” In Social Change and Urban Restructuring in Central Europe, ed. G. Enyedi. Budapest: Akadémiai Kiadó, 67–86. Lukacs, J. 1989. Budapest 1900. A Historical Portrait of a City and its Culture. New York: Weidenfeld and Nicolson. Pickvance, C.G. 1996. “Environmental and Housing Movements in Cities after Socialism: The Case of Budapest and Moscow.” In Cities after Socialism, eds G. Andrusz, M. Harloe and I. Szelényi. Oxford: Blackwell, 232–67. Probáld, F. 1974. A Study of Residential Segregation in Budapest. Budapest: Annales. Ruoppila, S. and Kährik, A. 2003. “Socio-economic Residential Differentiation in Post-socialist Tallinn.” Journal of Housing and the Built Environment 18(1), 49–73. Smith, A. and Timár, J. 2010. “Uneven Transformations: Space, Economy and Society 20 Years after the Collapse of State Socialism.” European Urban and Regional Studies 17(2), 115–25. Smith, N. 1996. The New Urban Frontier: Gentrification and the Revanchist City. London: Routledge. Standl, H. and Krupickaite, D. 2004. “Gentrification in Vilnius (Lithuania) – the Example of Uzupis.” Europa Regional 12(1), 42–51. Sýkora, L. 1999. “Processes of Socio-spatial Differentiation in Post-communist Prague.” Housing Studies 14(5), 679–701. Sýkora, L. 2005. “Gentrification in Post-Communist Cities.” In Gentrification in a Global Context. The New Urban Colonialism, eds R. Atkinson and G. Bridge. London: Routledge, 90–105. Sýkora, L. 2009. “Post-Socialist Cities.” In International Encyclopedia of Human Geography Vol. 8, eds R. Kitchin and N. Thrift. Oxford: Elsevier, 387–95. Szelényi, I. 1974. Urbanizáció és életmód alakulása Budapesten [Urbanisation and way of living in Budapest]. Budapest: Tanulmányok Budapest Múltjából. Szelényi, I. 1987. “Housing Inequalities and Occupational Segregation in State Socialist Cities.” International Journal of Urban and Regional Research 11(1), 1–8. Tosics, I. 2005. City Development in Central and Eastern Europe since 1990: The Impacts of Internal Forces. In Transformation of Cities in Central and Eastern Europe, eds. F.E.I. Hamilton, K. Dimitrowska-Andrews and N. Pichler-Milanovic. Tokyo: United Nations University Press, 44–78.

Residential Segregation in Budapest before and after Transition

215

Weclawowicz, G. 1998. “Social Polarisation in Postsocialist Cities: Budapest, Prague and Warsaw.” In Social Change and Urban Restructuring in Central Europe, ed. G. Enyedi. Budapest: Akadémiai Kiadó, 55–64. Weclawowicz, G. 2002. “From Egalitarian Cities in Theory to Non-egalitarian Cities in Practice: Changing Social and Spatial Patterns in Polish Cities.” In Of States and Cities, eds P. Marcuse and R. Van Kempen. Oxford: Oxford University Press, 183–99.

This page has been left blank intentionally

Chapter 10

The Limits of Segregation as an Expression of Socioeconomic Inequality: The Madrid Case1 Marta Domínguez, Jesus Leal and Elena Martínez Goytre

Introduction This chapter tries to show that in the Madrid Region social changes, in terms of socioeconomic cleavages and ethnic differences, within a framework of liberal urban policy and a welfare state typical of southern European cities involving the absence of social housing, are leading to an increase in social inequalities without necessarily increasing segregation. The aim of this chapter is, therefore, to analyze the relation between social change and the transformation of Madrid’s social morphology, with particular emphasis on the relation between social inequality and spatial segregation. The way social spaces are distributed in large cities represents an important indicator of the characteristics of an urban society and the changes in this distribution and the practices within urban space are important social processes. Cities are spaces, and the way in which they are perceived influences social practices in a subjective way, as shown by Simmel (1950), but also in a “social way,” as authors like Wirth (1938) argued. The study of the main social processes taking place in these cities’ spaces continues to be a major research aim. During the 1920s the Chicago School tried to establish the basic ecological processes that characterize human settlements. Segregation was considered to be a natural (biotic) process whereby individuals with similar characteristics tended to group together within specific places, differentiated from those of other groups. “Segregate,” in its etymological meaning, means to separate. The majority of texts dealing with urban segregation see it as the residential separation of different social or ethno-racial groups from a broader population within cities. The concept of segregation has very often been misinterpreted. This is partly due to the fact that it is not usually understood in its contradictory nature which relates it at the same time to separation and to social integration and cohesion. Segregated neighborhoods may show a high level of integration based on a high level of social solidarity among their inhabitants, which would be impossible if the people involved were spatially dispersed. However, the spatial perception of inequality lies at the heart of our mental images and our consideration of inequality and the segregation of individuals and groups in a given city is therefore considered an important aspect of this inequality, and thus social research turns frequently to the study of segregation from this perspective, measuring and comparing cases and forgetting that sometimes there is some beneficial effect of segregation process for the individuals. The internal solidarity that may exist in a segregated neighborhood usually applies to the less privileged. The determining factor in the analysis of the segregation process is the entrenchment 1  This chapter is the partial result of research supported by Spanish Plan Nacional I+d, number CSO2010-22117-C02-01.

218

Residential Segregation in Comparative Perspective

of its effect on individuals especially in terms of delinquency, low education and unemployment, insofar as they reinforce these differentiating attributes of the individuals who settle in a given area, which could be considered as a negative effect, known as the “neighborhood effect.” The grouping together of individuals with lower levels of education or higher unemployment rates may reinforce these levels. Conversely, gated communities or neighborhoods where higher status social groups are concentrated contribute to the improvement of their residents, increasing their opportunities and reinforcing inequalities relative to other neighborhoods. The initial assumption is that the social structure is somehow reflected in space. However, both the social structure and its spatial representation have different patterns and modes of articulation, and it is not clear that the social structure—considered as the distance of social groups in terms of income, occupation, education, ethnicity etc.—is immediately reflected in the way different types of households settle in urban space. A direct correlation between social structure and segregation patterns used to be an assumption in urban theory, but empirical research seems to challenge this assumption. The concept of social structure has changed over the last decades, and while this is not the place to discuss this broad concept, we need to state that it refers to differences between groups in terms of occupation (Goldthorpe 1987, Wright 1985). In addition to this, we will examine differences in incomes2 and education, since educational achievement has effects on the valuation of people and on their potential opportunities for a better occupation. Finally, differences in race or ethnic origin are also relevant for our purpose. All this has to be considered within the relative shrinking of the welfare state, with social expenditures falling from 28.7 percent of the GDP in 1993 to 25.2 percent in 2005, even though in absolute terms they increased by 42.3 percent against a higher growth of the GDP (55.6 percent). In recent years, the relationship between social structure and segregation has become a recurring theme in urban sociology (Wacquant 1993, Kempen and Ozuekren 1998, Musterd and Ostendorf 1998, Andersen 2004, Häussermann 2005, Préteceille 2006, 2009) and the analytical methods used have become increasingly refined. Social Space Measurement The distribution in space of the social categories that make up social structures gives places a distinctive social value as they come to be considered in terms of the characteristics of its residents or users. Hence the term social space. There is an assumption that urban space could have different social values and that these social values could even be expressed in different indicators, like average housing price, types of automobiles used by residents or general consumption patterns. Median income of urban areas could also give us an idea of this social space. In this chapter we try to build a measure of the social value of urban space, mainly based on the occupation of residents, without forgetting other components like income or education. This yields a “social score” that we shall call the social value of space which referes to the composition of the socioeconomic categories of the residents in a given area. 2  It would be better to take into account differences in income which are more accurate in the measure of inequalities, but the relevant and dependable information is harder to access. To compensate for this we also use household expenditure data which are considered more trustworthy than data on income. In southern European societies individuals are more reluctant to reveal their real income due to the greater role of shadow economy and fear of consequences.

The Limits of Segregation as an Expression of Socioeconomic Inequality

219

The Space of Madrid Since the walls that enclosed cities were dismantled, urban space delimitation became a problem because it was difficult to draw a line between the urban and the rural or to establish the real limits of an urban region, even more so after the great sprawl of Spanish cities in the last 10 years. Here we consider three spatial levels. First is the Madrid urban region comprising the Province of Madrid, whose official population in 2010 was 6,458,684 people living in 8,028 km2. Within this area there are 179 municipalities divided into 4,194 census tracts. Within the urban region we find the Metropolitan Area whose limits were defined in the late 1960s. It comprises 23 municipalities and 5,529,299 inhabitants and will be used to analyze the segregation process. Finally in the center of the Metropolitan Area is the municipality of Madrid, with 3,273,049 people, half of the regional population. The region has grown considerably in recent years because of the great economic growth from the mid-1990s to the economic crisis in 2008. During this period, several waves of immigrants arrived, mainly from Latin America, North Africa and Romania, constituting approximately 17 percent of the total population and contributing to changes in the social pattern and spatial structure of the region. There are difficulties in comparing the spatial units comprising this metropolitan region, especially when dealing with different time periods due to the changing boundaries of census tracts which are required to have a fixed population size. These difficulties compelled us to proceed to some spatial regroupings in order to make comparison. From the Production System to Social Change As has occurred in other cities in the past, the transformation from an industrial to a post-industrial society has given rise to a transformation of social structures and social space. According to Lawless et al. (1998), employment change is the fundamental force driving urban socioeconomic transformation, as has happened in Madrid after the economic expansion which brought about changes in the occupational structure. In the two decades from the end of the 1980s to the end of the first decade of the twenty-first century, Madrid has experienced a considerable change in its economic structure, being highly impacted by globalization with the decline of its industrial base and the growth of areas linked to new technologies, financial activities, insurance, communication, transport and commerce (Observatorio Metropolitano 2007). With only 12 percent of the Spanish population, the region is home to 33 percent of Spanish information technology companies. It is here where the major Spanish telecommunications and IT companies are located, as well as companies offering ancillary products and services to globalized sectors, such as logistics and those supporting production (legal, communication, design, financial management and accounting, among others). Madrid has thus become an attractive region for companies characterized by innovation, a highly qualified workforce and high productivity (Iglesias 2005). This change resulted in a growing number of executives and managers and in the proportion of professionals and technicians (Table 10.1). Concomitant with this growth is industrial decline, with the departure of many industrial activities to other regions or other countries, except for construction which grew enormously until 2007, the year before the economic crisis, developing a powerful system of infrastructures (highways, metro and railways lines, energy supplies etc.) and an unprecedented level of production: more than 70,000 housing units were built per year in the region (more than in London, Paris, and

220

Residential Segregation in Comparative Perspective

Berlin together). In 1995 the construction sector accounted for 6.6 percent of the region’s GDP, rising to 9.3 percent in 2005, and for 8 percent and 11 percent of the total labor force respectively. Growth in this sector stopped and from 2008 this trapped many foreign workers, who represented 47 percent of its labor force in 2010 (Encuesta de Población Activa 1995, 2005, 2010). The fall led to a financial crisis due to the indebtedness of the promoters who had bought land to continue profiting from the growth trend. Table 10.1

Proportional distribution of Socioeconomic Categories (CSE) in Madrid Region 1996–2010

Socioeconomic Categories (CSE) Agricultural Workers Self-employed professionals and technicians Business owners with employees Business owners without employees Members of non-agricultural cooperatives Directors and managers Salaried Professionals and technicians Public sector workers Other administrative and sales workers Other service sector workers Foremen Skilled and specialized workers Unskilled workers Military Unclassifiable Total

Percentage and change 1996

2010

1996–10

0.6 1.7 3.0 4.8 0.3 2.5 20.0 0.2 24.5 17.5 0.8 16.9 4.9 0.9 1.4

0.1 3.3 3.4 4.5 0.0 3.8 28.7 1.0 20.2 20.2 0.4 11.8 1.9 0.5 0.2

-0.5 1.6 0.3 -0.3 -0.2 1.2 8.8 0.8 -4.3 2.7 -0.4 -5.2 -3.0 -0.4 -1.2

100.0

100.0

Source: Encuesta de Población Activa (Labor Force Survey) 1996, 2010.

Unfortunately the categories used by the Spanish Statistical Institute differ from those of other European countries. The Spanish categories are based on three principal variables: occupation, sector of activity (rural, industrial, or services) and employment status (salaried, employer, independent worker). Where data on all three variables is missing, we rely on occupational categories. The social structure has undergone significant changes with the great expansion of middle and upper categories (managers and professionals) and service workers (mainly personal services workers) while lower categories, such as skilled and unskilled workers, decreased. It is the general upward mobility in the social structure that is affecting the pattern of its distribution in urban space. These changes in the distribution of social categories can be contrasted with the changes in income distribution between different groups. It can be expected that changes producing a greater inequality in the social structure will be related to similar changes in terms of income or salary distribution.

The Limits of Segregation as an Expression of Socioeconomic Inequality

221

The proportional decrease of social expenditure was compensated by higher capital gains and rising property values that both more than doubled between 1995 and 2006 and contributed to the growth of the GDP. The distribution of this growth was unequal with average pensions increasing by 18 percent, while salaries decreased by 2.4 percent. Despite the overall improvement in social conditions, inequality in terms of salaries remains evident and there is an increasing distance between the highest-income strata and the other categories, even without taking in account the impact of rising capital gains and property values. This change has nonetheless produced a decrease in the numbers of those with lower salaries, supporting Hamnett’s (2003) assertion that an increasing inequality in income is not necessarily accompanied by polarization. Directors and managers are the socioeconomic categories whose salaries have increased the most; at the same time the lower end of the structure also increased its salaries, due to the scarcity of skilled workers in a period of great economic expansion, and the increase in social expenditure. In contrast, the salaries of professionals remained practically stable between 1995 and 2006 lagging far behind average salary growth (Table 10.2). This can be explained by the significant increase in graduates, which led to their devaluation as a social category. This means that the gap between the middle and upper classes increased considerably in terms of salaries and also, probably, in terms of prestige, since being a professional nowadays does not carry the same distinctive value that it carried a few decades ago. The economic and social value accorded to this broad category of professionals is certainly conditioned by gender, age and nationality and one must take into account both its great expansion and the differences between its sub-divisions in order to understand its devaluation. Table 10.2

Percentage salary increases by National Occupation Categories (CON) in Madrid (1995–2006)

National Occupation Categories

%

Government Officials and Managers (over 10 employees) Professions associated with high university degrees Professions associated with lower-level University Degrees Technicians and supporting professions Administrative employees Hospitality sector workers and personal care services Sales assistants and similar workers Skilled construction sector workers Skilled workers in industries, metallurgy, machine manufacturing Skilled workers in graphic arts, textile, food preparation, carpenters, Industrial workers, machine operators, assembly workers and installers Drivers Unskilled workers in service sector Peasants, fishermen, construction, manufacturing and transport workers All occupations

36.7 0.4 1.3 15.7 10.1 33.1 12.9 31.2 17.7 27.9 14.1 20.8 11.1 35.8 14.7

Source: Salary survey 1995 and 2006.3 3  The social classification used by salary survey is not the Socioeconomic Categories (CSE) but the National Occupation Categories (CON); this limits our analysis.

222

Residential Segregation in Comparative Perspective

In much the same way, analysis of changes in salaries shows that between 1991 and 2005 the salaries of the top and the bottom groups increased the most (Table 10.3), while those of the middle groups increased the least. Moreover, the importance of wealth in recent years has overtaken that of income (Colectivo IOÉ Study 2008), reinforcing the growing gap between social and income groups. The increase in housing prices has led to an even greater gap between homeowners and tenants. In Madrid, 82 percent of households are homeowners and more than 20 percent own a second home. But there is also an increasing gap between households living in a house that is owned outright and those with a mortgage. One factor in the size of this gap is the average length of mortgages that rose from 10 years to 27 years between 1991 and 2008. Another is the differences in demographic profile between outright owners and mortgage payers. For Spain, from 1994 to 2006: although yearly income growth rate has remained quite steady at 3.7%, wealth grew twice as quickly (8% annual average), with a smaller gap being observed in the first few years of the study and great differences in later years, coinciding with the increase in housing market prices … In short, income from real estate and business-finance has grown at a higher rate than income from salaries, pensions and unemployment benefits, so much so that the material conditions of households depending solely on the latter sources of income are, in relative terms, worse off today than in the past. (Colectivo IOÉ Study 2008)

As a consequence, “the unequal distribution of wealth has risen greatly in recent years: the 75 percent richest households increased their wealth (€215,400 average, in constant currency) 77 times more than the poorest 25 percent households (€2,800)” (Colectivo IOÉ Study 2008). Therefore what is observed is that differences in wealth, which are strongly related to house ownership, are increasingly relevant in understanding inequality and the differences among social categories. Consequently, indices of inequality are increasing and in the case of Madrid the growing numbers of households paying a mortgage for their newly acquired houses has an impact on the way income inequality is evolving, when we estimate it by the spending capacity of households. The increase in the length and the value of mortgages contributes to an underestimation of the purchasing power of some groups, because mortgage payments are considered an investment and not as current expenditure. This especially affects occupational categories such as professionals, with many young people becoming homeowners, restricting their ability to spend. Incomes, salaries and spending for all social categories have improved as a result of the economic transformation, at least until the economic crisis, but at the same time there has been an increase in inequality in spending between 1990 and 2005 expressed by different indices that measure income inequality (Table 10.3). The increasing income inequality cannot be understood unless we do consider differences in education and age. New generations are much better educated than before and the proportion of university graduates is much greater than in previous generations. This explains both the increase in immigrants who take the jobs rejected by these new upwardly mobile generations as well as the stagnation of their salaries in relation to older generations of highly educated people. The difference in salary between young professionals and young skilled workers has been reduced, as well as differences in education and social prestige. In fact, the reason behind the increase of salaried professionals is the influx of young graduates into the labor market, indicating considerable intergenerational social mobility. In terms of urban distribution, these fragile new upper-middle strata are important as they feed the generalized tendency in southern European countries, for young professionals to locate near parents and friends. Most of the new real estate developed in the

The Limits of Segregation as an Expression of Socioeconomic Inequality

Table 10.3

223

Income inequality indices based on average household expenditure in Madrid (1990/91–2005) Inequality indices*

1991 2005

Standard deviation

Gini coefficient

Theil index

Atkinson index

0.21 0.22

0.29 0.31

0.16 0.17

0.13 0.14

Note: * The Gini is the most commonly used indicator to measure economic inequality and it is based on the Deviations model. It quantifies the degree of inequality graphically represented by the Lorenz curve; higher values indicate more inequality. The Theil is an index based on the Entropy model; it is a way to establish what percentage of the inequality corresponds to the difference between groups, by calculating the difference between maximum emtropy (perfect equality) and the actually measured entropy. The Atkinson index is based on the Social Welfare model and is defined as the ratio between the “equally distributed income” (income means to be shared by all people creates a comfort level similar to the present) and the average income of the economy. The inequality index is simply one minus the ratio and takes values between 0 and 1. A small value of the Atkinson index implies a more equal distribution. The Atkinson index (I) measures the fraction of income that can be sacrificed without loss of social welfare if income were distributed equally. For a detailed description of inequality indices, see Coulter (1989). Source: Mas et al. (2006).

late 1980s and 1990s near Madrid’s industrial localities in the South and the East, contains some high quality housing which was a viable option for the new middle-class generations aspiring to live in better conditions than their parents but not far from them. This pattern leads us to question the theoretical redistribution of urban space according to class (Map 10.1). Our argument is that segregation patterns can vary in different ways because the complex interrelation of social position, income and education, as well as of demographic features, may bring about a kind of disturbance to the expected correlation between social inequality and spatial segregation. Segregation, Social Categories and Social Space The socioeconomic changes described above have had a great impact on the urban landscape in terms of both residential space and public space. Madrid has witnessed a redefinition of its social space in recent years as a result of the dizzying economic growth that took place between 2000 and 2007. Segregation has persisted as one of the spatial manifestations of social inequality. It is generally thought that a greater social distance entails a greater spatial distance measured in residential terms, or put differently, that the greater the distance between social groups, the less likely it is that they will be found living in the same city area (or neighborhood). But empirical evidence has shown that there is no direct relationship between these two variables.4 In recent years social inequality has increased in many European cities, but this has not always been accompanied by an increase 4  Sabatini and Brain (2008) mention various studies that reject the direct correlation between these phenomena (social and spatial). One such study by Telles (1992) compares major cities of the United States

224

Residential Segregation in Comparative Perspective

in the level of social segregation (Leal 2007). Additionally, studies of residential segregation among immigrants in the cities of Spain and southern Europe (Arbaci 2008, Martínez del Olmo and Leal 2008, Musterd and Fullaondo 2008, Maloutas 2007) conclude that segregation is not a good indicator of inequality for immigrants. Recent trends in Spain have shown a decrease in segregation between natives and immigrants without evidence of a parallel decrease in income inequality. Hence, there may be other residential processes at work such as social exclusion linked to residential exclusion (Cortés 2008) or overcrowding, which reveal situations of inequality that segregation indicators conceal. The spatial manifestation of social inequality is more complex than a mere association with residential distance between social groups. Maloutas’ argument on vertical segregation as a spatial expression of these differences which mitigates horizontal segregation (Maloutas and Karadimitriou 2001) or Martínez del Olmo and Leal’s (2008) studies on sub-standard housing and overcrowding, show that the “spatial” can be understood at many different scales. Moreover, as White (1983) and Chamboredon and Lemaire (1970) suggest, people of varied social classes living in the same neighborhood may not display any meaningful interaction. Nonetheless, interest in the study of residential segregation results from the social mix present in a neighborhood, which supposedly favors interaction and the formation of social networks between different social groups. Living in proximity facilitates contact and interaction. Studies of neighborhood effects have produced uneven results, but some have shown that a high concentration and segregation of certain groups in a given space leads to worse conditions in social terms. Moreover, the concentration and self-isolation of advantaged groups in urban areas can generate positive consequences, as argued in a paper related to neighborhood effects and education by Gordon and Monastiriotis (2006). Less segregation does not in itself ensure increased interaction and it is advisable to examine underlying socio-spatial changes. In Madrid it is clear that increased social inequality does not always go hand in hand with increased residential segregation. To show this we analyze some of the key processes that underlie changes in segregation and the social restructuring of the Madrid Metropolitan Area, which is closely related to socioeconomic and urban processes. The first problem in the analysis of segregation of social categories was how to operationalize them to form groups with a real identity, i.e. whose members share common values and interests and which are different from members of other groups. In this chapter we use several measures. First, for a general view of social segregation we use the variable Socioeconomic Condition from official statistics. Second, we consider segregation related to education levels. Third, we consider the segregation of minorities by reference to their foreign origin and we draw comparative insight from other countries that have been receiving immigration flows for a long time. There are not many variables available at the census tract level to measure the segregation process. Here we concentrate on two of them: occupational categories and immigrant status. Among the variables available those that best suit the classification are based on socioeconomic conditions (CSE in its Spanish abbreviation) which provide information both on professional categories and occupational status. In some cases we had to use the Spanish Classification of Occupations (CNO in its Spanish abbreviation, similar to ISCO categories) to complete this information.

and Brazil and shows that although cities in the United States are more segregated in residential terms based on indices of dissimilarity, Brazilian cities show greater economic inequality.

The Limits of Segregation as an Expression of Socioeconomic Inequality

225

The Impact of the Housing Market Some other issues are related to demographic variations. Often there is not a great difference between the residential behavior of a married couple with children where the main income earner is a manual worker, and a household whose income earner is a young professional at the beginning of his career with a precarious job contract. On the other hand, two similar managers may express different residential preferences if one lives alone and the other lives in a family with children. The position of “professional” workers has changed over the years due to the increasing precariousness of their working conditions, as experienced by many young people today. Moreover, in terms of status and income this category is becoming increasingly differentiated internally. Furthermore, the changes in household living patterns affect all socioeconomic categories, increasing the diversity of residential strategies within each category. The increase in middleaged single-person households, single-parent households and even multi-person households and the decline in the absolute dominance of nuclear family households have induced socio-spatial processes such as the pressure on central areas of the city by middle class single-person households and couples without children, often leading to gentrification. The type of available housing in a given area of the city does not always suit different household types. The spatial distribution of housing types within a city greatly affects its social map. Larger houses, preferred by nuclear families rather than by single-person households, are usually found in peripheral areas, whereas smaller houses tend to be located in central areas. In addition to the size of the house, the amenities offered by each area and the environmental setting also condition housing choices. Therefore, the type of housing and amenities of a given area will influence housing choices, and impact on the shape of the city’s social space. However, beyond these changes in residential strategies, the resulting spatial distribution of socioeconomic groups in Madrid Metropolitan Region is quite clear. In fact, we can draw a line dividing the north-west region from the southeast, with the former containing areas of higher socioeconomic status and the latter containing lower-status neighborhoods. To obtain a clear map representing the distribution of social categories in urban space it is important to assign a unitary value to the small areas comprising the urban space. This unitary value was calculated using a factorial analysis (Principal Components Analysis—PCA) on the proportion of the principal categories of the Socioeconomic Condition in each census tract of Madrid Metropolitan Area. We used 12 categories related to social position as variables for the PCA (the others were excluded due to their minor social importance). The resulting factors—principal components in this case—that are a synthesis of the initial variables, account for a substantial part of the information (variance) contained in the initial dataset, with the first factor explaining 42 percent of the total variance. This first factor is significant not only for the percentage of variance explained, but also for the clarity of its meaning which is derived from its correlation with the initial variables (Table 10.4). To understand better the real meaning of this first factor, Table 10.4 shows its correlation coefficients with the initial variables. It is clear that there is a relation between the value of the correlation coefficient of each category with the first factor and the social status of the occupational categories. At one extreme of the factorial axis we find the professionals and at the other skilled and unskilled workers, while employees in commerce are situated in the middle. The hierarchy in this table does not follow exactly the expected social hierarchy (for example, skilled workers score lower than the unskilled) due to the fact that these scores are determined by the spatial distribution of categories rather than directly by their social status. The higher positive coefficient scores correspond to the higher social categories and the higher negative scores correspond to the

Residential Segregation in Comparative Perspective

226

Table 10.4

Correlation coefficients of socioeconomic categories with the first factor of a principal components analysis of their distribution in Madrid’s census tracts (2001)

Socioeconomic categories Professionals and technicians salaried Directors and Managers Self-employed professionals and technicians Business owners with workers Professionals from public sector Other sales workers Members of cooperatives Business owners without workers Other service sector workers Foremen Unskilled workers Skilled workers

Correlation with the first factor 0.944 0.871 0.857 0.521 0.391 0.021 -0.267 -0.466 -0.540 -0.614 -0.751 -0.894

Source: Compiled by the authors from 2001 Spanish Census (Instituto Nacional de Estadística).

lower social categories; intermediate categories have lower correlation scores either positive or negative. This means that the first factor represents the city’s systematic social division: different social categories with high positive scores have similar distribution patterns across the city’s census tracts, while categories with high scores but placed on opposite sides (positive or negative) tend to have dissimilar distributions. Categories with lower correlation scores have distributions that are less similar to the distribution pattern defined by this factor. Hence this factor may be taken as a measure of the social “value” of spatial entities. Depending on its social composition, each census tract comes with a correspondence coordinate on the first factor, calculated by the PCA. This means that all census tracts can be projected on the line determined by this factor, creating a hierarchy based on the scores of their coordinates. These coordinates on the first factor for each census tract are mapped for Madrid Metropolitan Area, in a shade ranging from black for the highest positive values to white for the highest negative values. Map 10.1 represents the synthetic pattern of social differences in the form of the unequal distribution of social groups in urban space. The shape of this pattern suggests that there is a strong correlation between this socio-spatial division and other variables like housing prices or household income. Map 10.1 represents the values (coordinates) of census tracts of the Metropolitan Region on the first factor of the Principal Components Analysis (distributed according with their standard deviation from the middle value) and shows what we call the “social value of space” at census tract level. Differences derive from data on occupational categories, but they could also be explained in terms of how the metropolitan area has expanded, the location of better services and jobs, the quality of the environment, the location of services (like hospitals and universities) as well as to other variables. The most striking feature of the socio-spatial structure of the city is the line differentiating the north and west (in dark shades), where the middle and upper socioeconomic categories are located, from the south and east (in lighter shades) where most of the lower-middle socioeconomic categories and workers live, more distant from better jobs and services, with a few exceptions. This demarcation gives rise to two different cities in respect to their social components,

The Limits of Segregation as an Expression of Socioeconomic Inequality

Map 10.1

227

Socio-spatial inequality in Madrid Metropolitan Area: Census tracts classified according to their distance from the mean socioeconomic profile (2001)*

Note: * Distances are derived from the coordinates of census tracts on the first factor of a Principal Components Analysis on the spatial distribution of occupational categories in Madrid that summarizes socioeconomic distance. Source: Compiled by the authors 2001 Spanish Census (Instituto Nacional de Estadística).

while in the city center there are some deviations from this dichotomy which have developed slowly in the last half century. In the 1950s and 1960s, the middle classes settled in the city center or in their immediate surroundings, while the working class lived in the periphery where they lacked services and transport facilities. In the following years the pattern changed with the improvement of service distribution and transport infrastructures, and the current differentiation between north and south in the Metropolitan area became stronger. Changes were rather slow and the first settlements of workers in the north of the city took their time to become, little by little, more bourgeois. An important question concerns the evolution of segregation patterns in recent years. Responding to this question is complicated due to changes in the content of some socioeconomic categories between 1991 and 2001 and changes in census tract boundaries. To explore the question of the decrease of segregation in Madrid during the 1990s, indices of segregation (IS) were calculated for each of the socioeconomic categories based on Socioeconomic Conditions for 1991 and 2001, using data at census tract level. As Table 10.5 shows, the result was that for most categories segregation had decreased, but most importantly, the greater decrease was observed for the higher and higher-middle categories, even though the three highest categories continue to be

228

Residential Segregation in Comparative Perspective

among the most segregated. We should stress that two of the categories which maintained the same level of segregation in this period, other service sector workers and skilled workers, included 47 percent and 19 percent of immigrants respectively (Encuesta de Población Activa 2010). Table 10.5

Segregation indices for socioeconomic categories in Madrid Metropolitan Area (1991 and 2001)5

Categories of the socioeconomic condition (CSE) Directors, managers and high public sector workers Self-employed professionals and technicians Salaried professionals and technicians Other administrative and sales workers Unskilled workers Business owners with employees Business owners without employees Skilled workers Other service sector workers Foremen Members of cooperatives

Segregation index 1991

2001

Difference

0.44 0.43 0.30 0.15 0.33 0.21 0.16 0.31 0.17 0.23 0.31

0.34 0.35 0.25 0.10 0.28 0.20 0.16 0.31 0.17 0.25 0.48

-0.10 -0.08 -0.06 -0.05 -0.05 -0.01 0.00 0.00 0.00 0.02 0.17

Source: Compiled by the authors from 1991 and 2001 Spanish Census (Instituto Nacional de Estadística).

The lower degree of segregation observed for business owners without employees and for administrative and sales workers is in part due to the social heterogeneity of these categories and their position in the middle of the social scale. In the case of business owners with employees, this varies greatly with respect to the latter’s number. Despite the decrease in segregation in the Madrid Metropolitan Area, one of the problems observed in these indices is that they are highly influenced by the changing numbers in each category which distorts comparisons between the different categories.6 Once the effect of the change in the population size of these categories was corrected, there were just two categories which exhibited an increase in segregation: skilled workers and other administrative and sales workers, but for these two cases the change induced by the variation in population size was small. These results are consistent with recent socioeconomic patterns and the assumed relationship between inequality and segregation. The groups that are most distant from the rest and which show the highest levels of residential segregation are those with the highest socioeconomic ranking. 5  The agricultural sector and the military professional categories were excluded from the analysis as they represent minor groups in terms of numbers. The category of professionals exclusively in the public sector is not comparable to that created for 2001 which includes public sector managers, sales representatives, and those working in company services; therefore, this category was eliminated as well. 6  Echazarra (2010) analyses the effects of changing population size on indices of segregation for different immigrant groups and shows that for smaller groups, the growth in population causes the index of dissimilarity to drop considerably. In consequence we recalculated the segregation index following the method employed by this author but in our case the result of this new analysis was not significantly different from the first one.

The Limits of Segregation as an Expression of Socioeconomic Inequality

229

However, while there has been a growing distance in social and economic terms, in spatial terms the predominant spaces of these groups have experienced some increase in social mixing. It is possible that the decrease in segregation of directors, managers and high public sector workers is the result of a changing pattern linked to their expansion throughout the period in question. Starting from a relatively small size and seeking alternatives for their expansion implies their settlement in different places where they mixed with other groups, mainly belonging to the upper-middle categories. In terms of space, these categories expanded from their initial central position and invaded the surrounding areas. Urban planning has influenced this process due to the increasing housing density in these wealthy central areas, but also due to the policies of certain local governments to attract such high-income households in order to increase the area’s social value, to support the local budget, and to attract political supporters. They planned new types of urban settlement able to attract high socioeconomic categories by developing single family housing, private schools and other private services (golf courses, shopping centers, etc.). The Settlement Patterns of Higher-status Professionals There are two segregation-related issues that make the higher-income and higher-ranking professional groups noteworthy: these groups are the ones whose increasing purchasing power accounts for increasing inequality, while at the same time they exhibit the highest level of segregation. However, they also underwent many changes in terms of segregation patterns, as they mixed with other groups in increasing numbers and over large areas.7 Census tracts are spatial divisions based on population size. Therefore, census units located in the city center suit the purposes of the analysis quite well since they do not change much in population size, due to the difficulties in increasing the number of dwellings in built-up areas. In peripheral places or less-dense metropolitan places, however, the spatial size of census units is larger and their social composition could be more complex as a consequence, made up of groups that do not really share living spaces due to spatial distance. Indices of segregation could be lower, but in this case the interaction between groups may not be as marked as in the smaller-sized central census units. At any rate, if the population of a given social category increases significantly it will saturate certain areas and will invariably invade areas in which it previously had little or no presence; these areas usually represent newly-developed residential districts or spaces undergoing socioeconomic transformation, largely driven by strategies of location which are specific to certain types of economic activity, facility and amenity. This expansion is accompanied at the outset by lower indices of segregation Maps 10.2 and 10.3 show not only how the weight of “directors and managers” has grown significantly, but also that it has invaded places which had not previously been inhabited by these groups, such as Madrid’s northern districts that underwent considerable business and residential development during the 1991–2001 period. These groups have also often settled in central districts that they previously did not occupy, in some cases through gentrification, as mentioned earlier. This gentrification process is driven by new offers on the housing market in places where there had often been new investment in public spaces, and at the same time it is a response to cultural

7  Préteceille (2003) concludes that in the case of Paris segregation usually comes from above, from the highest categories on the social scale.

230

Map 10.2

Residential Segregation in Comparative Perspective

Proportional distribution of managers and high civil servants in Madrid Metropolitan Area (1991)

Source: Compiled by the authors from 1991 Spanish Census (Instituto Nacional de Estadística).

and lifestyle changes in these higher categories that have been accompanied by a greater variety of residential location strategies. The increased residential mixing of professional elites does not automatically mean, as was explained above, an increase in social interaction. They invaded certain spaces, transforming them while sharing space with other social groups. But, these processes set off other mechanisms of exclusion and segregation, which favor the social closing-off of these classes, which cannot be measured using indicators of segregation or dissimilarity. In the majority of cases, segregation arises as a result of using exclusive and private services that are marketed as different social networks and urban living patterns. Analyzing the mechanisms that maintain or even increase a closing-off of classes within this context would help one to better understand the relationship established between the aforementioned social processes and space. On the other hand, we can also consider two processes that influence the decrease of segregation in low-income census tracts. First, the intergenerational social mobility, due to the general increase of academic achievement, mentioned before, which means that workers’ children become service workers, technicians or even professionals, is clearly observed in Madrid, but less clear in mediumsized cities. Second, the typical strategy for new households in southern European countries to settle near their family, due to the scarcity of public services, as described in Allen et al. (2004) and Maloutas (2004). This contributes to developing new social patterns in the old working class

The Limits of Segregation as an Expression of Socioeconomic Inequality

Map 10.3

231

Proportional distribution of managers and high civil servants in Madrid Metropolitan Area (2001)

Source: Compiled by the authors from 2001 Spanish Census (Instituto Nacional de Estadística).

peripheries and their surroundings with new real estate developments with better housing and a different social image adapted to new middle class households. Foreign Immigrants Madrid’s population growth is due mostly to the large influx of immigrants experienced between 1996 and 2007. During that period, the percentage of foreigners in the region’s total population increased from 1.5 to 17 percent. Almost 900,000 people, most of them from developing countries (North Africa, South America, and Eastern Europe) came in waves to work mainly in services and construction. Their location in the city changed over the years, beginning in the most deteriorated places of the city center and expanding to the periphery, following the “doughnut pattern” described by Arbaci (2008) for southern European cities. Settlement patterns of immigrant groups have transformed the social landscape of some areas of the city, affecting the residential strategies of other households. Furthermore, improved social conditions and higher levels of education have meant that greater segments of the Spanish-born population would forgo occupations in these high-demand service positions that were subsequently filled by foreign workers (Cachón 2002). The result was

Residential Segregation in Comparative Perspective

232

a parallel demand for occupations characterized by lower levels of specialization and credentials which contributed to increased internal differentiation of a category already marked by great heterogeneity. As stated by Garrido and Toharia (2004), the majority of immigrants in Spain work in low-skilled occupations, such as those associated with domestic and cleaning services, in construction and tourism, while 40 percent work in other services: 23 percent as skilled workers and 13 percent in administrative and sales occupations,8 which places them at the bottom of the socioeconomic scale. Any analysis of the residential behavior of these lower social categories and their levels of segregation would be quite marked by the behavior of immigrant groups. There are no data to test the effect of immigration in terms of income or social categories from 2001 in Madrid, but Table 10.6 shows some decrease in the segregation of immigrants in terms of dissimilarity indices, as happened with most social categories some years ago. This change has different dimensions for each nationality, as Latino-Americans are less segregated than Africans. The first immigrants of a group were often concentrated in deteriorated districts in central areas and in some places in the rich periphery, whereas later arrivals are located on the borders of working class districts. In the late 1990s and early 2000s, households composed of foreigners from developing countries were concentrated in places where there was a high proportion of rental housing, but in recent years their increased access to homeownership in the lower price localities of the periphery has introduced new patterns in their location patterns. Many of the foreign homeowners rent one or two rooms of their house to newcomers in order to pay their monthly mortgage. So, whether it is in the newly rented homes or in the purchased apartments on the periphery, overcrowding is the general practice, at is made clear by the fact that immigrants’ houses are in worse condition and, despite being smaller than average, contain larger household sizes (Martínez del Olmo and Leal 2008). Table 10.6

Segregation index by immigrant origin in Madrid Metropolitan Area (2001 and 2009) Segregation index for foreigners

Year

Total foreigners

Europeans (east and west)

Americans

Africans (north and south)

Asians

Oceania (Australians and NZ)

2001 2009

28.54 26.22

32.06 27.76

36.08 31.02

41.37 42.91

48.71 43.75

87.30 85.22

Source: Compiled by the authors from Spanish Census for 2001; Padron9 for 2009.

The residential strategy of immigrants tends to rely on overcrowding: their average household size in 2007 in Madrid was 3.9—greater than the figures for the whole Madrid population (2.8) and also for foreigners living in Spain (3.4)10—with a greater proportion of multifamily and 8  Instituto Nacional de Estadística, Encuesta de Población Activa (Labor survey) 2007. 9  “Padron” is the official annual register of population made by local authorities in Spain. A kind of census for small areas with only three variables: age, sex and nationality. 10  The data for foreigners originate from Encuesta Nacional de Inmigración (National immigrant survey) by the National Institute of Statistics (2007), and from Encuesta de Población Activa (Labor survey) (2007) for the Spanish-born.

The Limits of Segregation as an Expression of Socioeconomic Inequality

233

Figure 10.1 Distribution of Madrid’s total and foreign population by household size (2007) Source: Encuesta Nacional de Inmigrantes, 2007.

multinuclear households composed by foreigners. The distribution of the foreign and Spanishborn populations by household size reveals important differences (see Figure 10.1) even though household size in Spain is among the highest in Europe. A more detailed analysis based on interviews about the living conditions of the foreign population shows that spatial proximity between foreigners and nationals does not always imply greater communication and interaction between them.11 The main question for many foreign newcomers to Madrid is not where to live, but where to find housing for rent, due to the housing shortage. According to the National Immigrant survey, the percentage of foreign households renting their houses in the Madrid Region is 50 percent, while the average for all households is 13 percent. The concentration of housing for rent explains the spatial distribution of immigrant newcomers, regardless of whether these houses are in high social value neighborhoods, where they will rent apartments in poor condition (lack of light, damp, etc.) with low rents and will then overcrowd them in order to be able to pay the monthly rent. Another common strategy is to buy an apartment on the periphery and to share it with others, as mentioned before, subletting a room in order to be able to pay the mortgage. Overcrowding is a spatial expression of inequality, but quite different from segregation. Moreover, overcrowding may have contributed to the decrease of segregation in recent years. These housing practices of the foreign population explain why the areas with the lowest proportion of foreigners are in fact some of the old working class estates in the near periphery of the city. In these areas there is low residential mobility due to widespread homeownership, to the aging population and to the high cost of moving (high taxes and high prices of housing in general).

11  This is detailed in the forthcoming publication: Leal and Alguacil (2012).

234

Residential Segregation in Comparative Perspective

Conclusion From 1991 to 2007, Madrid experienced rapid social change linked to fast economic growth. The subsequent financial crisis led to a significant decrease in economic activity, which was felt in Madrid before other European capitals, due to the fall in housing demand. The rapid social change in the late 1990s and the early 2000s was characterized by a considerable growth of diversity due to the great influx of immigrants, by the huge increase of the upper-middle categories of professionals and technicians and, in particular, by the large increase in the size and income of higher occupational categories, mainly of directors and managers. These changes have resulted in increased inequality among socioeconomic categories in terms of income, accompanied, however, by a slight increase in the spatial mixing of different occupational categories. There is need for a more detailed analysis of spatial inequality associated with the comparative advantages and disadvantages of the city’s residential areas in terms of access to certain types of employment and services, in order to locate and evaluate other patterns of spatial inequality that may coexist with the decreasing segregation indices. In this chapter, we have tried to demonstrate the lack of a direct correlation between residential segregation and income and/or class inequality. Residential segregation may be generating or reproducing inequality due to the advantages and opportunities offered by certain spaces. Hence, the uneven access to certain services or amenities from different residential areas may give rise to new inequalities or perpetuate existing ones. However, a greater socioeconomic distance does not necessarily imply greater spatial distance; hence our argument that growing polarization or economic inequality is not necessarily accompanied by increased segregation. Processes that have brought about social mixing in recent years come hand in hand with the growth in numbers of the more segregated groups, i.e. immigrants and high-status professionals. These growing groups have saturated their original residential areas and invaded others as they expanded, leading to reduced values of segregation index. These processes have been favored by certain changes in ways of life as well as in the housing market. The diversification of household patterns and lifestyles gives rise to diverse residential preferences within social categories. Moreover, occupational categories have become more heterogeneous in terms of working conditions due to increased employment precariousness and flexibility and, therefore, residential options may also vary according to those conditions. Furthermore, rehabilitation and the construction of high quality housing into areas traditionally inhabited by the working class or which were simply of lesser social value, have favored the influx of households of higher socioeconomic level, but also enabled upwardly mobile households originating from the same working class areas to remain near their parental households and preserve family networks based on spatial proximity. The decrease in segregation is, partly at least, related to this increasing complexity in residential preferences. Strategies like overcrowding apartments even in wealthy areas or living in deteriorated houses in the city center as the only way to get shelter, result however in housing deprivation; and decreasing segregation does not provide protection from such parallel processes that deepen inequality. References Allen, J., Barlow, J., Leal, J., et al. 2004. Housing and Welfare in Southern Europe. Oxford: Blackwell Science.

The Limits of Segregation as an Expression of Socioeconomic Inequality

235

Andersen, H.T. 2004. “Spatial – Not Social Polarization: Social Change and Segregation in Copenhagen.” The Greek Review of Social Research, 113 A, 145–65. Arbaci, S. 2008. “(Re)Viewing Ethnic Residential Segregation in Southern European Cities: Housing and Urban Regimes as Mechanisms of Marginalisation.” Housing Studies 23(4), 589– 613. Cachón, L. 2002. “La formación de la España inmigrante: mercado y ciudadanía.” Revista Espanola de Investigaciones Sociologicas 97, 95–126. Chamboredon, J. and Lemaire, M. 1970. “Proximité spatiale et distance sociale: Les grands ensembles et leur peuplement.” Revue Francaise de Sociologie 1, 3–33. Colectivo IOÉ Study 2008. Barómetro Social de España: Análisis del periodo 1994–2006 a partir de un sistema de indicadores. Madrid: Traficantes de Sueños. Cortés, L. 2008. Nuevos y viejos problemas residenciales: Vivienda y exclusión. Madrid: Editorial Milenio. Coulter, P.B. 1989. Measuring Inequality. A Methodological Handbook. Boulder, CO: Westview Press. Echazarra, A. 2010. “Segregación residencial de los extranjeros en el área metropolitana de Madrid. Un análisis cuantitativo.” Revista Internacional de Sociologia 68(1), 165–97. Garrido, L. and Toharia, L. 2004. “La situación laboral de los españoles y los extranjeros según la Encuesta de Población Activa.” Revista Economistas 99, 74–88. Golthorpe, J. 1987. Social Mobility and Class Structure in Modern Britain. Oxford: Clarendon Press. Gordon, I. and Monastiriotis, V. 2006. “Urban Size, Spatial Segregation and Inequality in Educational Outcomes.” Urban Studies (Edinburgh, Scotland) 43(1), 213–36. Hamnett, C. 2003. Unequal City: London in the Global Arena. London: Routledge. Häussermann, H. 2005. “The End of the European City?”. European Review (Chichester, England) 13(2), 237–49. Iglesias, F. 2005. Por qué razones el empleo crece más en Madrid que en el resto de España? Madrid: Instituto de Estudios Económicos. Kempen, R. and Ozuekren, A. 1998. “Ethnic Segregation in Cities: New Forms and Explanations in a Dynamic World.” Urban Studies (Edinburgh, Scotland) 35, 1631–56. Lawless P, Martin R. and Hardy S. (eds) 1998. Unemployment and Social Exclusion: Landscapes of Labour Inequality. London: Jessica Kinglsey Publishers. Leal, J., ed. 2007. Vivienda y segregación en las grandes ciudades europeas. Madrid: Ayuntamiento de Madrid. Leal, J. and Alguacil A. 2012. “Vivienda e inmigración: las condiciones y el comportamiento residencial de los inmigrantes en España.” In La integración de la inmigración. Anuario de la inmigración en España (edición 2011), ed. Aja, E., Arango, J. and Oliver, J. Barcelona: Belaterra Spain. Maloutas, T. 2004. “Segregation and Residential Mobility: Spatially Entrapped Social Mobility and its Impact on Segregation in Athens.” European Urban and Regional Studies 11(3), 195–211. Maloutas, T. 2007. “Segregation, Social Polarization and Immigration in Athens during the 1990s: Theoretical Expectations and Contextual Difference.” International Journal of Urban and Regional Research 31(4), 733–58. Maloutas, T. and Karadimitrou, N. 2001. “Vertical Social Differentiation in Athens: Alternative or Complement to Community Segregation?” International Journal of Urban and Regional Research 25(4), 699–716.

236

Residential Segregation in Comparative Perspective

Martínez del Olmo, A. and Leal, M.J. 2008. “La segregación residencial, un indicador espacial confuso en la representación de la problemática residencial de los inmigrantes económicos: El caso de la Comunidad de Madrid.” ACE 8, 53–64. Mas, M., Aldás, J. and Goerlich, F.J. 2006. Gasto de las familias en las comunidades autónomas españolas. Pautas de consumo, desigualdad y convergencia. La Coruña: Fundacion Caixa Galicia. Musterd, S. and Fullaondo, A. 2008. “Segregación étnica y mercado de la vivienda en dos ciudades del norte y sur de Europa: Los casos de Ámsterdam y Barcelona.” ACE 8, 93–114. Musterd, S. and Ostendorf, W. eds. 1998. Urban Segregation and the Welfare State: Inequality and Exclusion in Western Cities. London: Routledge. Observatorio Metropolitano 2007. Madrid, ¿la suma de todos? Globalización, territorio y desigualdad. Madrid: Traficantes de Sueños. Préteceille, E. 2003. La division sociale de l’espace francilien. Typologie socioprofessionnelle 1999 et transformations de l’espace résidentiel 1990–99 [Observatoire Sociologique du Changement FNSP-CNRS]. Available at: http://hal.archives-ouvertes.fr/docs/00/13/02/91/ PDF/DivSocParis1999 Transfo.pdf [accessed 8 May 2011]. Préteceille, E. 2006. “La ségrégation contre la cohésion sociale: La métropole parisienne.” In L’épreuve des inégalités, ed. H. Lagrange. Paris: PUF, 195–246. Préteceille, E. 2009. “La ségrégation ethno-raciale dans la métropole parisienne.” Revue Francaise de Sociologie 50(3), 489–519. Sabatini, F. and Brain, I. 2008. “La segregación, los guetos y la integración social urbana: Mitos y claves. Santiago de Chile.” EURE. Revista Latinoamericana de Estudios Urbano Regionales 34(103), 5–26. Simmel, G. 1950. The Sociology of Georg Simmel. (K. Wolff, compiled and trans.). Glencoe, IL: Free Press. Telles, E.E. 1992. “Residential Segregation by Skin Color in Brazil.” American Sociological Review 57(2), 186–97. Wacquant, L. 1993. “Urban Outcasts: Stigma and Division in the Black American Ghetto and the French Urban Periphery.” International Journal of Urban and Regional Research 17, 366–83. White, M.J. 1983. “The Measurement of Spatial Segregation.” American Journal of Sociology 88, 1008–18. Wirth, L. 1938. “Urbanism as a Way of Life.” American Journal of Sociology 44, 1–24. Wright, E.O. 1985. Classes. London: Verso.

Chapter 11

Changing Dynamics of Residential Segregation in Istanbul Tuna Taşan-Kok

Introduction Istanbul lost its function as an imperial capital city with the collapse of the Ottoman Empire (1923), though it has remained as the financial center of Turkey and has slowly re-connected with the global networks. With continuous migration from the poorer regions of the country and the associated problems—e.g., illegal housing, gray economy—the city has grown into a large metropolis which in recent years has attracted much international investment capital. Focusing on the city’s path dependent development, this chapter presents the changing trends in residential segregation processes in the city, which over time have changed from ethnic to social through economic differences and spatial inequalities. In order to explain this transformation, the main social, economic, political and cultural events that influenced the segregation processes from a historical perspective will be analyzed. Throughout its history Istanbul has always been an important node of global trade, culture, politics and capital flow due to its geopolitical location between Europe and Asia. Its historical roots, which stretch back over 2,600 years, demonstrate the importance of Istanbul for Turkey and for the region (located between the Middle East, Eastern Europe and Russia) as a center of culture and economy. Multiculturalism became a fundamental characteristic of the city, and so it remains today. However, the composition and spatial representation of it changed with the spatially segregated non-Muslim ethnic groups being replaced by people from other religious, ethnic, cultural and socio-economic backgrounds over time. This multi-cultural atmosphere has brought with it changes that have, over time, triggered certain processes of spatial separation between diverse social and ethnic groups which were sometimes a result of explicit policy and planning targeting the physical separation of diverse groups of people; and sometimes due to the residential choices of individuals or groups. As world-city studies (Taylor et al. 2002) have shown, Istanbul is connected to the global economic networks through its service economy and large service employment; flexible labor; trade, capital, information, culture and media functions; large numbers of domestic and international immigrants; advanced producer services, such as banking and advertising; an advanced tourism sector; World-renowned cultural events; and a fairly large population (14 million inhabitants) and cultural diversity. The city’s integration into the global economy has played a crucial role in the residential choices and the quality of life of diverse social groups. First of all, the city has developed the dualistic economic conditions that are driven by the availability of both skilled and unskilled (cheap) human capital and infrastructure facilities for a service economy, as well as production sectors with relatively high technologies and a skilled labor force (Eraydın 2008a). Secondly, the massive migration from less-developed rural areas since the early 1960s has helped shape the informal economy and patron-client types of relations; while foreign investors have been attracted

238

Residential Segregation in Comparative Perspective

to the city’s land and property markets and the availability of scientific and technical labor (Eraydın 2008a). Thirdly, Istanbul demonstrates the importance of property rights regimes for all social groups. Property rights not only became a speculative tool for property developers and investors, but also an income transfer mechanism for the urban underclass, as disadvantaged groups made use of property as a means of raising their status by the redevelopment of squatters. In the absence of a social welfare system that provides all social classes with equal access to housing, Istanbul has witnessed the development of sharp contrasts in land and property market developments, such as gecekondu (squatter) neighborhoods, which exist next to super-rich gated communities; and traditional shopping streets adjacent to luxurious shopping malls. This path-dependent development has helped shape Istanbul’s social and ethnic mobility patterns and residential segregation trends. In the context of Istanbul, residential segregation is defined on the basis of spatialized social and economic inequality (in terms of access to housing, quality of space, and locational advantage) among groups living under different social (employment and education) and economic (income and property ownership) conditions. As Işık and Pınarcıoğlu (2009) also argue, the ties that drove ethnic segregation processes dissolved in time with the major amount of non-Muslim ethnic minorities leaving the country by the beginning of the 20th century and immigrants from poor regions of Turkey moving into the large cities to find employment opportunities. Thus, in Turkey a socio-economic rather than ethnic segregation exists. Intending to expand this argument, this chapter puts forward an argument that although the multi-cultural atmosphere has always remained as an important characteristic of the city, the sharp spatial boundaries between diverse social groups—defined by their social, religious, and ethnic backgrounds—have dissolved over time, to be replaced by socio-economic differences and legal and illegal property markets, which have begun to define the residential patterns of the diverse social groups. The dynamics of residential segregation are affected by the following interrelated processes. First of all, the position of the city in the world has changed over time. Up until the fall of the Ottoman Empire, Istanbul was a powerful city with a strong position in the world due to its geopolitical position in the region. Since then the city has slowly re-connected with the global economy, but this time on the strength of its service and industrial sectors, which demand qualified and unqualified human capital. Secondly, the social, ethnic and economic profile of the city has changed under new cultural, political, occupational, educational and residential conditions. The city lost its rich ethnic mix under the new Republic. The non-Muslim ethnic population rapidly declined in elite neighborhoods, while the migration of poor people from the Eastern part of the country brought about an increase in illegal housing settlements. As Istanbul reconnected with the global economy and saw a growth in its service sector, the employment profile of the city shifted from being unqualified and uneducated, to educated and qualified as a result of a rise in migration of scientific and technical personnel (128,000 in 1980, 237,000 in 1990 and 394,000 in 2000) from different regions of the country and abroad (Eraydın 2008b). As the profiles and spatial preferences of different social groups changed, they moved to different parts of the city. Thirdly, the land and property market provided diverse high-income groups and land and property market speculators with numerous opportunities to make profits, and low-income groups and poor immigrants with the tools for survival and prosperity. Moreover, global capital investment in the city has generated new (partly international and partly local) actors who compete in land and property development, contributing to the increasing spatial inequalities. Both the land and property markets and the new occupational, educational and residential potentials impacted upon the different residential preferences of diverse social groups. Finally, the constantly changing legislation and regulation related to spatial development and planning in Istanbul mirrored the

Changing Dynamics of Residential Segregation in Istanbul

239

political economic transformation periods. Up until the 1950s, it was central planning by the state that dominated the spatial development of the city but it was almost impossible for governments to control migration and illegal housing development in the city. Central planning authority became irrelevant and dissolved in stages around election periods until the beginning of 1980s. After 1984 elections, some reforms were introduced to alter the central planning system. The new government increased the authority and financial resources of municipal governments and entrusted them with the preparation and follow-up of master plans, though some planning responsibilities were left partly in the central government’s hands. As a consequence of these transformations the 1980s saw the domination of fragmented planning regulations, zoning plans and squatter regularization in the city’s planning processes to deal with the growth of illegal developments. After this period the planning regime in Istanbul became more market oriented, and the city implemented special regulations that allowed piecemeal property-led developments. As a result of these changes, spatial disparities were inevitable in a city like Istanbul, where there is a clear separation of gated communities, squatter areas and middle-income neighborhoods on the basis of income and social status. This chapter will provide a review of the areas where these “spatialized inequalities” (Eraydın 2008a) occur, focusing on four distinct areas in which residential segregation has changed over time: 1. 2. 3. 4.

The historical peninsula; The squatter areas; Neighborhoods with housing transformation projects; Neighborhoods with gated communities.

To understand the full picture, it is important to comprehend the historical roots of the shift from ethnic to socio-economic causes of spatial inequality. To this end, the chapter begins by analyzing these diverse periods of change before looking at the changing dynamics of residential segregation in the above-mentioned areas. Finally, the chapter offers some conclusions on the subject. Path Dependent Residential Segregation in Istanbul There are four main time periods in the gradual socio-economic transformation of Turkey: 1) The imperial period (Ottoman Empire era); 2) Post-war republican period; 3) Liberalization and transition period of the 1980s; and 4) Neo-liberal globalization period. While political-economic tendencies followed global trends in each of these periods, characteristic socio-spatial patterns and processes and path-dependent trends are evident in each period. It is therefore worth focusing first on the economic-political context of each era, with special emphasis on the socio-spatial influences (Table 11.1). It can be seen from Table 11.1 that in each period the reasons for socio-economic segregation changed in parallel with the changing characteristics of the economy. In some periods residential segregation patterns resulted from political decisions (like the immigrant population exchange after WWI); while in other periods economic reasons were the driving force (e.g. the liberalization and privatization of the economy in 1980s).

Main political-economic conditions

Imperial economy based on agriculture and taxation

Building up a new democratic, economic-political system

Transformation from a semicontrolled mixed economy towards a liberal market economic model

Period

The Ottoman Empire period

Post-war Turkish Republic

Liberalization and transition period of 1980s

Building up a multi-national society based on a mosaic of diverse ethnic groups

Main reasons for socio-ecnomic segregation

Forced exchange of ethnic population

Regularization of squatter areas and the provision of new building rights

Appearance of new entrepreneurial urban groups

Redevelopment of illegal housing becomes a profit-making mechanism for squatter owners

Clear separation of social groups on the basis of income and social status

Massive migration from rural to urban Emergence of new socio-economic areas, accelerating population growth groups in metropolitan areas Shortage of affordable housing for migrants from rural areas

Loss of capital city function

Multi-nationalism became the essence of the city Appearance of ethnic-based upper income groups and non-Muslim elites

Istanbul became the capital of the Ottoman Empire

Changes in metropolitan areas

Changing dynamics of residential segregation in Istanbul

Table 11.1

Continuous squatter developments with a changing socio-economic profile

New luxurious housing areas (firstgeneration gated communities) and new commercial facilities, housing cooperatives of the middle income groups

Emergence of gecekondu (illegal housing) neighborhoods with increasing numbers of uneducated, unqualified and poor urban residents

Decline of the previously ethnic neighborhoods

Clear separation of neighborhoods based on ethnic, religious and social background

Elements of residential segregation

Globalization

Ongoing neo-liberalization: marketoriented regulatory transformation and privatization of main state functions

Appearance of new private sector actors (partly international) shaping urban development

Increasing interest of international commercial property investors

Regularization of squatter areas Increasing volume of educated and becomes a profit-making mechanism qualified labor for other actors and increases land and property market speculation Increasing consumerism and dependence on individual credits Central government agencies have despite the declining economic an increasing role in redistribution conditions of middle income groups of welfare in the city with large scale housing development projects Lack of affordable housing provision for lower income groups Increasing amount of commercial property development (shopping centers, office-plazas, etc.) with mainly local investment

Privatization of state land and the introduction of new forms of urban development

Mass housing supply by the state housing development authority (HDA) serving middle- and uppermiddle income groups

Gentrification in neglected neighborhoods

Social profile of the gecekondu areas changes and isolated peripheral neighborhoods of urban poor (varoş) appear

Isolated up-market residential areas: Gated communities, guarded residential towers, etc.

242

Residential Segregation in Comparative Perspective

Changing Patterns of Residential Segregation in Selected Areas The current residential segregation patterns in Istanbul have been studied by several authors (Güvenç et. al. 2005, Eraydın 2008a-b, Pınarcıoğlu and Işık 2009). Through an assessment of a number of diverse indicators (demographic, educational, and occupational) Pınarcıoğlu and Işık (2009) indicate the presence of a clear segregation pattern that has existed throughout the 1980s, with the high-income groups living along the coast, the low income groups at the periphery of the city and middle income groups between the two. However, they note that these patterns underwent a slight change in the. On the basis of occupational and educational data, Eraydın (2008a) argues that white-collar workers preferred to live in the southern part of the city along the coast, while blue-collar workers opted to live mainly in the peripheral areas close to where the manufacturing firms were based. That said, these tendencies have changed in recent years for the reasons indicated above. This chapter will focus on four types of area (the historical peninsula, squatter areas, neighborhoods with housing transformation projects, and neighborhoods with gated communities). Located around the Golden Horn, the historical peninsula was home to ethnic and religious communities in segregated areas from the fifteenth century onwards; however the decline and transformation of the early non-Muslim neighborhoods would later trigger new socio-spatial segregation and gentrification. Squatter areas, which emerged as a result of the massive migration from rural Turkey from the 1950s onwards, also reflected the changing patterns of residential segregation in the city. The central and peripheral areas became the prime target for socio-economic transformation and residential segregation; which were for many years occupied by immigrants, but are today also home to middle-income groups who cannot afford decent housing. With the middle and lower-middle income groups in mind, the Housing Development Authority (HAD) began a series of squatter renewal programs (on public land), which extended to also serve the higher-middle income groups in the 2000s (Eraydın and Taşan-Kok 2010). In addition, large-scale housing development projects in some former gecekondu areas have also changed the profile of the lower-income neighborhoods; and finally, in contemporary Istanbul, residential segregation is most obvious in the case of gated communities of the ultra-rich urban groups (both local and international). Historical Peninsula: From Elite Ethnic Neighborhoods to Gentrified Areas Pınarcıoğlu and Işık (2009) argue that ethnic-based segregation has been replaced by class-based segregation in Istanbul. The Ottoman Empire’s invasion of Roman Constantinople resulted in a great mix of social and ethnic groups in the city, coupled with new spatial patterns. In the period after the invasion of 1453 the Muslim Turkish population was a minority, but in order to run the state machine and serve the palace the emperor (Mehmed II) needed a large and prosperous population in Istanbul. Thus, people were transferred from all parts of Europe and Asia to the city (both Muslims and non-Muslims), creating a multi-national mosaic (Mansel 1995: 8). This ethnic and social mobilization process, based on the forced transfer of populations, had a strong spatial dimension that resulted in socio-spatial elements that are still visible in Istanbul today. For one thing, each neighborhood (mahalle) retained the name of the city of origin of its inhabitants, and diverse ethnic groups generally lived in concentrations. Large Greek populations were also transferred to the city at this time, and there was no religious barrier to Greeks and Turks living together in the same neighborhoods (Mansel 1995: 9). By 1477 the Turkish Muslim population had reached 59 percent, but the Armenian and Orthodox populations retained a strong presence

Changing Dynamics of Residential Segregation in Istanbul

243

(34 percent). By the sixteenth century, the city had become home to Turks, Greeks, Franks, Kurds, Serbs, Arabs, Armenians, and many other nationalities but, over time, the proportions shifted in favor of Muslim Turks, who are now the predominant group in the city (99 percent) (Mansel 1995). In the second part of the nineteenth century the city went through a rapid growth and transformation process as a result of the Ottoman Empire’s economic and political success in the region. This period saw a plethora of social and institutional reforms and attempts at Westernization, including the artificial adoption of the French revolution vocabulary which in the end served the happy few (the elite) and helped the international commercial bourgeoisie to settle in the city (Çelik 1986). The neighborhood boundaries of ethnic, religious and social groups became more defined during this period; and the composition of the population changed drastically, in part due to the arrival of Muslim refugees escaping from Southeastern Europe and Russia, and due to the growing interest of non-Muslim foreigners who wanted to benefit from the privileges given to Western tradesmen and investors under commercial treaties (Çelik 1986). By the end of the 19th century the non-Muslim ethnic population (Greek Orthodox, Armenian Gregorian, Jewish, Catholic and Protestant, Bulgarian and Latin communities) constituted more than half of the total (56 percent, according to the 1885 census) (Çelik 1986: 38), and were able to regulate their own religious, educational and general administrative businesses due to the reforms enacted in the era. Within this social, political and economic sphere the city went through a very interesting sociospatial transformation process in terms of ethnic and social segregation, especially in the central areas. The most significant trend was the emergence of a particular ethnically, and later, socially defined settlement pattern. In the early nineteenth century the non-Muslim communities became concentrated on the northern side of the Golden Horn in Galata (Map 11.1), turning the area into an isolated up-market area and resulting in the arrival of wealthier Muslims and senior government officials (Çelik 1986). The Galata neighborhood and its northern extension (Pera) had from the earliest years of the capital been home to non-Muslims. When Constantinople was conquered in the fifteenth century Galata was a Genoese Colony, and was occupied by an Italian Christian community called Cenovalılar (from the Republic of Genoa). Over the following centuries the area retained its multinational character, despite the fact that the proportion of the Muslim-Turkish population was slowly increasing (Kuban 2000). As İslam (2003) and (Çelik 1986) indicate, the political center of the Empire had moved from Topkapı Palace to the Çırağan, Dolmabahçe, and Yıldız Palaces; and upper-income Muslim neighborhoods naturally developed around these new areas (Çelik 1986). The internationalization of the population was strongly affected by the number of foreign embassies and residences that started to be established in and around the Pera area, which also had a pull effect on upper-income groups, turning Beyoğlu into a socially and spatially isolated urban area (İslam 2003). Although the sharp boundaries of the ethnically defined neighborhoods of Galata in the fifteenth century became less distinct in time, the non-Muslim-dominated neighborhoods extended toward the northern side of the Golden Horn (Map 11.1) and Beşiktaş up to Rumelihisarı. Greek, Armenian and Jewish groups, as well as Italians, Bulgarians and Romanians, also became quite mobile in later years, opting to live in their own concentrated zones. According to the 1885 census, the populations of Pera, Galata and Tophane were about 80 percent non-Muslim; while Beşiktaş and other villages up to Rumelihisarı were 57 percent nonMuslim (Çelik 1986). At this time, non-Muslims constituted 45 percent of the population on the whole historical peninsula. Çelik (1986: 38) points out that this resulted in a particular settlement pattern, with non-Muslims living to the north of the Golden Horn and Muslims in other Istanbul neighborhoods. Later, the non-Muslim groups also moved to other parts of Istanbul, mainly to build summer houses on the other side of the Bosporus or on the islands, however, their tendency to

244

Residential Segregation in Comparative Perspective

concentrate in one area continued. This sharp ethnic definition of urban neighborhoods disappeared over time, initially due to the bloody events of 1821 when non-Muslims in Galata were killed and their properties looted under suspicion that the resident Greeks, Armenians and Jews would rise against the Empire; and secondly, as a result of the forced emigration and expulsion of the Greek population in 1922–24. From 1923 onwards, as the newly established Republic needed a new modern capital city to bolster its image of a modern state, enormous effort was put into the planning of the new capital city of Ankara, while Istanbul was left to its own devices. Istanbul became poor and neglected, and the cosmopolitan atmosphere vanished (Kuban 2000). In this period, spatial inequalities were produced by three broad transformations: the decline of the ethnic non-Muslim population in the city, increased migration from rural areas, and the emergence of new social groups within the framework of the modernization of the city in the 1950s. An important transformation was experienced in the profile of urban society in the first years of the Republic with the forced exchange of the Greek population, and then again in the 1940s due to the further decline of the ethnic non-Muslim population in the city. This was followed by an increase in the arrival of poor and unqualified immigrants from the Eastern Anatolian regions to the city. The 1930s plan of Henry Prost, part of the industrialization drive of the new Republic, also contributed to spatial mobility processes in later years after suggesting a new industrial development axis along the Golden Horn (Haliç) to replace the derelict and run-down neighborhoods in which the nonMuslim communities had lived. This area would also become the subject of urban regeneration processes in the years to come with the replacement of highly polluting industrial activities. A further decline of the non-Muslim population was triggered with the implementation of a new tax (varlık vergisi) in 1942 under the new political climate, which imposed an additional capital tax (up to 230 percent of their capital) on non-Muslim ethnic businesses (namely Armenian, Greek and Jewish) to help the growth of Turkish firms. The tax not only brought about the extinction of longestablished firms overnight, but also a decline in the attractiveness of Istanbul for international migration (Mansel 1995: 424). Tension among the remaining Greek population in Istanbul grew, leading to the bloody riots of 6–7 September 1955 that were triggered by the troubles between the Greeks and Turks in Cyprus; and in the same era (1948), a large proportion of the Jewish population left the city for the newly established State of Israel. While the city was losing increasing numbers of non-Muslims starting in the late 1940s, it was quickly gaining an immigrant population from the poorer rural areas of the country, changing the spatial patterns of social and ethnic segregation in the city. The previously elite ethnic neighborhoods like Balat (which had predominantly Jewish and, to a lesser extent, Armenian populations) and Fener (home to a predominant Greek population) went into decline for many decades, and would later become ideal areas for the new socio-spatial mobility patterns in the city. The decline and transformation process of the non-Muslim neighborhoods as a result of these processes would trigger new socio-spatial segregation and gentrification dynamics in the years to come. In time the previously ethnically segregated neighborhoods (such as Galata, Kuzguncuk, and Cihangir—Map 11.1) became derelict zones and were regenerated later on. Some gentrifiable areas were also created as a result of urban renewal projects (Fener and Balat districts—Map 11.2), or under urban rehabilitation and renewal plans and regulations issued by local governments (Ortaköy—Map 11.1). Depending on the type and place of gentrification, different levels of displacement occurred in these areas, changing the spatial inequalities one again in the city. Explaining the different patterns of gentrification, Uzun (2001) claims that the changing economic conditions affected the residential choices of households in the city, as most of the middle- to high-income groups moved out to peripheral areas. One group of people, which included

Changing Dynamics of Residential Segregation in Istanbul

Map 11.1

245

Districts of Istanbul in the historical peninsula and the city center

Source: Prepared by Sıla Özdemir, based on Google Maps.

professionals, managers, academics, artists and white-collar workers, showed a different pattern of behavior, preferring to search for options closer to the central city to save time and money, to be close to entertainment and other services, and to live in prestigious areas. It was these people who initiated the fragmented and individual gentrification processes in central neighborhoods like Cihangir, and caused the transformation of the existing housing stock and even the building of new apartment buildings on empty plots of land (Uzun 2001). İslam (2005) argues that the nearby Galata and Cihangir neighborhoods were subject to gentrification in different phases: Galata followed a slower pace due to the presence of traditional commercial spaces and industrial workshops, while Cihangir transformed into an upper/middle-income neighborhood very rapidly throughout the 1990s. On the historical peninsula the poor neighborhoods of Fener and Balat (former non-Muslim neighborhoods), which had been the former settlement areas of non-Muslim residents, were also occupied by rural immigrants from the eastern parts of the country. Existing houses were subdivided into cheap apartments, and became home mainly to poor, traditional and religious groups. This decline was supported by the industrialization of the Golden Horn initiated by the Henry Prost Plan of 1930s; and later, both the urban rehabilitation plans funded by the European Commission and the urban planning initiatives to replace the industrial land uses elsewhere and decontaminate the polluted land and water created ideal conditions for gentrification in these areas. As İslam (2005) argues, this latest wave of gentrification differed from earlier examples, in that it was not triggered by spontaneous or individual processes, but more by institutional initiatives. As is often the case, the gentrification process caused the replacement and even displacement of lower and disadvantaged groups by middle- and upper-middle income groups in the city. İslam and Enlil (2006) point out that the new dynamics of gentrification in Galata were due to investment

246

Residential Segregation in Comparative Perspective

by development companies and large-scale investors rather than individuals, arguing that this speculative process caused the displacement of low-income groups living in the area. İslam (2005: 130) presented an overview of the profile of the gentrifiers in some areas in Istanbul, in which he revealed that the gentrifiers were from similar social and economic backgrounds and high levels of education, although they appear to be older than the usual gentrifier prototype. He also noted that the new residents were predominantly architects, journalists and artists, as well as workers in the financial and managerial sector workers, thus confirming the earlier analysis of Uzun (2001). Illegal Housing Areas: From Squatter Areas to Middle-income Neighborhoods Kinship (hemşerilik) and family networks were very important among the immigrants from the poor regions of the country in their selection of areas for the construction of illegal housing (gecekondu), resulting in squatter neighborhoods with concentrations of communities based on shared ethnic, religious or cultural backgrounds. Large-scale gecekondu neighborhoods sprung up on the periphery of the city outside the city limits, and in smaller concentrations dispersed throughout the city on steep slopes or on land already slated for development (Tekeli 1994). The immigration of the poor, unqualified and uneducated labor force and the growth of self-made illegal houses and neighborhoods (gecekondu) from the 1950s onwards mark the urban landscape of the city to this day (Map 11.2), along with the associated problems. The modernizing state aimed to apply the principles and practice of a welfare state, which protects the well-being of citizens and provides them with public services such as affordable housing, education and health, though not adequately. The concept of a welfare state was introduced by the Constitution of 1961 after the military intervention of May 27, 1960 in Turkey but it did not really accomplish affordable housing provision. Owing to the lack of housing, immigrants from the

Map 11.2

Squatter areas (unauthorized settlements) with and without improvement plans

Source: Prepared by Sıla Özdemir on the basis of Eraydın and Taşan-Kok (2010).

Changing Dynamics of Residential Segregation in Istanbul

247

poorer regions invaded public land (or whatever land they could find) and built their own shelters in order to survive. These illegal developments triggered the emergence of a dualistic land and property regime. Another important reason for socio-economic segregation can be found in the transformations in the city that resulted from the political and economic modernization seen between 1923 and 1960 in Turkey, when a Republican elite was slowly emerging. The planning of Istanbul gained momentum in the 1950s with the new government’s modernity project to rebuild the city with high-rise buildings and wide avenues (Kuban 2000). The import substitution-oriented Turkish economy was grounded in a strong industrialization policy, and the state played a prominent role in economic development, especially between 1940 and 1950. During this period, the local business environment was isolated from external markets, and foreign direct investment was discouraged (Taşan-Kok 2004). The deep economic downturn that hit at the end of the 1970s (in the wake of the world oil crisis) caused high inflation, an unbalanced distribution of income and a decrease in production; and the result was political and social instability. At this conjuncture the spatial patterns of the city were also undergoing dramatic change. Firstly, the previously ethnic neighborhoods began to deteriorate in line with the declining population. The new urban residents, the migrants, managed to survive the lack of affordable housing by building squatters and fought for jobs in the booming industrial and commercial sectors. However, as Pınarcıoğlu and Işık have noted (2009: 471), they never became passive residents of the city; rather, by using kinship relations and personal networks based on their places of origin, they quickly adapted to the new conditions, using to their advantage the informal economy and the informal property market. The rapidly improving economic conditions helped these lower income groups to achieve upward social mobility based on land rents and profits from informal economic activities. Thus, the poor were never trapped spatially or economically, as was the case in the ghettos and no-go areas in the Western world; instead, new urban neighborhoods were established through the invasion of available public land. Life in these areas was quite rural-like (single-storey, rapidly built, but quite proper family houses), but in time they were improved and even redeveloped, creating a profit-making mechanism in the city. The opportunity to change one’s social, economic and spatial conditions allowed upward social mobility, and in fact, on top of the opportunities created for entrepreneurs through the gray economy, the regularization and legalization of illegal housing areas became common practice thanks to populist political campaigns during the runup to elections. The concentration of illegal housing areas dramatically changed the morphology of urban areas (Tekeli 1994); and the legalization of illegal housing became such a profitable practice and a mechanism for income transfer that in the following decades the characteristics of the people living in those areas also changed (Taşan 1996). Thus, the property market became a provider not only of opportunities for lower income groups, but also for speculators. Under these favorable conditions the construction sector became an area of strong economic activity for new actors (contractors, real estate agents, developers, investors, insurance services, architects, etc.) involved in urban transformation, and caused more wealth transfer in the following period. Secondly, the mobility of the newly emerging urban middle-income groups had spatial impacts. The growing middle-income groups began to move into the old summer house areas of the nonMuslim minorities on the Asian side of Bosporus (Kadıköy), turning them into permanent housing neighborhoods (Kurtuluş 2005a). With good railway and sea-transport connections, these areas have provided an opportunity to the middle-income groups who work in the city center to live in more affordable, but more remote areas (Kurtuluş 2005a); and subsequently, the empty areas between these leapfrog developments and the central city attracted further new housing developments.

248

Residential Segregation in Comparative Perspective

While the economy was improving as a result of a strong privatization and market-oriented transformation throughout the 1980s, three important changes brought new spatial inequalities in this period in Istanbul. First of all, the spatial separation of social groups on the basis of income and social status became more visible; secondly, the regularization of illegal housing areas under new legal arrangements, especially before elections, created a profit-making mechanism in the form of redevelopment projects. This process changed the profile of these neighborhoods as the rural migrants began to move out to other affordable places, and other urban lower income groups began to move into these areas. Finally, a new entrepreneurial urban rich class appeared with particular demands and lots of money to spend. Regarding the high levels of social mobility during the 1980s, Pınarcıoğlu and Oğuz (2009) claim that the urban poor were not excluded from urban socio-spatial processes, and retained their status with the hope of upper social mobility. This was, of course, created and maintained thanks to the land and property markets, both in the regularized gecekondu areas, and in the areas where the new urban rich were concentrated in new forms of housing, namely gated communities. In the 1980s, unauthorized developments and amnesty laws went hand-in-hand. For political reasons, gecekondu developments were legalized around election times through a series of laws and regulations (Taşan 1996), thus changing the profile of the population living in these areas. The regularization of the squatter areas usually brought about fragmented redevelopments throughout the city, as squatter renewal plans suggested higher density developments. This tendency caused two main trends in the city: First of all, unlike the first wave of rural poor that had flocked to urban areas, new lower-middle income groups began to move into these areas to take advantage of the affordable housing. Secondly, realizing the profitability of the squatter renewal processes, new actors (other than the squatter owners themselves) became involved in the redevelopment processes in these areas. Both trends contributed to the transformation of the previously poor immigrant squatter areas into lower-middle income group areas, accompanied by a new set of social issues related to the socio-economic inequalities in society. Consequently, the boundaries between the rich and poor became more visible when compared to the transitional pre-Republican era, while the squatter areas became more fragmented, widespread and divided due to increasing unemployment and the income disparities that emerged under the new liberal state policies (Bozkulak 2005). As a result, the profile of the squatter residents began to be defined in terms of a combination of income, education and labor patterns. The land and property markets played a key role in the redistribution of wealth in Istanbul, though over time this role has shifted from providing social benefits to disadvantaged groups toward the redistribution of speculative commercial property-led urban development-based profits among private sector actors and local governments. Between 1983 and 1988 several laws and regulations were issued to regulate the transformation of squatter development areas (Erkan 2009). These began after the 1984 elections, when reforms were introduced in the central planning system. The new government increased the authority and financial resources of municipal governments and charged them with the preparation and followup of master plans; However, some planning responsibilities were left in the hands of the central government, which gave power to central government agencies to create profitable commercial developments in the city without interference or control from local governments (Taşan-Kok 2004). The government then passed the Law on Mass Housing and initiated several mass housing projects that served only the middle- and upper-middle income portion of society. Subsequently, the Law on Regularization of the earlier illegal housing areas was adopted for low-income squatter areas, offering them higher building rights. The preparation of improvement plans as a means of regulating unauthorized settlement areas meant a transfer of wealth to certain groups in Istanbul, since they provided high building rights and special planning conditions to those people

Changing Dynamics of Residential Segregation in Istanbul

249

who had either invaded public lands, or who had built their houses in areas without planning or building permission (Eraydın and Taşan-Kok 2010). The squatter redevelopment processes also created socio-economic disparities among the squatter residents, as not all were able to take advantage of the opportunity and profit from the renewal. Those who could become part of the profit-making mechanisms related to squatter renewal were able to move out of poverty, while others remained poor (Bozkulak 2005). These processes even created another group of squatter owners who took advantage of the redevelopment as a profit-making mechanism, moving from one squatter area in the city to another (Taşan 1996). Squatter regularization processes have also transformed unauthorized areas into a form of slum neighborhood (varoş) (Işık and Pınarcıoğlu 2001, Bozkulak 2005). Housing Transformation Project Areas: From Decline to High Density Housing Development The new political and economic climate of the 1980s brought about economic reforms that included the adoption of flexible exchange rates, export-oriented production, public enterprise reform and privatization, financial liberalization, import liberalization and the promotion of foreign direct investment (Arıcanlı and Rodnik 1990). These policies had a strong impact on retail and wholesale trade, finance, insurance, real estate and business services, as well as the manufacturing functions concentrated in and around the CBD. As a result of the rapid population growth and the rising share of service jobs in the CBD, new scattered sub-centers emerged and Istanbul was transformed from a monocentric into a polycentric city (Dökmeci and Berköz 1994). Up until the beginning of the 2000s the number of foreign property development firms in Istanbul had been very limited as a result of their hesitation about entering into this extremely dynamic and risky market, and consequently the urban landscape of the city was re-shaped on the basis of local capital accumulation processes in this period. There were a few other reasons behind this particular market-led transformation: First of all, local competition was strong and local actors were resilient to economic crises (Taşan-Kok 2004, 2008); secondly, the labor market was able to provide a young and motivated workforce in both qualified and unqualified (cheap) forms. In the years that were to follow, it was this young labor force that would increase the competitive edge of the city, combined with a high availability of infrastructure facilities such as high-tech office space, a comprehensive transport network and residential facilities for the service economy, as well as production sectors with low- (textile) and high-technology (automotive and electrical and non-electrical) industries (Eraydın 2008a). Thirdly, although massive privatization activities had taken place, the strong role of the state was still evident in the main urban development decisions. Especially after 2000, the central government became a main decision maker in certain parts of the Istanbul Metropolitan Area through the activities of the Privatization Agency and the Housing Development Administration (HDA)/(TOKI in Turkish), which function under the Prime Ministry. These combined to foster a clear spatial separation between different socio-economic groups. As Table 11.2 shows, the most striking change in the occupational composition of the city occurred in the finance and banking sector between 1980 and 2000. Eraydın (2008b) argues that the spatial segregation of occupational groups at the neighborhood level was particularly evident: people with white-collar jobs (from the finance and banking sector, including the clerical work force) lived in the southern part of the city along the coast, while people with blue-collar jobs lived in the north, close to the large factories. This trend continued with the inflow of increasing numbers of newcomers with high scientific and technological skills into the white collar neighborhoods in the central areas and along the coast (Eraydın 2008b).

250

Table 11.2

Residential Segregation in Comparative Perspective

The change of the occupational composition in Istanbul from 1980 to 2000 1980 Number

Manufacturing Construction Trade Transport and communication Finance and banking Public services Others* Total

1990 %

Number

2000 %

Number

%

526,490 111,690 279,699 104,929 82,715 333,587 422,479

33.7 7.1 17.9 6.7 5.3 21.3 27.0

834,888 224,126 486,177 167,467 179,558 456,245 646,531

32.8 8.8 19.1 6.6 7.1 18.0 25.5

1,097,051 215,925 650,295 221,298 283,404 696,033 994,405

31.6 6.2 18.7 6.4 8.2 20.0 28.6

1,563,939

100.0

2,539,963

100.0

3,471,400

100.0

Note: * Covers people working in agriculture, mining, public infrastructure services and in undefined activities. Source: TSI (1980; 1990; 2000), Eraydın 2008a.

While all these changes contributed to new forms of residential segregation, some of the older tendencies remained, although in different forms. For instance, the regularization of the squatter areas continued to be a profit-making mechanism for squatter residents, but the HDA’s increasing involvement in squatter renewal projects moved this process out of the scope of the provision of affordable housing for lower income groups, becoming a profitable business providing of housing for middle- and upper-middle income groups. Moreover, many other state enterprises, such as the Turkish Maritime Enterprise, and the General Directorate of State Railway, Harbour and Airport Construction began to act as land and property market entrepreneurs, transferring the ownership rights of their existing facilities and land. As a result, in the early 2000s Istanbul became a playing field of different authorities and actors (Eraydın and Taşan-Kok 2010) in which the profile of the residents in the former squatter areas changed, transforming them into higher density and quality housing areas for middle-income groups. The provision of different forms of housing increased (Map 11.3) for the middle- and upper-middle income groups; and all of these projects helped to create clear physical boundaries based on income and social status. Although migration from the rural areas to the city had decreased, the lower-income groups continued to live in squatter areas and peripheral underclass settlements (varoş) around the city. The HDA and other local government agencies initiated housing renewal projects to serve the higher income groups; however, the disadvantaged populations were displaced, resulting in new spatial inequalities. The recent renewal project to transfer some of the Roma population from the Istanbul neighborhood of Sulukule to remote organized housing areas on the periphery is a perfect example of this kind of public sector-led spatial inequality. The areas with a high density of transformation projects in Map 11.3 correspond to settlements occupied mainly by high-income white-collar workers (scientific and technical employees, managers, administrators, and people working in financial and commercial activities) in the 2000s. According to studies by Güvenç et al. (2005) and Eraydın (2008a) a large proportion of these areas are occupied by social groups with high education levels who work in the service sector. Some studies (Güvenç and Işık 2002, Eraydın 2008a, Pınarcıoğlu and Işık 2009) show that the distribution of the population by education level is in parallel with the income distribution in the

Changing Dynamics of Residential Segregation in Istanbul

Map 11.3

251

The distribution of new housing, renewal and transformation projects by HDA, metropolitan municipality and district municipalities

Source: Prepared by Sıla Özdemir on the basis of Eraydın and Taşan-Kok (2010).

city (Eraydın 2008a), meaning that the highly educated groups also (partly) prefer neighborhoods with a high concentration of housing transformation projects. Another important trend has been the increase in the purchasing power of people due to the availability of individual credits among the middle-income groups. Newly available individual consumer credits, mortgages, credit cards and favorable interest rates have triggered, supported and increased consumerism in society, and thanks to these conditions commercial property development (shopping centers, office plazas, entertainment facilities, etc.) boomed throughout the 2000s, attracting luxury residential developments around them. These projects have not only contributed to the changing urban landscape, but have also attracted more international capital. These new international businesses and private sector actors have been influential in shaping the built environment, providing housing for higher income groups with high levels of education, mainly in the form of gated communities and luxury residential projects. Gated Communities: From Top-end Housing Provision to a Common Practice New luxury housing areas began to appear to meet the demand of new high income groups in the city from the 1980s onwards. The growing income disparities of the mid-1990s began to be reflected in the creation of super luxury residential developments and commercial property developments built with international capital, while local governments also became more open to property-led urban development projects. The development of first-generation gated communities also contributed to a clear definition of the boundaries between the better and worse off parts of urban society.

252

Residential Segregation in Comparative Perspective

Supported by the increasing capital accumulation in the city, the popularity of gated communities gained in momentum in the mid-1980s in line with the new consumerist lifestyle (Kurtuluş 2005a); but the real boom in such enclosed housing areas (not only in the form of villas with large gardens, but also of gated residential towers) took place from the late-1990s onwards (Kurtuluş 2005a). Although the first cases were designed exclusively for the highest income groups, the concept has since become so popular in urban society that entering into the 2000s, different forms of guarded communities at lower prices have also been developed for the middle- and upper-middle income groups as well (Kurtuluş 2005a). Baycan-Levent and Gülümser (2007) refer to the existence of approximately 400 gated communities in Istanbul, housing some 60,000 people (Pérouse 1999, 2001, Tempo Dergisi 2003). Although this may not seem to be a large proportion of the total population of the city, gated communities are highly visible in the physical fabric of the city, and have changed the urban landscape dramatically as places of spatial segregation. Istanbul’s gated communities are scattered throughout the city, springing up wherever large areas of land become available. Baycan-Levent and Gülümser (2007) made a classification of gated communities in Istanbul on the basis of their physical characteristics, and defined four categories: gated towers, gated villa towns, gated apartment blocks, and gated towns with mixed categories of housing. On the basis of the income criteria, Kurtuluş (2005b: 169–70) defines three types of gated communities in the city: 1) Gated communities of the new middle-income groups located in the peripheral sub-regions (e.g. Bahçeşehir); 2) Luxury gated communities to serve the rapidly growing urban rich groups that are seeking secure investment areas and ways to satisfy their consumerist expectations (e.g. Acarkent); and 3) Radically closed communities that serve the ultra-rich parts of society and have their own facilities (e.g. Beykoz Konakları and Kemer Country). These projects have become isolated islands of higher income groups, and are sometimes even located next to derelict neighborhoods and squatter areas. Geniş (2007) provides insight into the socio-economic conditions of the residents of Kemer Country (which is among the most prestigious), claiming that the house prices range from between $300,000 to $2 million, with monthly rents averaging between $2,500 and $10,000. Yönet and Yirmibeşoğlu’s (2009) study of 20 gated communities (67 households) in the city details the socio-economic profile of the residents of these areas, reporting that almost 93 percent are highly educated (55 percent of whom are university graduates); 80 percent are homeowners; 60 percent have other property in Istanbul and 51 percent have summerhouses. Pınarcıoğlu and Oğuz (2009) emphasize that the impact of the scattered gated villa towns and gated towns (which consist both of villas and luxurious apartment blocks) at the fringe of the city is not too visible in these large peripheral areas. However, the gated towers and apartment blocks in the dense inner city neighborhoods form islands of higher socio-economic groups that are visibly and explicitly isolated from their surroundings. In recent years the prestige of living in this type of housing has increased the popularity of gated (or guarded) housing in the city for middle- and upper-middle income groups as well (Kurtuluş 2005b). Conclusions The characteristics and patterns of the segregation processes have changed in time from ethnically defined segregation toward socio-economic segregation; and socio-economic differences are the driving force behind residential segregation in Istanbul. Moreover, local and central governments also create and/or support residential segregation in urban plans and legislation, such as in urban renewal programs, squatter redevelopment plans and regeneration projects. Three main processes

Changing Dynamics of Residential Segregation in Istanbul

253

should be emphasized to explain residential segregation in Istanbul. First of all, particular pathdependent social, economic and political development in the segregation processes exists; secondly, throughout history these spatial inequalities have been based on land and property market dynamics; and finally, social groups change locations in the city based on land and property prices. The case-specific events can be best noticed in the changing profile of urban society resulting from the change from the Ottoman Empire to the democratic Turkish Republic. The sharp boundaries between ethnic groups experienced in the late-nineteenth and early-twentieth century in Istanbul completely disappeared with the dissolving and declining non-Muslim population; however new sharply defined boundaries were established between high- and low-income groups, with gated communities and poor squatter neighborhoods being at the opposite ends of the spectrum. The spatially segregated elite ethnic groups had been replaced by relatively poor new urban residents by the beginning of the twentieth century, and were in turn replaced by the gentrifiers. Poor squatter residents were replaced by middle- and upper-middle income groups, and, in some cases, by the residents of gated communities in the latter years of the twentieth century. In addition, spatial inequalities were also initiated as a result of state-led urban renewal projects, which had brought about the displacement of marginalized and low-income groups by the beginning of the 2000s. Coming to the twenty-first century, new highly skilled immigrants moved into the city. As different studies have underlined (Eraydın 2008a and 2008b, Güvenç and Işık 2002, etc.), with the growth of the financial sector, the attractiveness of neighborhoods in close proximity to the CBD increased, and a transformation of the social structure of these areas, from a blue-collar-dominated occupation structure to a more mixed pattern, occurred. The historical review of events presented in this chapter also shows the density of the interconnected processes that lead to segregation in a city. Moreover, it shows that if one is to comprehend the real processes behind spatial inequalities in a city, then the path-dependent events that lead to the spatial and social mobility of diverse groups in the city needs to be analyzed. The particularities of the social and cultural geography of Istanbul of course merit a deeper analysis than this brief review of the main events. Acknowledgments The author wishes to express her gratitude to Assoc. Prof. Dr Duruöz Uzun for her help and support on this chapter, and to Sıla Özdemir for digitalizing the maps. References Arıcanlı, T. and Rodnik, D. 1990. “An Overview on Turkey’s Experience with Economic Liberalization and Structural Adjustment.” World Development 18(10), 1343–50. Baycan-Levent, T. and Gülümser, A.A. 2007. “Gated Communities in Istanbul: The New Walls of the City” EURODIV Paper 51, Fondazione Eni Enrico Mattei. Bozkulak, S. 2005. “Gecekondu’dan “varoş“a: Gülsuyu mahallesi” (From squatter to slum: Gülsuyu neighbourhood). In Istanbul’da Kentsel Ayrışma: Mekansal Dönüşümde Farklı Boyutlar (Urban Segregation in Istanbul: Different Dimensions in Spatial Change), ed. H. Kurtuluş. İstanbul: Bağlam Yayıncılık, 239–66. Çelik, Z. 1986. The Remaking of Istanbul: Portrait of an Ottoman City in the Nineteenth Century. Washington: The University of Washington Press.

254

Residential Segregation in Comparative Perspective

Dökmeci, V. and Berköz, L. 1994. “Transformation of Istanbul from a Monocentric to a Polycentric City.” European Planning Studies 2(2), 189–201. Eraydın, A. 2008a. “The Conditional Nature of Relations between Competitiveness, Social Cohesion and Spatial Inequalities: The Evidence from Istanbul.” In Cities Between Competitiveness and Cohesion, eds P. Ache, H.T. Andersen, T. Maloutas, et al. Dordrecht: Springer, 99–115. Eraydın, A. 2008b. “Cohesion and Spatial Segregation in Istanbul the Impact of Globalisation on Different Social Groups: Competitiveness, Social Cohesion and Spatial Segregation in Istanbul.” Urban Studies (Edinburgh, Scotland) 45(8), 1663–91. Eraydın, A. and Taşan-Kok, T. 2010. Urban Social Movements and Property Right Regimes in Different Modes of Metropolitan Governance: The Evidence from Istanbul. Unpublished manuscript. Erkan, N.E. 2009. “Kentsel yenileme ve Istanbul: Kente teslim olmak, kenti teslim almak” (Urban renewal and Istanbul: Surrender to the City, to Conquer the City). In Gecekondu, Donusum, Kent (Squatter, Transformation, the City), eds S. Kayasu, O. Işık, N. Uzun, et al. Ankara: ODTÜ Mimarlık Fakültesi Yayınları, 219–40. Geniş, S. 2007. “Producing Elite Localities: The Rise of Gated Communities in Istanbul.” Urban Studies (Edinburgh, Scotland) 44(4), 771–98. Güvenç, M. and Isık, O. 2002. “A metropolis at the Crossroads: The Changing Social Geography of Istanbul Under the Impact of Globalization.” In Of States and Cities: The Partitioning of Urban Space, eds P. Marcuse and R. van Kempen. Oxford: Oxford University Press, 203–20. Güvenç, M., Eraydın, A., Yoncacı, P. and Özbay, S. 2005. “Istanbul’un eylem planlamasina yonelik mekansal gelisme stratejileri arastirma ve model gelistirme calismasi” (Research study for modelling and developing spatial development strategies for Istanbul’s action plan), Unpublished Research Report Section 1/B, ODTU: Ankara. Işık, O. and Pınarcıoğlu, M. 2001. Nöbetleşe Yoksulluk: Sultanbeyli Örneği (Poverty In-turn: The Case of Sultanbeyli). Istanbul: Iletisim Publications. Işık, O. and Pınarcıoğlu, M. 2009. “Segregation in Istanbul: Patterns and Processes.” Journal of Economic and Social Geography (TESG) 100(4), 469-484. İslam, T. 2003. Istanbul’da Soylulastirma: Galata Ornegi (Gentrification in Istanbul: Case of Galata), Unpublished PhD manuscript submitted to Yildiz Technical University. İslam, T. 2005. “Outside the Core: Gentrification in Istanbul.” In Gentrification in a Global Context: The New Urban Colonialism, eds R. Atkinson and G. Bridge. London: Routledge, 121–36. İslam, T. and Enlil, Z.M. 2006. Evaluating the Impact of Gentrification on Renter Local Residents: The Dynamics of Displacement in Galata. Paper to the 42nd ISoCaRP Congress: Istanbul, September 2006. Kuban, D. 2000. Istanbul: Bir Kent Tarihi. (Istanbul: An Urban History). Istanbul: Tarih Vakfi Yurt Yayinlari. Kurtuluş, H. 2005a. “Bir ütopya olarak Bahçeşehir” (Bahçeşehir as an utopia). In Istanbul’da Kentsel Ayrışma: Mekansal Dönüşümde Farklı Boyutlar (Urban Segregation in Istanbul: Different Dimensions in Spatial Change), ed. H. Kurtuluş. İstanbul: Bağlam Yayıncılık, 77– 126. Kurtuluş, H. 2005b. “Istanbul’da kapalı yerleşmeler ve Beykoz Konakları örneği” (Gated communities in Istanbul and the case of Beykoz Konakları). In Istanbul’da Kentsel Ayrışma: Mekansal Dönüşümde Farklı Boyutlar (Urban Segregation in Istanbul: Different Dimensions in Spatial Change), ed. H. Kurtuluş. İstanbul: Bağlam Yayıncılık, 161–86. Mansel, P. 1995. Constantinople: City of the World’s Desire, 1453–1924. Middlesex: Penguin.

Changing Dynamics of Residential Segregation in Istanbul

255

Pérouse, J.F. 1999. Les Figures du Risque Dans le Discours Publicitaire des Cites Privées: le Cas d’Istanbul. Unpublished manuscript, Istanbul. Pérouse, J.F. 2001. “Les Cites Sécurisées des Terrirories périphériques de l’Arrondisemen d’Eyup (Istanbul) ou les Mirages de la Disticntion.” L’Information Geographique 68, 139–54. Pınarcıoğlu, M. and Işık, O. 2009. “Segregation in Istanbul: Patterns and Processes.” Tijdschrift voor Economische en Sociale Geografie 100(4), 469–84. Taşan, T. 1996. Urban Populism: Redistribution of Urban Land Rents. Unpublished Masters Thesis, Middle East Technical University: Ankara. Taşan-Kok, T. 2004. Budapest, Istanbul, and Warsaw: Institutional and Spatial Change. Delft: Eburon Publishers. Taşan-Kok, T. 2008. “Spatial and Institutional Consequences of Globalization in Istanbul: Focus in Commercial Property Markets.” In Dutch Windows on the Mediterranean: Dutch Geography 2004–2008, ed. V. Mamadough, S.M. de Jong, F. Thissen, J, van der Schee and M. van Meeteren, NGS: Utrecht, 376–84. Taylor, P.J., Walker, D.R.F. and Baeverstock, J.V. 2002. “Firms and their Global Service Networks.” In Global Networks, Linked Cities, ed. S. Sassen. New York, London: Routledge, 93–115. Tekeli, İ. 1994. The Development of the Istanbul Metropolitan Area: Urban Administration and Planning. Istanbul: IULA-EMME. Tempo Dergisi 2003. “Siteler Sosyal Çatısma Yaratacak” (Gated communities will create social unrest). Tempo, 16–22 January 2003, 44–5. TSI (Turkish Statistical Institute) 1980. Census of Population; Social and Economic Characteristics of Population, Turkey. State Institute of Statistics Prime Ministry Republic of Turkey: Ankara. TSI (Turkish Statistical Institute) 1990. Census of Population; Social and Economic Characteristics of Population, Turkey. State Institute of Statistics Prime Ministry Republic of Turkey: Ankara. TSI (Turkish Statistical Institute) 2000. Census of Population; Social and Economic Characteristics of Population, Turkey. State Institute of Statistics Prime Ministry Republic of Turkey: Ankara. Uzun, N. 2001. Gentrification in Istanbul: A Diagnostic Study, Utrecht. KNAG/Faculteit Ruimtelijke Wetenschappen Universteit Utrecht (Netherlands Geographical Studies No. 285): Utrecht. Yönet, A. and Yirmibeşoğlu, F. 2009. Gated Communities in Istanbul: Security and Fear of Crime. Paper to the ENHR Conference: Prague, July 2011.

This page has been left blank intentionally

Chapter 12

Social Polarization and De-segregation in Athens Thomas Maloutas, Vassilis Arapoglou, George Kandylis and John Sayas

Introduction This chapter provides evidence regarding the patterns, level and trends of residential segregation in Athens, in terms of class and ethnicity, based on census data (especially from 1991 and 2001). Explanations of these patterns, level and trends—which largely contradict the expectations according to dominant theories of socio-spatial polarization—are sought in the changing local labor market profile, the residual welfare state especially in housing provision, the broader welfare system involving a crucial role for families and their networks, and the built environment in terms of specific housing forms and property structures that dominate in the city’s land and housing markets. A Brief Outline of the Athenian Context Although its site has been inhabited for several millennia, Athens was not more than a big village when it became the capital city of the Modern Greek state in the 1830s. The total absence of building stock from the early Modern period—unlike in most historic European cities—testifies to Athens’ interrupted urbanity for very long intervals. The type of growth that Athens has experienced since the end of the nineteenth century has not been on the model of the industrial city. To a large extent, external events accounted for this growth rather than the dynamism of its modern economy. In the 1920s, several hundred thousand refugees settled in the periphery of Athens following the unsuccessful invasion attempted by Greece in Asia Minor; in the late 1940s there was another important migration wave from rural areas following the Civil War (1946–49) and the ensuing difficult survival of the vanquished (i.e. those that followed the communist Left) in village communities; in the 1950s and the 1960s the rural to urban migration continued with increased force, marking the gradual demise of the agrarian economy rather than the attractiveness of urban industries, as witnessed by the parallel process of an equally important emigration wave from Greece to the Fordist labor markets of Western and Northern Europe; in the 1990s a very large and spontaneous immigration wave from post-communist Balkan countries—mainly from Albania—have brought, within a decade, the city’s immigrant population from less than 2 percent to almost 10 percent of the total in 2001; and, more recently, “undocumented” refugees from war zones in the Middle East and elsewhere are also gaining access to the city in perilous ways. Today Athens sprawls across the central plain of Attiki, which encompasses the most densely populated region in Greece with a metropolitan population of about 3.8 million people. A rather short period of industrialization and the limited presence of large-scale manufacturing, particularly in capital goods production, has deeply affected the urbanization as well as the structuration of urban society in Greece (Lipietz 1987, Tsoukalas 1987). Industrial capital has

258

Residential Segregation in Comparative Perspective

historically been weak and less internationalized than financial, shipping and commercial capital (Mouzelis 1986, Tsoukalas 1987). On the eve of the twentieth century, the Greek Diaspora, consisting of commercial capitalists, bankers and ship-owners, directed its investment to the city center, symbolically illustrating its economic power, while, in a speculative spirit, landowners and local lower middle classes were exercising pressure for urban expansion (Burgel 1976). An industrializing process was sustained only during the decade preceding World War II and the first three post-war decades. In the early pre-World War II period minimal urban and transport infrastructures were laid and manufacturing activities were scattered along the railway lines to the port of Piraeus and on the western periphery of the city. Working class settlements grew out of selfpromoted housing and accommodated the incoming refugees from Asia Minor, at a distance from middle class central city areas and suburbs (Leontidou 1990). During the first post-war decades, Athens became a large territorial pole, monopolizing investments in manufacturing, transport and trade and consolidating the centralized political and administrative functions of the state. The construction industry was booming to supply the houses and infrastructures for rapid urban growth (Economou 1988). Small-scale production and informality facilitated the reproduction of capitalist relations until the late 1970s when the international economic crisis initiated de-industrializing trends (Hadjimichalis 1987, Leontidou 1990, Vaiou and Hadjimichalis 1997). The 1980s signaled a process of restructuring and a definite shift to the service economy, with tourism, trade and communications becoming the main vehicles of growth. In fact, the limited presence of large industrial firms has prevented Athens, and the other major Greek cities from becoming spatially organized on the industrial principle with activity zoning, collective housing for the labor force, planned urban transport, etc. On the social level the existence of only a handful of large firms has affected the class structure by providing a rather weak representation of industrial capital in the bourgeoisie and an equally small segment of industrial workers in the broader working class (Leiulfsrud et al. 2005). At the same time, the local orientation of the fragile industrial economy contributed to establishing a social regulation model that comprised less state intervention in the form of welfare provision and a considerable organizational input from the family acting as a functional equivalent of the welfare state (Maloutas and Economou 1988, Allen et al. 2004, chapters 4 and 5). The whole post-war period has been one of unprecedented social mobility, marking the rapid transformation from an agrarian to a service economy. Rural migrants and working class offspring did not contribute greatly to reproducing and stabilizing the working class; they rather fuelled intermediate occupational categories ranging from traditional petite bourgeoisie positions to public employment and from lower service to professional jobs. They contributed, in fact, in sustaining an occupational structure with an expanded and diversified middle, i.e. a non-polarized structure.1 Athens played a major role in this process as the par excellence space that engendered social mobility in the post-war period. On the one hand, the expansion of small firms and self-employment spread as both the secondary and the tertiary sector created porous lines along the working class and the lower middle classes. On the other, Athens as capital city expanded the public sector, distributed services, trade and communications, and massively provided higher education. Segregation in Athens has developed within an urban context that has started to be ethno-racially diversified only in the last 20 years; a context characterized by substantial social mobility—that declined in the 1990s and has remained severely checked since the end of the twentieth century— and by a non-polarized social structure reproduced by a family-centered welfare system (Maloutas 1  Greece has been for a long time, and remains, the OECD country with the highest percentage of selfemployed (UNECE 2003).

Social Polarization and De-segregation in Athens

259

2007a and 2010). At the same time, the material locus of this context (i.e. the built environment) has almost exclusively been produced recently and, to a large extent, in unplanned ways through the aggregation of haphazard, individual, piecemeal and small-scale operations (Economou 2004). In terms of housing, in particular, Athens is completely devoid of large housing estates since there is no social rented sector, while low-cost home ownership social housing projects are also scarce. The effect of neoliberal globalization on Athens has been comparatively small until recently. The international role of the city is shaped by historically established channels in the Balkans and the Near East (Beavestock et al. 1999, European Commission 2004, Taylor and Derruder 2004) which were rediscovered in the process of the European Union enlargement and liberalization of capital markets. Athens connects to its neighbors and cities like Bucharest, Sofia or Istanbul. With the end of the Cold War the internationally-oriented fractions of capital, traditionally formed out of the leading financial and shipping capital and increasingly today out of telecommunications, were trying to redefine their role in the region by adjusting to the neoliberal agenda of the European Monetary Union (EMU), ‘the Growth and Stability Pact’ and the ‘Lisbon Strategy’ (a detailed exposition of the neoliberal agenda in Herman 2007, Bieling 2003 and Featherstone 1999 with reference to Greece). In the 1990s, becoming part of the European Monetary Union was the cornerstone of Greek policies, rolling out the institutional mechanisms for the liberalization of financial markets and the privatization of public companies and services. Hosting the 2004 Olympic Games in Athens became the flag for consolidating public consent and for providing local linkages to the dominant fractions of capital (Stathakis and Hadjimichalis 2004, Beriatos and Gospodini 2004, Maloutas et al. 2009). As Peck, Theodore and Brenner (2009) highlight, political mediation, contextual embeddedness and path-dependency shape the outcome and contradictions of urban neoliberalization. In the Greek case, the Olympics offered ample opportunities for cronyism in the exploitation of EU funds, public assets and property, actually advancing oligopolistic and politically collusive private capital in banking, media and telecommunications, transport and constructions (“intermeshed interests” in popular discourse). Although many of their structural powers were undermined, small landowners, estate agents and developers profited from concessions regarding the dispersal of projects, the expansion of private credit, and the tolerance of tax evasion and undocumented labor. Under favorable conditions in international financial markets a moderate acceleration in economic growth rates during the same period sufficed to advance the city’s position within South-Eastern Europe but not to converge with metropolitan areas of the European core (European Commission 2004, Kandylis et al. 2008). International transactions and economic restructuring concerned a handful of firms mainly in banking and finance, telecommunications and transport, which expanded to the Balkans and the Near East (IOBE 2006). Yet, Greek capital is a secondary, although important player, in the wider region. Moreover, about 10 percent of Greece’s outward FDI is directed to Cyprus for real estate and financial intermediation while another 10 percent is stocked at offshore financial centers and Luxembourg. Thus, the prospects for the city as an international business center are limited, because its financial transactions mainly consist of speculative and tax-evasion activities. With no traditional heavy industry and with no considerable local markets; with most economic activities being introverted and regulated in complex ways; with an urban landed property that is extremely segmented and with a work-force that does not create a major advantage in any particular sector, the situation of the Greek capital did not attract much inward foreign direct investment.2 2  Inward FDI has been fluctuating around 1 percent of the GDP since the 1980s; the largest investments were directed to energy, telecommunications, the food and the cement industry and distributive services.

260

Residential Segregation in Comparative Perspective

The international restructuring of the retail sector has, however, had an important impact on the local retail market and distributing services, especially in food and clothing. In particular, the arrival of international firms and the setting up of very large and geographically dispersed outlets has produced a new socio-spatial landscape in Athens. The suburbanization of large outlets, shopping centers and malls, has created significant economic activity poles in competition with the traditional center of Athens; a process which was reinforced by the dispersion of sports and recreation facilities for the Olympics (Beriatos and Gospodini 2004). Large-scale building projects of this kind strengthened capital concentration in the construction industry and family-based firms increasingly relied on subcontracting and the utilization of cheap migrant labor. During the same period the transformation of Greece into a migrant-receiving country reflects both the dynamics of domestic socio-economic change as well as the changing nature of global migration. The Greek experience of migration reinforces King’s (2000) identification of a distinct Southern European immigration model. However, immigration to Greece is overwhelmingly related to geographical proximity. It is linked to the process of transition of Eastern Europe and it originates from countries aspiring to be part of the European Union (Labrianidis et al. 2004, Cavounidis 2002). The societal impact of such complex processes diverges from theoretical expectations concerning the articulators of capital in the world economy (Maloutas 2007a, Arapoglou and Sayas 2009). The non-contextualized schema of polarization in global cities (Sassen 1991, Mollenkopf and Castells 1991) forecasts the decline of Fordist occupations, including the diminishing of both the working class and intermediate categories, like clerks and salespersons, in the organization of production and distribution of goods. In contrast, the Athenian case highlights the significance of the construction industry for stabilizing the size of the working class, while changing its ethnic composition through the recruitment of male migrant labor. It also shows that the shrinkage of middle strata does not apply to intermediate and lower-middle positions in administrative and distributive services, which here expanded and gave rise to female employment. The decline of the middle strata mainly concerns the exit of small entrepreneurs and family businesses from a variety of industries and particularly retail trade where they are unable to compete with outlets and franchises of international distributors. Moreover, professionalization and the provision of skilled services to business is moderate and does not constitute a significant route for upward mobility or a channel for the inflowing of a high profile workforce and polarization. Yet, there is an expansion of low status services related to the commodification of leisure, personal and social care; migrant women undertake a significant but not exclusive portion of casual employment in such services. Athens fails also to confirm the “dual city hypothesis” in spatial terms, with international elites isolated in exclusive spaces and the service proletariat confined to ghettos, barrios and deprived areas (cf. Mollenkopf and Castells 1991, Sassen 1991). In the sections to follow we depict a more complex picture of changing social distances and proximities in urban space. We report the lessening of overall levels of segregation due to the inflow of migrant workers into traditional middle class areas and to dispersal of lower middle occupations across the city fabric. Greek industrial workers and upper middle class categories are spatially distributed in very unequal ways, but not so pronounced as to threaten the cohesion of the city. It is only very recently that Greece has started tasting the impact of globalization in a more direct way, following its sovereign debt crisis and the agreement with its creditors on policies that are definitely oriented toward restructuring on a purely neoliberal rationale. The stability program for Greece includes not only a reduction in public expenditure and wages, but also flexiblization Greece became a net exporter of capital in the late 1990s, but FDI outflows never reached more than 2 percent of the GDP.

Social Polarization and De-segregation in Athens

261

of labor markets and a set of reforms in health care, pensions, education, and local government. This means that several barriers and disincentives to foreign and domestic large investments may be rapidly undone including all kinds of protective measures that nurtured the high social mobility, the reproduction and consolidation of intermediate social strata as well as their spatial corollaries, i.e. the relatively limited urban segregation. It is not surprising that the government’s strategy relies on moralizing discourses for a ‘clean start’ to impose market rules and secure the support of supranational institutions (EU and IMF). It is certainly too early to assess what these new policies, which reflect a higher globalization pressure, will eventually bring in terms of socio-spatial changes; the policy direction is quite clear, but outcomes are never predetermined or inevitable. Segregation Patterns and Measures Segregation may be relatively reduced in Athens, but it clearly exists. There are wealthy neighborhoods, and poor ones; and there are neighborhoods with high and low percentages of immigrants and refugees (see Figures 12.1 and 12.2).

Figure 12.1 Census tracts in the metropolitan area of Athens3 according to the percentage of Large employers and higher grade professionals and managers (ESeC 1) and Lower technical and routine occupations (ESeC 8 and 9)4 within the active population (2001) Source: EKKE-ESYE (2005). Only CTs with at least 400 economically active inhabitants are represented.

3  Although there is no official definition, the metropolitan area of Athens can be taken to include 125 municipalities (local government entities) with a population of about 3.8 million and extending approximately over 3,800 km2. The central city of Athens comprises the largest municipality with 745,000 inhabitants. 4  European Socio-economic Classes (ESeC) are a socio-economic classification model that has been elaborated as the successor of EGP classes (see Rose and Harrison 2007).

262

Residential Segregation in Comparative Perspective

Figure 12.2 Census tracts in the metropolitan area of Athens according to the percentage of immigrant and native population (2001) Source: EKKE-ESYE (2005). Only CTs with at least 500 inhabitants are represented.

Figure 12.1 shows that in the great majority of census tracts the percentage of large employers, higher grade professional and managers (ESeC 1) is close to the metropolitan average; this percentage is combined with variable percentages of lower technical and routine occupations (ESeC 8 and 9) that follow, however, a roughly contrasting pattern. There is a limited number of census tracts where these two socio-economic groups live quite isolated from each other, while cases of a polarized social structure (i.e. high concentration of both) are also rare. Figure 12.2 indicates that immigrants are quite evenly distributed in metropolitan space, with few exceptions of predominantly native Greek areas and even fewer areas where immigrants’ presence is significantly higher than its metropolitan average. Types of Segregation Social segregation  Before the 1990s, residential segregation in Athens existed only in terms of class—apart from the long-lasting segregation of Roma groups who continue to suffer a high level of social and spatial marginalization. The higher occupational categories traditionally resided in and around the city center, while the working class was mainly located at the periphery, and especially in the western part. Since the mid-1970s the geography of social segregation in Athens has started to change (Maloutas 2000). The new generations of the numerically expanding higher and uppermiddle occupational groups opted massively for residence in the north-eastern and south-eastern suburbs; this suburbanization trend has continued during the first decade of the twenty-first century. At the same time, the shrinking native working-class (see Figure 12.3) has been residentially much less mobile and has remained increasingly within traditional working-class strongholds (i.e. the working-class suburbs on the western part of the metropolis; Maloutas 2004).

Social Polarization and De-segregation in Athens

263

Figure 12.3 Percentage of major occupational categories in the active population in Athens (1961–2001)

Source: Leontidou (1986: 97) for 1961 to 1981 and EKKE-ESYE (2005) for 1991 and 2001.

Map 12.1 shows a synthetic image of class segregation in Athens for 2001. A classification of the city’s census tracts into five different social types was chosen to convey this image with rather clearly demarcated class features between the different types/clusters (see Table 12.1).5 According to this classification, 10 percent of the city’s residents live in areas dominantly inhabited by higher, and 22 percent by higher and intermediate occupational categories (clusters 1 and 2), while 17 percent reside in areas mainly inhabited by lower ones (cluster 5). The other half live in socially mixed areas (clusters 3 and 4), which sometimes appear slightly polarized as they contain increased shares of both higher and lower occupational categories (cluster 3). The location of the two higher clusters covers most of the eastern part of the city, including the traditional strongholds of the upper socio-economic classes in the city center. The lower cluster 5  We constructed a very broad social typology of residential space in Athens using a combination of a binary correspondence analysis on 17 class variables to determine the main social axes of their spatial variation, and a K-means clustering to classify the city’s 3,500 census tracts (average population: 1,000; average active population: 450) according to their position on the axes that resulted from the correspondence analysis. The class variables are a disaggregated version of the nine fundamental ESeC classes that (a) differentiates between managers and professionals; (b) uses separate categories for large employers and artists; and (c) distinguishes the members of intermediate and lower socio-economic classes employed in industry, construction and transport from those working in the services. Agriculture-related occupational categories have not been taken into consideration due to their limited numbers and their wide spatial variation, which is not relevant to the present analysis. The variables used are directly available in the database application EKKE-ESYE (2005).

264

Map 12.1

Residential Segregation in Comparative Perspective

Social types/clusters of residential areas in Athens, census tract level (2001)

Source: Maloutas (2007a: 753).

covers the western, working-class part of the city and most of the residential areas very distant from the center. The slightly polarized cluster is located around the center in areas that are densely populated, and the mixed cluster is mostly, but not exclusively, located in the best residential areas of the western (working-class) part of the city (see Map 12.1), which gather the socially –but not residentially– mobile offspring of the local population (Maloutas 2004).

18 24 20 21 15 20

50

34

30 19 13 25

Source: EKKE-ESYE (2005).

Higher (cl 1) Higher and intermediate (cl 2) Polarized (cl 3) Mixed (cl 4) Lower (cl 5) All clusters

Intermediate and lower technician occupations

Large employers, professional, administrative and managerial occupations

11 17 16 16

16

15

Small employer and self-employed occupations

19 24 26 21

16

9

Lower white collar and skilled occupations

21 19 31 18

11

8

Routine occupations

6 46 17 100

22

10

Percent distribution of the active population

Social typology of residential space in Athens (2001), cluster composition in terms of major occupational groups (values above average in bold)

Social Clusters (occupations)

Table 12.1

Residential Segregation in Comparative Perspective

266

Ethno-racial segregation  Immigration flows that led to noticeable ethno-racial diversity in Athens started in the early 1990s. Among the immigrant groups, Albanians are by far the largest and most of the others are also from neighboring Balkan countries (see Table 12.2). It was a spontaneous immigration wave that infiltrated Greece through its long northern border. During the last decade the flows have changed with more asylum seekers from the broader Middle East gaining access through Turkey and the Aegean. Table 12.2

Percentage of the 10 major groups within the immigrant population in Athens (2001)

Country of origin

%

Albania Poland Russian Confederation Bulgaria Romania Ukraine Pakistan Turkey Egypt Philippines

51.1 2.8 2.7 2.6 2.4 2.3 2.1 1.9 1.6 1.5

Source: EKKE-ESYE (2005).

Before these flows, official statistics showed that immigrants came mainly from developed economy countries, usually occupied high status jobs and resided in high status neighborhoods. In 2001 things had radically changed: immigrants in Athens represented 10 percent of the total population—against 1.9 percent in 1991—and 12.6 percent of the active population; and 82 percent of the immigrants in 2001 originated from former socialist or from developing economy countries (EKKE-ESYE 2005). The large immigrant inflow of the 1990s also radically changed the spatial distribution of immigrants’ residential location (see Map 12.2). Immigrants were mainly situated in the densely built areas around the center, but also in many areas of the distant periphery. They were relatively absent from the ‘first ring’ bourgeois (and petit-bourgeois) suburban zone as well as from the traditional working class suburbs. The aggregate pattern of immigrant location is dominated by the pattern of the Albanians; other groups have very diverse patterns, depending on their economic activity. Immigrants from the Indian peninsula, for instance, are usually located on the north-south highway axis, away from the center, since this brings them near their workplaces in transportation, warehousing and agriculture, while those from the Philippines and Sri Lanka—mainly women working in domestic services—are often found in high status suburbs or in areas with direct access to these suburbs (Maloutas 2007a, Kandylis et al. forthcoming).

Social Polarization and De-segregation in Athens

Map 12.2

267

Relative concentration of immigrants in the residential space of Athens by census tract (2001)

Source: Maloutas (2007a: 748).

The Dynamic of Segregation The first question is whether residential segregation in Athens, according to class or ethnicity, is increasing or decreasing. Is empirical evidence from Athens corroborating the theoretical expectation of the dual or divided city that is increasingly polarized socially as well as spatially? The Dynamic of Class Segregation In order to evaluate the dynamic of residential segregation in terms of class, we have compared indices of dissimilarity for some major occupational categories6 according to their residential location, between 1991 and 2001 (see Table 12.3). 6  Due to important changes in the occupational categories used in the Greek censuses of 1991 and 2001, we have been obliged to use sub-categories that remained constant and recompose partly the aggregate major categories.

268

Table 12.3

Residential Segregation in Comparative Perspective

Indices of dissimilarity (ID) for occupational categories in the Athens Metropolitan Area; census tract level in 1991 and 20017 1991 ID

Professionals (legal professionals; medical doctors and dentists;   architects, urban planners, civil engineers) Teachers (primary and secondary education) Managing proprietors and managers of small businesses in   commerce, hotels and restaurants Technical assistants (in natural science, engineering, biology and   architectural design); nursing professionals Office clerks (accounting, cashiers, inventorying, libraries,   mailing, telecommunications (telephone operators), travelling,   transport, other) Service providing personnel (cooks and waiters; hairdressers,   cosmetics counsellors, other practitioners in similar services;   fire persons, police and other security personnel) Sales persons in shops Skilled workers in building Skilled workers in industry (cabinet makers; textiles and garment   industries; machine tool operators in chemical industries) Unskilled workers (sales persons in open markets and door to   door; domestic personnel, cleaners etc.; garbage collectors, road   cleaners, unskilled workers in building, industry and transport) Unemployed

2001

count

ID

count

36.1

47852

35.1

55543

20.2

42533

20.2

45085

13.3

97141

15.0

73298

25.0

15853

22.4

23737

12.7

142992

11.2

147803

13.8

62388

14.3

75072

13.7 27.0

58867 51821

13.3 24.9

92981 52315

23.1

51918

26.7

30023

25.1

55261

22.3

118960

12.9

107061

12.8

146846

Source: EKKE-ESYE (2005).

Table 12.3 shows, first of all, that indices of dissimilarity are rather low and have remained stable or decreased during the 1990s. The few exceptions (small employers and skilled workers) consist of numerically shrinking and aging categories that usually tend to regroup in spaces where their presence was high. On the other hand, the table shows clearly that higher occupational categories present the highest indices, followed by the working-class, while intermediate categories and lower intermediate positions in the services are least segregated. Very low segregation levels characterize the unemployed, indicating that unemployment is less related to space than it is to class. Table 12.4 shows the difference between the dissimilarity indices for each pair of major occupational categories in 1991 and 2001. In this case, indices have not been calculated for each occupational category in respect to the sum of the economically active population—as in Table 12.3—but in respect to each other. The table shows that the aggregate stability or decrease shown in Table 12.3 is structured by lower order relations in a rather systematic way: The higher and upper-intermediate categories have increased their mutual spatial distance and the distance to all other categories except the skilled workers in building and the unskilled workers; the latter have reduced their distance to all other categories. The overall tendency is, therefore, fuelled by 7  The islands belonging to the region have been excluded as well as census tracts with less than 600 inhabitants; 2.701 CTs were used for 1991 and 3.129 for 2001.

Social Polarization and De-segregation in Athens

269

the combination of more segregation among the higher and upper-intermediate categories with desegregation among the lower occupational categories. Table 12.4

Percentage point difference of dissimilarity indices between occupational categories (1991–2001), positive values (in bold) denote the increase of spatial distance 1

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 11

2

3

4

5

6

7

8

9

10

11

1.8

0.6 4.9

2.2 7.5 1.3

4.9 5.2 1.9 -0.5

2.8 7.5 1.6 1.0 -1.1

1.5 5.5 1.9 1.1 -2.6 -1.3

1.3 7.2 0.4 0.5 -2.4 -2.7 -1.2

-3.6 -0.5 -2.5 -0.9 -7.8 -5.3 -4.3 -3.8

1.1 5.1 2.1 3.2 -2.6 -0.5 1.4 1.6 2.7

-6.2 -2.6 -2.6 -1.7 -7.9 -4.7 -4.2 -2.4 -3.2 1.1

Professionals Artists Teachers Managing proprietors and managers of small businesses in commerce, hotels and restaurants Technical assistants and nursing professionals Office clerks Service providing personnel Sales persons in shops Skilled workers in building Skilled workers in industry Unskilled workers

Source: EKKE-ESYE (2005).

The Dynamic of Ethno-racial Segregation There is no reliable empirical evidence, at present, concerning the trend of ethno-racial segregation in Athens. Census data for 1991 refer to a small number of immigrants and do not differentiate among ethnicities or racial characteristics. The 2001 census will remain the only reliable source for the spatial distribution of ethnic groups until the 2011 census becomes available. It is, therefore, difficult at this point to evaluate the segregation trend for Athens in ethno-racial terms. There is, however, some indirect evidence, which refers to the desegregation impact that the presence of immigrants has produced in Athens during the 1990s in class terms. Table 12.5 shows that, when immigrants are taken into account, the level of segregation for most class categories is reduced. Since immigrants in Athens have almost exclusively occupied lower status jobs, the effect of their presence was to desegregate the lower occupational categories. This assumption is corroborated by the data in Table 12.4, where desegregation is clear for workers in building and for unskilled

270

Residential Segregation in Comparative Perspective

workers (i.e. jobs very frequently occupied by immigrants), but not for skilled workers in industry where jobs remained in the hands of natives as they shrunk in number. It seems, therefore, that the presence of immigrants has not boosted the levels of occupational segregation in Athens, contrary to the theoretical expectation that the precarious status in the labor market entails spatial marginalization as well. Table 12.5

Dissimilarity indices (ID) for ESeC classes in Athens (2001), total active population and Greek population IDs in respect to the residential location of higher grade professionals

(1) Higher professionals and administrators/   managers (2) Lower professionals and administrators/   managers (3) Intermediate (4) Small self-employed (5) Self-employed in agriculture (6) Lower supervisory and lower technician   occupations (7) Lower services, sales and clerical   occupations (8) Lower technical occupations (9) Routine occupations

Within the total active population (1)

Within the Greek population (2)

(2)–(1)

4.4

4.4

0.0

16.8

17.0

0.2

24.1 29.1 59.6

24.2 29.6 64.9

0.1 0.5 5.3

29.6

29.6

-0.1

31.9

32.4

0.4

40.8 39.6

42.2 42.7

1.4 3.2

Source: EKKE-ESYE (2005).

This kind of evidence seems to run against the evidence of apparently high levels of segregation between individual migrant groups and the native population as well as among the different migrant groups in 2001 (see Table 12.6). It should be noted, however, that the dissimilarity index for the very large immigrant group (i.e. the Albanians) is low (0.33). Indices for the rest of the groups vary between 0.50 and 0.83, with the highest scores for groups from the Indian peninsula and the Middle East. But, even these very high indices do not imply the existence of group ‘ghettoes’, in the sense that there are no areas—even at census tract level—where one of these groups forms the majority of the population; and, with very few exceptions, this is also true for the immigrant population as a whole (see Figure 12.2). The high dissimilarity indices indicate, in fact, the tendency of these groups to coalesce in space, due to the ways they operate in the housing market (i.e. through kinship and friends) and to the need for solidarity networks that depend on spatial propinquity. At the same time, due to their small numbers, these groups are almost non-existent in most of the city’s census tracts, and this increases substantially their dissimilarity index (see Dominguez et al. in this volume, chapter 10). On the other hand, the high dissimilarity indices among migrant groups show that individual ethnic concentrations have very diverse patterns. Migrant enclaves

Social Polarization and De-segregation in Athens

271

in a limited number of city quarters have a plural ethnic composition and rarely does one group dominate an area. Table 12.6 Country of origin Greece Albania

Dissimilarity indices for selected immigrant groups in the Athens Metropolitan Area (2001) in respect to the Greek population and to each other AL

BG

DE

UK

IRQ

MLD

UKR

PAK

P

RO

RUS

SYR

PHI

33

51 46

57 65

56 64

83 80

65 60

51 48

75 71

62 53

57 51

55 59

70 62

79 77

65

64

82

56

46

80

49

53

65

64

72

52

89

72

65

87

74

74

73

81

76

90

71

62

89

75

74

72

80

72

85

83

79

82

81

87

80

91

57

83

58

60

72

67

74

52

54

62

63

67

50

84

74

79

83

91

54

71

60

77

65

65

78

75

82 80

4523

6063

Bulgaria Germany UK Iraq Moldavia Ukraine Pakistan Poland Romania Russia Syria Count

206446 10645

5568

9016

6416

3865

9461

9639

11200

9960

11277

Source: EKKE-ESYE (2005).

The Spatial Forms of Segregation Neighborhood Segregation Due to its contextual origins and history, the concept of segregation implies mainly—if not exclusively—neighborhood segregation (see Maloutas and Fujita in this volume, chapters 1 and 13 respectively). The social profile of neighborhoods in Athens developed in the following way: the higher social categories resided in and around the city center from the 19th century to the mid 1970s, when they started gradually to suburbanize following the rapid degradation of living conditions in central areas (Leontidou 1990, Maloutas 2007a). The working-class and other lower occupational categories increasingly expanded the city’s periphery, especially on the western part of the agglomeration, as they massively migrated from rural areas and smaller cities during the first three post-war decades. The suburbanization of higher and intermediate social categories, that has continued for more than 30 years, has gradually reshaped the social map of the city: from a contrast between a rich center and a poor periphery to the contrast between wealthy enclaves in the east and in parts of the center and lower status areas in the western immediate periphery as well as in most parts of the outer periphery of the metropolis (see Map 12.1). Another feature of neighborhood segregation in Athens is the rarity of contiguous socially contrasted residential areas. The rather smooth transition from one extreme social type of area to

272

Residential Segregation in Comparative Perspective

the opposite is related to the piecemeal constitution of the urban tissue and to the small scale of operations that did not favor the emergence of important social breaks between neighboring spaces. In terms of ethno-racial neighborhood segregation, immigrant groups are mostly located around the city center, and to a smaller degree in the relatively distant periphery. Their presence in these areas is often considered as dominant, even though their numbers do not corroborate such claims. It is mainly the rapidity with which migrant presence has developed in these downgrading, formerly middle-class areas, and the very visible difference of some groups from the native population that have assisted the perception of immigrant presence as dominant. Immigrant presence is concentrated in most of the overbuilt and downgrading areas around the city center (see Map 12.2). Deterioration in terms of the built environment and infrastructure (that had started well before the immigration inflow) has contributed to another dominant perception: that of a causal link between immigrants’ settlement and degradation. At the same time, immigrants in these areas are subject to a systematic spatial hierarchy that is referred to as ‘vertical segregation’. Vertical Segregation Vertical segregation or vertical social differentiation are terms used to denote the social and ethno-racial hierarchy within the small condominiums of five to eight floors, that are the virtually exclusive house building form in the densely built areas around the city center. Apartments within these condominiums have different attributes according to floor: on the lower floors apartments are smaller, darker, noisier and usually destined for the private rented market; on higher floors, they have more space, light, view and less noise, and they are usually owner-occupied. This systematic

Figure 12.4 Percentage of households in the Ampelokipi neighborhood by social rank of head of household and floor of residence (1998) Source: Maloutas and Karadimitriou (2001: 706).

Social Polarization and De-segregation in Athens

273

Figure 12.5 Percentage distribution of Greek and foreign citizens in the Ampelokipi neighborhood by floor of residence (1998) Source: Maloutas and Karadimitriou (2001: 706).

differentiation of apartment attributes and quality in respect to floor level has produced over the years a vertical social and ethno-racial hierarchy within the housing stock of these neighborhoods. Figures 12.4 and 12.5 show the clear social and ethnic hierarchy within the condominiums of a typical overbuilt neighborhood (Ampelokipi) near the city center.8 Vertical segregation is a process that counteracts further development of neighborhood segregation (Leontidou 1990, Maloutas and Karadimitriou 2001), especially since it induced a large part of migrants with lower status jobs to live in the same neighborhoods as middle and lower middle-class natives. Gentrification The process of gentrification is related to segregation in an ambivalent way. In one sense, it leads to more segregation since it eventually takes residential areas away from their former inhabitants— the working-class and other lower class categories—to be regenerated and appropriated by uppermiddle class categories. In another sense, gentrification desegregates, at least for some time, since it increases the social mix in formerly rather homogeneous working-class areas. Athens did not follow the model of the Anglo-American industrial city, where the elites left the city center and eventually returned in the post-industrial era through the process of gentrification. The ‘belated’ suburbanization in Athens is still ongoing and the large wave of immigration was mainly directed to the lower status areas in and around the city center that would, otherwise, be potential targets for gentrification. It is not surprising, therefore, that until 2001 there was no trace of neighborhood 8  Areas of ‘vertical’ segregation comprise almost one third of the city’s population (Maloutas and Karadimitriou 2001).

274

Residential Segregation in Comparative Perspective

profile change in the center of Athens that could be associated with gentrification processes (Maloutas and Alexandri, forthcoming). It is, however, true that during this decade—especially following the successful unveiling of the center’s potential after the Olympic Games in 2004—the concern for what happens in the city center has dramatically increased and the trend within some of the downgraded central neighborhoods may be pointing toward gentrification (Alexandri and Tzirtzilaki 2010). And, although the scale of intervention remains small and individualized, the political climate that links the problems in the downgraded neighborhoods to the presence of immigrants (Pavlou 2009, Kavoulakos and Kandylis 2012) may work as a catalyst in this direction. The Major Factors Affecting Segregation Patterns and Intensity There are a number of important factors that shape the form and level of urban segregation in capitalist societies. The labor market creates and reproduces unequal means of access to different residential locations. The accessibility to these locations is also dependent on the hierarchy of places created and reproduced by the land and housing markets. On the other hand, policies usually act to change the accessibility of financially and otherwise unequal households to different segments of land and housing supply defined by the market. This effect varies from tempering the problems of adjustment between supply and demand—by providing, for instance, social rented housing—to exacerbating housing inequalities or reproducing discrimination in terms of race or ethnicity—by allowing or not opposing, for instance, practices like redlining or blockbusting. Broader urban policies, may also have significant effects on segregation, sometimes unintended. The Labor Market During the first four post-war decades the labor market in Athens offered substantial opportunities that led to a prolonged wave of social mobility, enabling rural migrants and working-class offspring to access intermediate occupational positions (Tsoukalas 1987). This has produced a social structure relatively expanded in the middle with numerous selfemployed positions, many of them being crucially integrated in spatially diffuse manufacturing networks (Vaiou and Hadjimichalis 1997), large numbers of professionals even in the prestigious professions (medicine, law, engineering) and large numbers of positions in public employment. Thus, the polarization that could be assumed from Figure 12.3 is, in fact, misleading. A large part of the higher and upper-middle occupational positions are in reality nearer the middle of the social hierarchy than the top. Professionals and managers in Athens are rarely part of the corporate elite, and an even smaller part is part of the international corporate elite. Therefore, the expansion of the higher pole of the occupational hierarchy has not substantially affected the housing market in the ways that an expanding demand for top quality housing usually does (pressure for gentrified spaces, for instance). As mentioned above, the lower occupational pole has, in contrast, been shrinking since the 1960s (Figure 12.3). During the 1990s the immigration wave significantly reduced the shrinkage of the lower occupational pole and restructured its content at the expense of skilled jobs in industry and in favor of routine jobs, especially in service industries. In principle, this change at the lower pole could have produced more segregation; but, in fact, it did not because the affordable housing stock was situated in socially mixed and densely built areas around the center, which complemented the city’s ‘vertical’ segregation in ethno-racial terms.

Social Polarization and De-segregation in Athens

275

The Housing Market and the Built Environment The housing market and the broader housing allocation system have largely configured the city’s segregation level and pattern. During the first two post-war decades, social mobility was related to the large immigration from rural areas. The South European type of metropolitan labor market— i.e. not based on industrial development—made housing the pivotal element for integrating the newcomers into urban societies (Allen et al. 2004, chapter 3). At the same time, the lack of industrial development, as the fundamental principle for urban organization, reduced the pressure for collective working-class housing from all the agents involved and led to the provision of low cost housing in various forms. In Athens, low cost housing was provided by individual operations of self-promotion and, often, irregular self-construction. This substandard housing was similar to housing in rural areas and had the advantage of not demanding regular payments (rent) that the settlers’ position in the labor market could not guarantee. In the 1950s and 1960s almost half a million people became homeowners at the city’s periphery in this way (Maloutas 1990). A second housing system—under the name antiparohi—provided a large volume of low cost housing in the 1960s and the 1970s (Antonopoulou 1991). Through this system, that literally filled the areas around the center with small condominiums, there was abundant provision of modern apartments in relatively low prices. Operations were also small scale (rarely exceeding one building) and entailed the joint venture of a small landowner and a small builder who got a share of the built property at the end of the works, according to an initial agreement. The popularity of this system was due to generous tax relief that made any other form of condominium production not competitive. Almost 35,000 units of five floors or higher were produced between 1950 and 1980 in a city that, before that period, had less than 1,000 such buildings (Maloutas and Karadimitriou 2001). Both these house provision systems have greatly affected the form and level of segregation. The traditional working-class areas in the western periphery of the city were developed and consolidated through the first method that gave them a systematically inferior built environment and has ‘entrapped’ several generations of lower occupational categories in substandard housing as homeowners in those areas. This entrapment refers both to the low residential mobility that homeownership entails in a rather inflexible housing market (where every transaction is heavily taxed) as well as to the formation of family self-help networks that were greatly facilitated through this house provision system. At a later stage, the new generations in these traditional working-class areas usually avoided relocation to other areas in order to remain part of their family’s network, even when they were socially mobile (Maloutas 2004; see also Dominguez in this volume, Chapter 10 and Leal 2004). This was enabled by the growing provision of improved housing units in most traditional working class areas. In this sense, the self-provision system of the early post-war period has had a contradictory impact on segregation: it has increased segregation by producing and consolidating rather large working-class areas in the city’s periphery; on the other hand, it has produced a kind of built environment and a type of social relations within it that increasingly became a substantial impediment to further shifting and sorting. The impact of the second system on segregation worked in the opposite direction, but was equally contradictory. At first, it desegregated the center, where the higher and intermediate social strata were initially located, by massively providing low cost and abundant housing. Following the overbuilding trend of the 1960s, the central municipality’s population increased by 50 percent with a considerable desegregating effect. However, the lucrative overbuilding of the central areas also led to rapidly deteriorating living conditions and, within a decade, led to the gradual suburbanization of higher and upper-intermediate social strata. The old resort areas and certain

276

Residential Segregation in Comparative Perspective

new suburbs that emerged as new residential locations in the eastern periphery have increasingly grown to be the most socially homogeneous parts of the city. The features of the city’s housing stock, produced mainly in the first three post-war decades, have also affected segregation in respect of the important immigrant wave of the 1990s. Due to the complete absence of social housing for rent, immigrants’ housing demand was primarily directed to the abundant stock of small apartments on lower floors in the densely built areas around the center, that remained underused for some time; moreover, immigrants were also driven to certain areas in the distant periphery that combined proximity to jobs in construction and in agriculture as well as to affordable, but very substandard housing stock. In both cases, segregation has not been increased by immigrant presence either because immigrants settled in small apartments within socially mixed areas in conditions of ‘vertical segregation’ or because they settled in small privately rented units in the periphery, often near the residential areas of their employers. The Welfare System and the Family The housing solutions that prevailed in post-war Athens were closely related to the South European type of welfare system characterized by the scarcity of direct state intervention in the provision of social services and the major role of the family as an organizing agent of social reproduction processes. In fact, the lack of social services induced an increased family role, which was assisted and promoted by a host of derogatory state policies and practices within the clientelist political system (Charalampis 1986, Petmezidou-Tsoulouvi 1987). Parties in, or aspiring to, power were after the family vote, which they tried to secure through targeted concessions and favors (i.e. hiring in the public sector, exceptional treatment in heath services, leniency regarding illegal housebuilding practices, etc.). In this way, families became an important locus for the accumulation of resources and strategic planning regarding their intergenerational distribution and use. This led to the formation of a family-centered urban society, with introverted solidarity and with a relatively atrophied appreciation of the public interest and estrangement from societal common goals. In terms of the spatial organization of the city, the importance of family solidarity is reflected in the extensive and active family networks, mainly created during the rapid urbanization period. Rural migrants followed the path opened by their predecessors to whom they were related by kinship and/or common origin. The aggregation of such movements created networks of proximity in the urban fabric, which developed further into networks of solidarity. Proximity to one’s parents and relatives remained the first consideration in the choice for residential location at least until the late 1980s (Maloutas 1990). Family networks were an impediment to segregation since, in an era of rapid urbanization and, therefore, great geographic mobility, they prevented migrants from changing radically their residential location once they settled in the city. Family networks thus prevented the free shifting and sorting process that the combination of social mobility and the housing market would normally entail (Maloutas 2004). This is true mainly for working-class and other lower occupational categories, but is much less so for higher and upper-intermediate categories. For the latter, family solidarity may be equally important, but it is much less dependent on spatial proximity since it can take a monetary form and, therefore, be despatialized. The massive suburbanization of the new generations of these categories since the mid-1970s—away, that is, from their centrally-located parental households—bears witness to the socially diversified impact of family networks on residential segregation.

Social Polarization and De-segregation in Athens

277

Urban Policies The main feature of urban policies in Athens during most of the post-war period has been the absence of a comprehensive and effective planning frame as well as the absence of large-scale operations that could reshape the course of processes like segregation. The urban fabric was largely shaped by the aggregation of the numerous unplanned and uncoordinated individual initiatives of old and new settlers (Economou and Petrakos 1997, Economou 2003, 2004). These initiatives were regulated through housing policies that, however, had only unintended outcomes on the city, as their main objectives were either macroeconomic (to help economic growth that was largely dependent on housing construction) or political (to facilitate access to homeownership). During the first post-war decades, the piecemeal character and very small scale of house-building operations were not only helped by the lack of large scale public operations, but were also intentionally protected against the potentially disruptive effect of large construction firms and private banks (Economou 1988, 2004). Things changed considerably during the 1990s, and especially after the city was chosen to organize the 2004 Olympic Games and started preparing for the event (Beriatos and Gospodini 2004, Maloutas et al. 2009). Large-scale projects emerged all over the city, mainly to redress the chronic deficit in urban infrastructures: they produced two modern metro lines, a new airport, a peripheral highway and a general upgrading of the urban highway system, two tramway lines, Olympic-size sporting facilities, the unification of major archeological sites in the city center for pedestrian use, etc. These large operations reshaped the map of the city and increased or revealed the potential of different areas within it, and especially of its center. The changes in urban policies during the 1990s—even if they were circumstantial—ended the slow pace that formed the city’s socio-spatial maze and used to obstruct in many ways the segregating, shifting and sorting of people and places. New metro lines and stations and new highways have reshaped the map of land and housing values in an uncontrolled way—i.e. the positive impact of the new infrastructures has been to the unrestricted profit of the adjacent landowners. The scale of the changes cannot yet be measured, but segregation pressure must have got stronger during the last 15 years due also to the abundance of low cost housing loans since the mid-1990s, combined with the rapid, and geographically differentiated, increase of land and housing prices (Emmanuel 2004). However, segregation pressure is resisted by processes that seem to run in the opposite direction. The improvement of the downgraded parts of the center and their potential gentrification, for instance, is opposed by the existence of a formerly largely abandoned housing stock that found willing users among the immigrants of the last 20 years. Such contradictory processes have put segregation and gentrification on the political agenda and have recently attracted media attention for the first time in the post-war period. The dominant discourse does not conceptualize these issues in terms of segregation or gentrification, but in terms of regeneration, upgrading, tourist development, safety, etc. presaging the kind of urban policies that may be expected. The post-war laissez-faire in urban development may be coming to an end as political support is shifting from measures that ensured the integration of large numbers of new settlers into urban society and the enhancement of social mobility, to the quality of urban life expected by the majority of social strata, i.e. by the broader middle-class groups, and by transient and less transient investors. However, this new policy direction may not be unaffected by the current severe crisis which undermines the reproduction of these broad middle-class groups.

278

Residential Segregation in Comparative Perspective

The Impact of Segregation Segregation is generally regarded as negative since it is part and parcel of social inequalities and discrimination. Its negative connotation is also related to the assumption that it is an important factor in the reproduction of social inequalities. The Importance of Residential Location or Neighborhood Effects There is a growing literature on neighborhood effects, i.e. on the importance of residential location in respect to life chances and social mobility (Buck and Gordon 2004, Ginther et al. 2000, Musterd et al. 2003, Ostendorf et al. 2001). The most interesting part of this literature refers to the overestimation of this effect that serves to legitimate the spatial redistribution of poverty or to legitimate gentrification in order to combat the negative effects of segregation. The authors who criticize this overestimation claim that, even though neighborhood effects may not be negligible, the importance of personal and household characteristics on their own is overwhelming. In terms of required policies this means that assistance to persons and households should have precedence over the spatial reorganization of social inequalities. The importance of neighborhood effects is not universally the same. On the one hand, neighborhood effects are socially diverse and for higher and middle social groups they are positive rather than negative.9 On the other hand, such effects are also geographically diverse. They are definitely far more important in South Chicago than in Stockholm’s Tensta. Athens probably lies somewhere in-between. It is less segregated than Chicago and not necessarily much more segregated than Stockholm, but it certainly lacks the quality and the socio-spatial distribution of Scandinavian welfare services. There are no systematic research findings on neighborhood effects in Athens; there are indications, however, of a higher social mobility for the same broad occupational categories as one ascends in the social hierarchy of residential areas, and a reduced mobility if one moves down the social hierarchy (Maloutas 2004). Segregated Services It is usually expected that a person or a household will move to a better residential area if their financial situation improves in order to escape the negative or to enjoy the positive neighborhood effects of a higher status area. In Athens, this theoretically-expected relocation movement is inhibited by the importance of family networks that induces, as in the rest of Southern Europe, low levels of residential mobility. In such contexts, the groups that cannot relocate in order to find the positive effects they seek, are developing alternative strategies that mainly involve exclusivity in services; and exclusive services for the higher and upper-intermediate groups usually implies segregated services for the lower end of the social hierarchy. Family strategies in terms of school choice function as an alternative to residential relocation near better schools and compensate, for several groups, for the lack of sufficient urban segregation in Athens (Maloutas 2007b). Strategies of exclusivity for the higher groups, that mainly involve opting for select private schools, trickle down a similar antagonistic spirit throughout the social hierarchy in terms of school choice (Van Zanten 2001). Eventually, this leads to very poor quality schools, segregated at the lower end. In this sense, the relatively reduced segregation of the lower 9  See Gordon and Monastiriotis (2006) on the greater importance—in terms of education—of positive effects for the middle-classes rather than the negative for the working-class.

Social Polarization and De-segregation in Athens

279

occupational categories and of immigrants in Athens may be a positive indication, but it does not necessarily imply more equity and less discrimination.10 Less spatial distance does not necessarily mean less social distance (Chamboredon and Lemaire 1970). Segregation, therefore, is not a problem that can be dealt with out of context; it needs to be addressed within the particular forms and processes it is associated with and in respect to the effective impact it has on inequalities and discrimination. Conclusion: Global and Local Parameters in Defining Socio-spatial Change in Athens Economic restructuring in Athens has been comparatively gradual and limited in its impact due to the archetypal southern European character of the city’s post-war development, where Fordist industry has never been the principal driving force. The steadily-growing tertiarization and the switch from international emigration in the first post-war decades to intense immigration inflows since the early 1990s have led to a certain degree of social polarization—which shows, however, substantial differences to the global city model. Nevertheless, Athens’ polarization has not led to more segregation, either in terms of class, ethnicity or race; on the contrary, segregation indices have decreased in the 1990s, even though residential neighborhoods at the social extremes have increased their social homogeneity and members of distinct immigrant groups—especially the smaller ones—tend to be concentrated in space following clear residential location patterns. Segregation indices have decreased in the 1990s even though, during the last three decades, the pattern of the city’s segregation has been changing in ways that usually favor an increase in segregation indices. The higher class categories have been steadily relocating to more peripheral locations in suburbs of increasing social homogeneity, while gradually abandoning the center where they clustered almost exclusively until the mid-1970s. At the same time, processes of vertical segregation, spatial entrapment of social mobility in traditional working-class areas as well as the embryonic state of gentrification, indicate that segregation in Athens is influenced by a host of different factors. The level and pattern of segregation in Athens are much more related to contextual parameters like the type of post-war urbanization, the structure of the built environment and the housing stock, the local model of the welfare system, the considerably extended period of social mobility and—at the same time—the low degree of residential mobility, and the clientelist policies favoring the reproduction and consolidation of intermediate social groups. Globalization has certainly affected the city’s socio-spatial condition, mainly through the large wave of immigration and, more recently, through the policies imposed by the EU and the IMF after Greece’s sovereign debt crisis. However, the effect of globalization on the city’s socio-spatial condition and on segregation in particular, cannot be read on the ground as it is certainly not dominant and the final outcome often runs against the theoretically expected direction.

10  See Arbaci (2007, 2008) for a discussion of reduced segregation that does not prevent discrimination in southern Europe

280

Residential Segregation in Comparative Perspective

References Alexandri, G. and Tzirtzilaki, E. 2010. Breaking Off the Boundaries in Athens’ City Centre. Paper presented at the European Network of Housing Research (ENHR) Conference: Urban dynamics and housing change, Istanbul, July 4–7. Allen, J., Barlow, J., Leal, J., Maloutas, T. and Padovani, L. 2004. Housing and Welfare in Southern Europe. Oxford: Blackwell. Antonopoulou, S. 1991. Postwar Transformation of the Greek Economy and the Residential Phenomenon. Athens: Papazisis (in Greek). Arapoglou, V.P. and Sayas, J. 2009. “New Facets of Urban Segregation in Southern Europe: Gender, Migration and Social Class Change in Athens.” European Urban and Regional Studies 16(4), 345–62. Arbaci, S. 2007. “Ethnic Segregation, Housing Systems and Welfare Regimes in Europe.” European Journal of Housing Policy 7(4), 401–33. Arbaci, S. 2008. “(Re)viewing Ethnic Residential Segregation in Southern European Cities: Housing and Urban Regimes as Mechanisms of Marginalisation.” Housing Studies 23(4), 589– 613. Beavestock, J.V., Taylor, P.J. and Smith, R.G. 1999. “A Roster of World Cities.” Cities 16(6), 445–58. Beriatos, E. and Gospodini, A. 2004. “‘Glocalising’ Urban Landscapes: Athens and the 2004 Olympics.” Cities 21(3), 187–202. Bieling, H.J. 2003. “Social Forces in the Making of the New European Economy: The Case of Financial Market Integration.” New Political Economy 8(2), 203–24. Buck, N. and Gordon, I. 2004. “Does Spatial Concentration of Disadvantage Contribute to Social Exclusion?” In City Matters. Competitiveness, Cohesion and Urban Governance, eds M. Boddy and M. Parkinson. Bristol: The Polity Press, 237–54. Burgel, G. 1976. Athens, the Development of a Mediterranean Capita. Athens: Exantas (in Greek). Cavounidis, J. 2002. “Migration in Southern Europe and the Case of Greece.” International Migration 40(1), 45–70. Chamboredon, J. and Lemaire, M. 1970. “Proximité spatiale et distance sociale. Les grands ensembles et leur peuplement.” Revue Francaise de Sociologie 1, 3–33. Charalampis, D. 1986. Clientelism and Populism: Non-institutional Consensus in the Greek Political System. Athens: Exadas (in Greek). Economou, D. 1988. “Land and Housing System.” In The Problems of Welfare State Development in Greece, eds T. Maloutas and D. Economou. Athens: Exandas, 57–114 (in Greek). Economou, D. 2003. “New Forms of Geographical Inequalities and Spatial Problems in Greece.” Environment and Planning D, Society and Space 11(5), 583–98. Economou, D. 2004. “Urban Planning in Greece: Structural Features and Contemporary Trends.” In City and Space from the 20th to the 21st Century, eds D. Economou, G. Sarigiannis and K. Serraos. Athens: National Technical University and University of Thessaly, 371–82 (in Greek). Economou, D. and Petrakos, G. 1997. “Urban Development Policies and Urban Planning in Greece.” In The Development of the Greek City, ed. D. Economou and G. Petrakos. Volos: University of Thessaly Press, 413–46 (in Greek). EKKE [National Centre for Social Research] – ESYE [National Statistical Service of Greece] 2005. Panorama of Greek Census Data 1991–2001. Database and mapping application under experimental and restricted use at the EKKE.

Social Polarization and De-segregation in Athens

281

Emmanuel, D. 2004. “Socio-economic Inequalities and Housing in Athens: Impacts of the Monetary Revolution of the 1990s.” The Greek Review of Social Research, special issue: Social Change and Segregation Trends in European Cities, A(113), 121–43. European Commission 2004. Interim Territorial Cohesion Report: Preliminary Results of ESPON and EU Commission Studies. Luxembourg: Official Publications of the European Communities. Featherstone, K. 1999. “Europeanization and the Centre-periphery: The Case of Greece in the 1990s.” South European Society & Politics 3(1), 23–39. Ginther, D., Haveman, R. and Wolfe, B. 2000. “Neighborhood Attributes as Determinants of Children’s Outcomes: How Robust are the Relationships?” Journal of Human Resources 35(4), 603–42. Gordon, I. and Monastiriotis, V. 2006. “Urban Size, Spatial Segregation and Educational Outcomes.” Urban Studies 43(1), 213–36. Hadjimichalis, C. 1987. Uneven Development and Regionalism: State, Territory and Class in Southern Europe. London: Croom Helm. Herman, C. 2007. “Neoliberalism in the European Union.” Studies in Political Economy 79, 61–89. IOBE 2006. Greece in International Investment Markets. Athens: Foundation for Economic and Industrial Research. Kandylis, G., Arapoglou, V. and Maloutas, T. 2008. “Immigration and the Competitiveness – Cohesion Equilibrium in Athens.” In Cities between Competitiveness and Cohesion. Discourses, Realities and Implementation, eds P. Ache, H. Andersen, T. Maloutas, et al. Netherlands: Springer Verlag, 119–34. Kandylis, G., Maloutas, T. and Sayas, J. (Forthcoming). “Inequality and Diversity. Socio-ethnic Hierarchy and Spatial Organization in Athens.” European Urban and Regional Studies. Kavoulakos, K. and Kandylis, G. 2012. “New Forms of Urban Conflict in Athens: The Sociopolitical Context, Discourse and Practices of Rejecting the Immigrants.” In Inequality in Times of Crisis, ed. A. Afouxenidis. Athens: Propombos, 79–99 (in Greek). King, R. 2000. “Southern Europe in the Changing Global Map of Migration.” In Eldorado or Fortress? Migration in Southern Europe, eds R. King, G. Lazaridis, and C. Tsardanidis. London: Macmillan, 1–26. Labrianidis, L., Lyberaki, A., Tinios, P., et al. 2004. “Inflow of Migrants and Outflow of Investment: Aspects of Interdependence between Greece and the Balkans.” Journal of Ethnic and Migration Studies 30(6), 1183–208. Leal, J. 2004. “Segregation and social change in Madrid metropolitan region.” The Greek Review of Social Research 113, 81–104. Leiulfsurd, H., Bison, I. and Jensberg, H. 2005. Social Class in Europe: European Social Survey 2002/3, Research Monograph. Italy: NTNU, University of Trento. Leontidou, L. 1986. “In search of lost labour: The urban sociology of post-war Greece.” The Greek Review of Social Research 60, 72–109 (in Greek). Leontidou, L. 1990. The Mediterranean City in Transition. Cambridge: Cambridge University Press. Lipietz, A. 1987. Mirages and Miracles. London: Verso. Maloutas, T. 1990. Housing and Family in Athens. An Analysis of Post-war Housing Practices. Athens: Exandas (in Greek). Maloutas, T. 2000. Social and Economic Atlas of Greece: The Cities. Athens-Volos: NCSR and University of Thessaly Press (in Greek). Maloutas, T. 2004. “Segregation and Residential Mobility: Socially Entrapped Social Mobility and its Impact on Segregation in Athens.” European Urban and Regional Studies 11(2), 171–87.

282

Residential Segregation in Comparative Perspective

Maloutas, T. 2007a. “Segregation, Social Polarization and Immigration in Athens: Theoretical Expectations and Contextual Difference.” International Journal of Urban and Regional Research 31(4), 733–58. Maloutas, T. 2007b. “Middle Class Education Strategies and Residential Segregation in Athens.” Journal of Education Policy 22(1), 49–68. Maloutas, T. 2010. “Mobilité sociale et ségrégation à Athènes: Formes de séparatisme social dans un contexte de mobilité spatiale réduite.” Actes de la Recherche en Sciences Sociales 184, 2–21. Maloutas, T. and Alexandri, G. (Forthcoming). “La gentrification dans les villes non gentrifiables? Rénovation urbaine et changement des structures sociales dans le centre d’Athènes.” In Gentrification, transformations des quartiers populaires et couches moyennes; notion unique, processus pluriels, eds C. Rhein and E. Préteceille. Paris: La Découverte. Maloutas, T. and Economou, D. eds. 1988. The Problems of Welfare State Development in Greece. Athens: Exandas (in Greek). Maloutas, T. and Karadimitriou, D. 2001. “Vertical Social Differentiation in Athens: Alternative or Complement to Community Segregation?” International Journal of Urban and Regional Research 25(4), 699–716. Maloutas, T., Sayas, J. and Souliotis, N. 2009. Consequences of the 2004 Olympic Games on the Socio-spatial Structure of Athens. Paper presented at the RC21 Conference: “Inequality, Inclusion and the Sense of Belonging,” Sao Paulo, August, 23–25. Mollenkopf, J.H. and M. Castells, eds. 1991. Dual City: Restructuring. New York: Russell Sage Foundation. Mouzelis, N. 1986. Politics in the Semi-Periphery: Early Parliamentarism and Late Industrialisation in the Balkans and Latin America. London: Macmillan. Musterd, S., Ostendorf, V. and De Vos, S. 2003. “Neighbourhood Effects and Social Mobility: A Longitudinal Analysis.” Housing Studies 18(6), 877–92. Ostendorf, V., Musterd, S. and De Vos, S. 2001. “Social Mix and the Neighbourhood Effect. Policy Ambitions and Empirical Evidence.” Housing Studies 16(3), 371–80. Pavlou, M. 2009. “Discourse and Policies on Immigrants.” In Immigrants and Minorities, eds M. Pavlou and A. Skoulariki. Athens: Vivliorama, 21–68 (in Greek). Peck, J., Theodore, N. and Brenner, N. 2009. “Neoliberal Urbanism, Models, Moments, Mutations.” SAIS Review XXIX(1), 49–66. Petmezidou-Tsoulouvi, M. 1987. Social Classes and Social Reproduction Mechanisms. Athens: Exandas (in Greek). Rose, D. and Harrison, E. 2007. “The European Socio-economic Classification: A New Social Class Schema for Comparative European Research.” European Societies 9(3), 459–90. Sassen, S. 1991. The Global City: New York, London, Tokyo. Princeton: Princeton University Press. Stathakis, G. and Hadjimichalis, K. 2004. “Athens International City: From the desires of the Few to the Reality of the Many.” Geographies 8, 13–29 (in Greek). Taylor, P.J. and Derruder, B. 2004. “Porous Europe: European Cities in Global Urban Arenas.” Tijdschrift voor Economische en Sociale Geografie 95(5), 527–38. Tsoukalas, K. 1987. State, Society, Labour in Postwar Greece. Athens: Themelio (in Greek). UNECE 2003. Trends in Europe and North America. The Statistical Yearbook of the Economic Commission of Europe 2003, ch. 4 (employment), Available at http://www.unece.org/stats/ trend/register.htm#ch4 [accessed 25.2.2009].

Social Polarization and De-segregation in Athens

283

Vaiou, D. and Hadjimichalis, K. 1997. The Sewing Machine in the Kitchen and the Polish in the Fields: Cities, Regions and Informal Labour. Athens: Exandas (in Greek). Van Zanten, A. 2001. L’école de la périphérie. Scolarité et ségrégation en banlieue. Paris: PUF.

This page has been left blank intentionally

Chapter 13

Conclusion: Residential Segregation and Urban Theory Kuniko Fujita

Residential Segregation and Class Inequality What sociological conclusions can be drawn for urban theory from this context-based study of residential segregation in 11 cities around the world—Athens, Beijing, Budapest, Copenhagen, Hong Kong, Istanbul, Madrid, Paris, Sao Paulo, Taipei and Tokyo? This 11 city study shows the divergent patterns of residential segregation along the lines of class, race and ethnicity. The most significant finding is that residential segregation has a profound meaning in some cities but not in others and that residential segregation and class inequality are highly correlated in some cities but not in others.1 This conclusion analyzes relations between residential segregation and class inequality from a comparative institutional perspective by focusing on forces that either intensify or counteract residential segregation and on institutions that either intensify or counteract class inequality (see Table 13.1). Before comparison and contrast are made among the 11 cities, a well-established case needs to be introduced for relations between residential segregation and class inequality. The most important perspective on the relations comes from empirically oriented American residential segregation studies. Starting with the Chicago School of Sociology’s investigation on European immigrants in the city of Chicago at the turn of the 20th century (Park 1915, Thomas and Znaniecki 1918, McKenzie 1924, Park et al. 1925, Wirth 19282 and 1938),3 American researchers long pursued the relations with strong empirical and census-data supports. They found racism to be a dominant force behind residential segregation (Drake and Cayton 1945, Hughes and Hughes 1952, Taeuber and Taeuber 1969, Forman 1972, Wilson 1987). They firmly established the case that racially discriminatory process has led to high levels of concentration by ethnic and racial minority groups in neighborhoods where residents were subject to higher rates of high school dropout, lower college attendance rates, higher rates of teen pregnancy, higher rates of crime and social disorder, higher unemployment and lower earnings, lower public health and services, higher mortality rates, and higher environmental injustice than counterparts in white neighborhoods (Philpott 1991, Massey and Denton 1993, King 1995, Kawachi and Berkman 2003, Pichett 2003, Bullard 2007, Wilson 2009). They have historically phrased this segregation process and pattern as “separate 1  Class inequality is simply defined as income differences made by income classes which are highly correlated to education levels and the kinds of occupations. 2  Louis Wirth wrote this book when he was still in Germany. Wirth came to the US in 1936. 3  See more about the Chicago School of Sociology in Ballis Lal (1990) and Matthews (1976 and 1989). The Chicago School of Sociology has greatly influenced American sociology and urban sociology where segregation and race relations have become dominant fields. The Chicago School’s emphasis on ethnographic and participative methods was seen in Herbert Gans’(1962) work later and is still in Harvey Molotch’s work including his recently edited book (Molotch and Noren 2010) today.

286

Residential Segregation in Comparative Perspective

and unequal,” meaning that black Americans in segregated neighborhoods are not only spatially separated but also face an unequal access to education and job markets that might enhance their life chance and social mobility.4 The “separate and unequal” pattern has come to mean that residential segregation is one of the major factors leading to class inequality since residential location provides differential access to valued resources that are critical to social mobility and class reproduction. US metro areas are starkly stratified by income and education as a whole (Iceland et al. 2002, Hacker 2003, Dreier et al. 2004, Logan et al. 2004, Street 2005, Hartman and Squire 2009). The fast-growing Hispanic and Asian segments of population have recently changed the dynamic of residential segregation in US metro areas and regions. The growing diversity in 20 most multiethnic metropolitan regions has spawned new neighborhoods with representation of all major racial and ethnic groups—which Logan and Zhang (2010) call global neighborhoods. Half of the white population of these metros lives in such neighborhoods, up from just 20 percent in 1980, while only 3 percent of whites live in the remaining white dominant neighborhoods (Logan and Zhang 2011). Consequently, Logan and Stults (2011) report that residential segregation between whites and blacks measured by dissimilarity index has declined invariably in 50 metro areas and regions in the past three decades between 1980 and 2010. Residential segregation, for instance, declined from 82.8 to 62.8 in New York City and from 85 to 66.9 in Los Angeles over the same period (Logan and Stults 2011). The growth of Hispanic and Asian populations is, they argue, the primary cause of declining residential segregation between whites and blacks and thus the primary factor for segregation decline as a whole. Residential segregation between whites and blacks, nonetheless, they also point out, continues to be persistent in traditional black ghetto belts of Midwest and Northeast. Housing integration policies and school reforms in the aftermath of civil rights movements and institutional changes such as the abolition of Jim Crow Laws have brought a significant decline in residential segregation. Yet, contemporary policies and reforms have proved to be inadequate for a further reduction in residential and school segregation (Gold 2007, Polikoff 2007, Payne 2008, Turner et al. 2008). Black and Hispanic households still live in neighborhoods with more than one and a half times the poverty ratio of neighborhoods where the average non-Hispanic whites live (Logan 2011a). Black and Hispanic children continue to face the low quality education in segregated public schools (Kozol 2006, Martin 2006, US Congress Senate Committee 2010, Logan 2011b) and fail in their human capital development.5 Studying residential and school segregation and income inequality6 among US 50 metro areas between 1980 and 2010, Logan (2011a)

4  The phrase “separate and unequal” is a twisted use of the “separate but equal” Supreme Court decision that was enacted into law in 1895 (Goldstone 2011). The separate but equal law was de fact segregation law until 1954 when the Supreme Court overturned it in the Brown vs. Board of Education case. Ten years later, the Civil Rights Act came into being, abolishing the application of separate but equal in all areas of public accommodations. 5  The lack of quality education means that the children are missing education which refers to the broad set of capabilities—including health, knowledge and intelligence, attitude, social aptitude, and empathy that make a person a productive member of society. Formal education plays perhaps the most important role in forming an individual’s human capital. Differences in attitude to education emerge in early childhood as a result of varying nutrition, learning environments, and behavioral expectations (Rajan 2010: 30). 6  Income inequality refers to the extent to which income’s distribution in an uneven manner among a population.

Conclusion

287

concludes that racial segregation itself is the prime predictor of which metropolitan regions are the ones where minorities live in the least desirable neighborhoods.7 The separate and unequal pattern thus continues today. Logan’s new findings from the 2010 census may, nonetheless, add to the complexity of neighborhood inequality measured by racism, class and nativity (2011a). First, racial segregation and income disparities in America’s 50 metro areas do not directly correlate. Some high income minorities—Blacks, Hispanics and Asians— live in high-income White neighborhoods. But, simultaneously, other high-income minorities often live in their own ethnic neighborhoods. Second, residential segregation between whites and both Hispanic and Asian groups has increased. But income levels do not correlate with increased segregation. The Hispanic and Asian population—including the growing number of foreign born and recent immigrants in both low and high ends of labor markets—tend to live in their own ethnic neighborhoods whose size now expands to a city-wide in most metropolitan areas like Miami, San Francisco, New York, Los Angeles, and Washington, DC.8 Are they self-segregated because of comforts and benefits that their cultural and language affinities bring about? Or, as Alba and Nee (2005) argue, will they eventually follow a mainstream immigrant pattern that immigrants have climbed social ladders as their second and third generations have been assimilated into the American culture? The case of low skilled immigrants may support Alba and Nee as their limited resources require them to live in neighborhoods where they can depend upon socially supporting networks. But there is no need for high skilled immigrants to be self-segregated. The traditional mainstream immigrant concept may need to be re-examined. Self-segregation may also become meaningless in minority majority cities. The American separate and unequal pattern of residential segregation has serious implications for the 11 city study and contemporary urban theories that we now turn to. Comparative Institutional Perspective and Divergent Residential Segregation Patterns How divergent are the residential segregation patterns of 11 cities compared with the American separate and unequal pattern? The American separate and unequal pattern shows that residential segregation is one of major factors that produce class inequality and has been shaped in the American context. The 11 city study in this volume has also made it clear that the divergent residential segregation patterns are sustained by contextual differences in housing and labor markets, politics and policies, the role of the state, the welfare state, family networks, built environment, laws and values, and customs and norms. But, a question to be raised here is whether these divergent patterns are linked to one of the major factors leading to class inequality in the same way as the American pattern. The contextual focus is not new in cross-national residential segregation studies. But, context usually means just national or regional differences, while actual institutional differences are seldom

7  John L. Logan and his team have done great public services by providing free access to valuable data on residential segregation (1980–2010), income inequality (1990–2010) and school segregation (1980–2010) in 50 American metro areas and regions at the webpage “US 2010 Census: Discover America in a New Century.” http://www.s4.brown.edu/us2010/Data/Data.htm (accessed 25 July 2011). 8  New York City, San Francisco and Boston changed from white majority to minority majority (Blacks, Hispanic and Asians) cities. Los Angeles reversed the ratio of white and Hispanic population: 27.5 percent for whites and 48.5 percent for Hispanic (mostly Mexican) groups by 2010. Miami also became an overwhelmingly 70 percent Hispanic (mostly Cuban) city.

288

Residential Segregation in Comparative Perspective

analyzed.9 From the viewpoint of twenty-first-century capitalism where nation-states have taken different trajectories of capitalist development, one must go beyond national and local specificities to explore the institutional differences that constitute various forms of capitalist societies and markets. Sako Musterd and other Western European researchers, for instance, tend to emphasize the segregation of low income people in dilapidated neighborhoods caused by the failure of individuals in the labor market (Hamnett 1996, Musterd and Ostendorf 1998, Ginther et al. 2000, Schnell and Ostendorf 2002, Musterd 2005, Musterd et al. 2006). They, therefore, focus on welfare state policies that help low income households and the unemployed improve their living conditions and access to the labor market.10 They believe welfare state intervention in housing and labor markets helps keep poverty low and prevent dilapidated neighborhoods in Western European cities from developing into stigmatized neighborhoods like traditional American black ghettos. They see American ghettos as the direct result of labor market failure and racism in the absence of welfare state intervention. The main contextual differences between Western Europe and the US come down to the welfare state in Western Europe and racism and the inadequate welfare state in the US. The welfare state approach to residential segregation, nonetheless, downplays segregation processes connected to welfare state policies, themselves. The explosion of urban riots in segregated neighborhoods and suburbs in Amsterdam, London, Paris, Stockholm, and other Western European cities in the last decade has disclosed that the welfare state does not cover all segments of the population. It was, first of all, the welfare state that led to the creation of segregated neighborhoods where immigrants, particularly those with Muslim backgrounds from the Middle East and Africa, were increasingly concentrated along with low income households. Segregated immigrants have missed out on job opportunities and social mobility for generations (Bjornberg 2010, BodyGendrot 2010). They have lacked political representation (Dancygier 2010) and suffered from cultural stereotypes (Muller and Smets 2009). Western European researchers have also focused on the evaluation of social mix policy at neighborhood levels within the EU policy framework (Marlier et al. 2009). But recurrent urban riots in London and other UK cities in 2011 have put the EU social mix policy into question. As many European countries have changed from emigrant to immigrant nations over the past three decades, the strategies of the welfare state, social inclusion and multiculturalism may, Paul Scheffer (2011) claims, be inadequate to deal with immigrants. The welfare state approach to residential segregation needs more discriminatory approach to welfare state institutions that may intensify or counteract residential segregation than the general approach that has so far taken to the welfare state. Comparative Institutional Perspective An analytical framework is necessary for comparison and contrast between the American separate and unequal pattern and the divergent residential segregation patterns of the 11 cities and among the 11 cities themselves. The framework adopted here is a comparative institutional perspective which refers to various forms of capitalism and regionalism. The comparative institutional 9  Marcuse and van Kempen (2000) have, for instance, taken contextual differences into consideration in evaluating the effects of globalization on urban space. Yet, contextual differences in their edited book refer simply to local and national specificities and fail to identify institutions in which specific local labor and housing markets are nested. 10  Welfare state policies include the provision of public housing or housing allowances for low income households to live in private housing, the subsidization of schooling and trainings that the unemployed might enhance their accessibility to the labor market, and relatively high minimum wages.

Conclusion

289

analysis allows us to see how institutions shape or dampen residential segregation, how residential segregation is one of the major factors that produce class inequality, or how institutions intensify and counteract the effects of residential segregation in developing into class inequality, or whether residential segregation has any meaning in the 11 cities. Various forms of capitalism  Various forms of capitalism are derived from Hall and Soskice’s notion of “varieties of capitalism” (2001a) and regulation theory’s notion of “social systems of production” (Hollingsworth and Boyer 1997a) which are extended here to include developing and transition economies,11 covering all 11 cities accordingly. Unlike neoclassical economics theory that makes the direction of market changes unilateral by assuming that markets are rational and self-regulated and independent of national diversity,12 markets in various forms of capitalism are defined as social and historical creations (Polanyi 1943, Esping-Andersen 1990, North 1990, Hollingsworth and Boyer 1997a, Hall and Soskice 2001a) Liberal capitalism like the US and UK is also referred to as a social and historical creation. The American residential pattern of “separate and equal” is, therefore, seen as the product of liberal American capitalism. Liberal capitalism emphasizes legal contracts and market competition and reveals more class confrontational approaches to the economy and society (Hall 1999, Wood 2001). Athens, Hong Kong, Madrid, and Paris embrace the liberal labor market but institutions nested in their national and regional configurations mean that these cities belong to different forms of capitalism. Athens and Madrid belong to Mediterranean clientelist capitalism (Manolopoulos 2011), while Hong Kong exemplifies developmental state capitalism (Applebaum and Henderson 1992). Similarly, Paris is nested in statist capitalism (Boyer 1997 and 2002, Loriaux 1999, Hancké 2001). By contrast, coordinated capitalism emphasizes cooperation between labor, management and the state and takes relational approaches to many parts of the economy and society (Hall and Soskice 2001b). Copenhagen, Taipei and Tokyo belong to coordinated capitalism. Yet, institutional differences make Copenhagen Nordic social democratic capitalism (Esping-Andersen 1990, Jessop 1991, Wood 2001, Pederson 2006) and Taipei and Tokyo developmental state capitalism (Johnson 1982, Wade 1990, Pempel 1999). Institutional difference in Beijing and Budapest in transition economies and Istanbul and Sao Paulo in developing economies also mean that these cities belong to other forms of capitalism as will be seen later. Institutions  The institutional perspective starts from a basic recognition that human activities are embedded and framed within larger institutional schemes. Institutions are rules, laws, social norms and values in culture, society, polity and economy (North 1990: 3). It is these institutions that make various forms of capitalism. Institutions constrain and shape human interaction and define and limit the set of choices of individuals. The structures of a nation are inextricably bound up with its history in two respects. They are created by actions, statutory or otherwise, that established formal 11  The comparative perspective emphasizes national divergences whether it refers to varieties of capitalism or various social forms of production (Albert 1993, Berger and Dore 1996, Hollingsworth and Boyer 1997a, Dore et al. 1999, Hall 1999, Kitschelt et al. 1999, Soskice 1999, Dore 2000, Hall and Soskice 2001a, Morgan et al. 2005, Mahoney and Thelen 2010, Morgan et al. 2010); the welfare state (EspingAndersen 1990 and 1999, Iverson 2001, Pierson 2001, Haggard and Kaufman 2008); finance (Zysman 1996, Iverson et al. 2000, Aoki 2001, Deeg 2010, Englen and Konings 2010); governance (Crouch 2005 and 2010, Hanké 2009). Hall and Soskice’s varieties of capitalism and regulation theory’s various social systems of production cover only developed economies like Germany, Japan, Denmark, the UK, the USA, etc. 12  Neoclassical economic theory refers to claims and theories made by F.A. Hayek, Milton Friedman, Merton Miller, Robert Lucas and others mainly in the Chicago School of Economics, University of Chicago.

290

Residential Segregation in Comparative Perspective

institutions and their operating procedures, on the one hand. On the other, repeated historical experience builds up a set of common expectations. This can be culture that consists of informal rules and common knowledge (Hollingsworth and Boyer 1997b). Nations are thus embedded in their particular democratic (nondemocratic) class struggle in very different institutional arrangements, and this means that similar structural impulses can produce more variety than grand theory would permit. Many of the most important institutional structures depend on the presence of regulating regimes that are the preserve of the nation state (Hall and Soskice 2001b: 4). Institutional variation  Institutional variation refers to different ways of coordinating markets by taking into account the contextual embeddedness of economic rules, players, organizations, or behaviors (Boyer and Hollingsworth 1997a). Skill development in the labor market, for example, varies from Copenhagen and Tokyo to Paris and Hong Kong. Coordinated labor markets in Copenhagen and Tokyo consist of diverse relational and non-market institutions. Skill training is not done in the labor market but in vocational schooling in Copenhagen and intra-corporate job training in Tokyo. Coordinated labor markets are the historical product of cooperative relations among firms, labor unions, schools, employer associations and governments and retain and train skilled workers to meet competitive demand for specific skills (Kume and Thelen 1999, Thelen 2004, Streek 2010). Non-market coordination has important economic implications for those who are academically weak. A highly coordinated training system based on vocational schools and intra-corporate training programs offers the best opportunities for students to acquire skills and promising career prospects and provides a stable economic future to those students who are not academically strong and do not have any academic credentials (Estevez-Abe et al. 2001). By contrast, liberal labor markets in Hong Kong and Paris are based on individual human capital. Wages are differentiated according to education, skills and social attributes. Individuals take responsibility for their own human capital development. Wage differentials are high between skilled and unskilled workers. Greater wage differentials then influence purchasing power of housing in neighborhoods of choice and are eventually translated into class-based residential segregation. Only state and urban policy intervention can reverse the result of liberal market outcomes as we have seen in the case of Hong Kong. Institutional variation also refers to different modes of coordination in the economy and society. For instance, firms’ success depends on the ability to coordinate effectively with a wide range of actors. In the liberal market, firms coordinate their activities primarily via hierarchies and competitive market arrangements. By contrast, in the coordinated market, coordination is extensive and takes place in industrial relations, corporate governance, innovation, inter- and intra-firm relations. These institutional differences in coordination have implications for social and spatial integration and exclusion. Institutional complementarity  Institutions, organizations, and social values tend to combine with each other into a fully-fledged system. An institutional logic in each society leads institutions to coalesce into a complex social configuration (Hollingsworth and Boyer 1997b). This occurs because the institutions are embedded in a culture in which their logics are symbolically grounded and organizationally structured (Hall and Gingerich 2009). The institutional configuration usually exhibits some degrees of adaptability to new challenges, but continues to evolve within an existing style. Hall and Soskice (2001b) argue that states in both liberal and coordinated capitalism tend to make policies and laws that preserve their liberal or coordinated capitalism respectively. For instance, state policy to support unemployment insurance programs is extensive in liberal labor

Conclusion

291

markets like Athens, Madrid, Hong Kong and Paris where market performance directly regulates employment and cyclical layoffs are frequent (Esping-Andersen 1990). In contrast, coordinated labor markets in Copenhagen, Taipei and Tokyo retain the workforce through various social and corporate mechanisms when the economy goes bust. State policy is also inclined to support the use of incentives to retain workers by corporations in Taipei and Tokyo. The extensive job sharing program that the state supports maintains the overall employment level in Copenhagen. Institutional complementarity suggests that state-market relations are complementary in the real economy and labor market. Institutional change  Social change is explained through politics and policies which in turn imply institutional changes (Campbell 2004, Morgan et al. 2005, Streek 2009a, 2009b, Thelen and Streek 2009, Hall 2010). Institutional change is based on an accumulation of changed values and practices and therefore means an extended period of slow change. Institutional change occurs when new rules displace existing ones; when new rules are layered on top or alongside existing ones or when existing rules change their impact due to shifts in the environment (such as shifts in power and distributional resources); and when existing rules change to enable strategic redeployment (Streek and Thelen 2005). Institutional changes take place over time and their effects accumulate over time. Accordingly, institutional change is evolutionary (North 1990). Historical institutionalism adds policy legacies and institutional pasts to the set of variables that were routinely considered when trying to account for the structures and outcomes of politicaleconomic institutions (Hall 1999 and 2010, Hall and Thelen 2009, Mahoney and Thelen 2010, Streek 2010). Institutional change is path dependence that institutions are—and are not—subject to self-reinforcing “lock-in.” Path-dependent lock-in is a rare phenomenon, opening up the possibility that institutions normally evolve in more incremental ways (Mahoney 2000, Mahoney and Thelen 2010: 3). Path dependence also means various slow-moving casual processes like cumulative causes, threshold effects, and causal chains (Pierson 2004) and that the cost of changes is compared to their expected returns in the economy (Arthur 1994).13 Furthermore, institutional change is basically conceived of as an unending process of learning about the inevitably imperfect enactment of social rules in interaction with a context and unpredictable environments (Streek 2010: 665). Structures or systems must be reconceptualized as processes. In sum, institutional change is slow, gradual, evolutionary, cumulative, historical and path dependent. Regionalism  Lastly, regionalism refers to a single geographical region in which cities share some common characteristics. Regionalism also refers to the regional flow, rather than the global flow, of trade and capital (Stallings 1995, Wade 1996, Zysman 1996, Mauro et al. 2011). Beijing, Hong Kong, Taipei and Tokyo are examples of East Asian developmentalism where the state leads economic development (Johnson 1982, Wade 1990, Amsden 1992, Hatch and Yamamura 1996, Woo-Cumings 1999) and influences residential segregation patterns through public policies. The extensive welfare state does not exist in East Asia. But the developmental state plays the surrogate role for the welfare state by making public housing available to all income groups in Hong Kong, by creating jobs through urban and industrial growth policies in Taipei, and by maintaining a full employment strategy through fiscal investment policies in Tokyo. Athens and Madrid embody Mediterranean familism (Rhodes 1997, Gal 2010, Héritier and Rhodes 2011) where family networks make up for the weak welfare state, and enable families and relatives to live within a 13  Path dependence here does not refer to the punctuated equilibrium model of change that is embedded in conceptualizations of path dependence.

292

Residential Segregation in Comparative Perspective

short distance and contribute to socially and racially integrated neighborhoods. Athens, Istanbul, and Madrid also share Mediterranean clientelism and class structures so that the pre-democratic oligarchy dominates the economy (Lynn 2011, Manolopoulos 2011, Streek 2011). The clientelist relationship between interest groups and the state is strong in Southern Europe. The oligarchs have organizations geared toward fostering close contacts with ruling politicians and buying favors (Manolopoulos 2011: 11). Sao Paulo shares Latin American populism with other cities in Latin America (Evans 1995, Schneider 1999, Domingues 2007, Levitsky 2007). Latin American populism which is akin to Mediterranean clientelism keeps Sao Paulo from reforming its century old colonial class structures and residential segregation pattern.14 Copenhagen is an example of Scandinavian social democracy where the Social Democratic Party and the unions have established a strong welfare state regime which has helped reduce class cleavage between capital and labor (Jessop 1991, Esping-Andersen 1999, Wood 2001, Iverson et al. 2000, Crouch 2005, Pederson 2006) and contributes to influence the degree of residential segregation.15 In a nutshell, comparative institutional perspective based on the various forms of capitalism is evolutionary, historical, path-dependent and regional. The comparative institutional perspective is particularly important in understanding the 11 city cases and the American separate and unequal pattern and among 11 cities. The focus of comparison is on forces that intensify or counteract residential segregation and on institutions that intensify or counteract class inequality, or institutions that strengthen or weaken the outcomes of residential segregation (see Table 13.1). Table 13.1

Cities grouped by levels of residential segregation and class inequality and by the forces that intensify or counteract them Forces that either intensify (A) or counteract (B) residential segregation

Institutions that either intensify (C) or counteract (D) class inequality

1. Highly separate and unequal Beijing

Household registration (A)

Istanbul

Religious divide legacy (A) Emerging capitalist economy (A) Colonial legacy and state policy (A)

Sao Paulo

Communist party politics (C) Bureaucratic networks (C) State clientelism (C) Pre-democratic oligarchy (C) Latin American populism (C) Pre-democratic oligarchy (C)

2. Moderately separate and unequal Budapest

Communist legacy (A) Emerging capitalist economy (A)

Bureaucratic networks (C)

14  Populism in Latin America has both an economic and political edge. Populism has mostly addressed “inclusiveness” of the poor against the backdrop of highly unequal societies. The state plays the key role as in institution in mediating between traditional elites and the “people” in general. In appealing to the means of poor people to gaining power, populists may promise widely demanded food, housing, employment, basic social services, and income redistribution. Once in political power, they may not always be financially or politically to fulfill these promises (http://en.wikipedia.org/wiki/Populism#Latin_America). 15  The European Union is not seen here as regionalism as EU member nations widely differ in their forms of capitalism.

Conclusion

Forces that either intensify (A) or counteract (B) residential segregation

293 Institutions that either intensify (C) or counteract (D) class inequality

2. Moderately separate and unequal Paris

Historical class divide (A) State and urban policy (A)

Liberal labor market (C) Urban housing policy (C) Racism (C) Education (C)

Historical class divide (A) State and urban policy (A)

Coordinated capitalism (D) Welfare state (D)

Athens

Mediterranean familism (B) Expansion of public sector (B)

Hong Kong

State and urban policy (B) Developmental state (B) Mediterranean familism (B)

Mediterranean clientelism (C) Pre-democratic oligarchy (C) Exclusive professional associations (C) Education (C) Liberal labor markets (C) Education (C) Mediterranean clientelism (C) Pre-democratic oligarchy (C) Education (C)

3. Separate but equal Copenhagen 4. Together but unequal

Madrid

5. Together and equal Taipei Tokyo

Developmental state (B) Ruling class politics (B) Developmental state (B) State institutions (B)

Coordinated capitalism (D) State and urban policy (D) Coordinated capitalism (D) Compressed wage system (D) State and urban policy (D)

How Divergent are the 11 Cities’ Residential Segregation Patterns? Residential segregation and class inequality are closely linked as seen in the American separate and unequal pattern. How does class inequality and residential segregation play out in the 11 cities then? Table 13.1 shows that the 11 cities cluster according to different relations between residential segregation and class inequality. “Separate” or “together” refers to residential segregation, meaning that some cities have substantial residential segregation but others do not. “Equal” or “unequal” refers to class inequality, meaning that class inequality is mild in some cities but strong in others. Residential segregation and class inequality in turn are highly correlated in some cities but are not in others. “Separate and unequal” means that residential segregation and class inequality are highly correlated and residential segregation is one of the major factors that produce class inequality, while “together and equal” means that residential segregation and class inequality are not highly correlated and residential segregation is not one of the major factors that produce class inequality. There are various combinations between “separate and unequal” and “together and equal.” The first group of the “strongly separate and unequal” category means that both class residential segregation and class inequality are high and intricately intertwined and that residential segregation is one of the major sources of class inequality. Beijing, Istanbul and Sao Paulo are in the first group. The second group is “moderately separate and unequal”: among these cities there is a high

294

Residential Segregation in Comparative Perspective

correlation between residential segregation and class inequality but other institutions like the welfare state mediate the effects of class inequality. These cities are Budapest and Paris. The third group is “separate but equal” category and refers to cities where residential segregation exists but class inequality is mediated by other institutions, like coordinated labor markets, so that reduced class inequality alleviates the degree of residential segregation. Copenhagen belongs to the third group. The fourth category “together and unequal” refers to cities where residential segregation is low but class inequality is high and residential segregation is not one of the major factors leading to class inequality. These cities are Athens, Hong Kong and Madrid. Finally, among cities in the “together and equal” category residential segregation is absent and income inequality is not strong enough to lead to class-based residential segregation. These cities are Taipei and Tokyo. Separate and unequal: Beijing, Istanbul and Sao Paulo  Beijing, Istanbul and Sao Paulo are closer to the American separate and unequal pattern. Household registration in Beijing, religious divide and the emerging capitalist economy in Istanbul, and the Brazilian metropolitan model in Sao Paulo are major forces that drive residential segregation in these cities in the same way as institutionalized racism in American metro areas. Residential segregation is also one of the major factors that produce class inequality in these cities just as in US metro areas. In Beijing, caste-like residential segregation based on household registration determines the life chances and social mobility of local residents with urban hukou and migrants with agricultural rural hukou. Urban residents are state and ex-state officials and their families and relatives and have better access to education, jobs, and housing than rural residents who migrate to Beijing and live at the urban fringe. The Communist Party dominates the state and markets and designs institutions to benefit its members and state bureaucrats (Tsai 2007). In the stock market, the state remains the largest holder of shares and has majority ownership of the major companies (Walter and Howie 2011). Stock market gains mainly go to state agencies. In labor and housing markets, the privileged class of state officials benefits from living in the city core where high paying jobs are concentrated and housing conditions and neighborhood amenities are better than anywhere else. State officials can purchase their flats in state official private housing in the city core. Wealthy professionals can purchase their flats or recently built housing on the open market in and near the city core. But wealthy professionals come from mostly state and ex-state officials and their families and party members. In the urban fringe and countryside, the absence of private property rights has let local leaders like party members control local hinterlands that the state owns. The local leaders play the role of power broker between villagers and the state bureaucracy (Hsing 2010) and benefit from this institutional complementarity. Residential segregation is systematic and historically rooted in the city.16 The urban and rural division in Beijing did not start with the Chinese Communist regime. It is centuries old. As China’s shift to market economy drives rapid mass migration to urban areas from the countryside, the historic division may be at a turning point. Yet, Logan and Li (2012) write that it is unlikely that the Chinese government will abolish the household registration policy any time soon. Uprising and discontent have grown in rural areas. But the demand to abolish household registration is still weak among rural household registration holders. Martin Whyte (2011a) tells us that rural people who have lived too long in poverty have no idea about how privileged urban residents are and are 16  There certainly exists a difficulty to measure segregation and income among neighborhoods in these developing and transitional cities. Partly because their informal economy is too large to measure. Partly because they have developed institutional innovation that fills the absence of formal institutions. And the lack of research on institutional innovation makes it difficult to measure and infer connections between segregation and class inequality.

Conclusion

295

content with limited resources that living in the urban fringe of big cities like Beijing brings to them. Fifty-four percent of China’s 1.3 billion people are still effectively bound to rural household registration (Bell 2012). Apart from exception of a few cities (Bell 2012), recent reforms and institutional changes are unlikely to alter household registration that divides people into rural and urban castes with different rights and opportunities (Whyte 2011b). Beijing is part of East Asia developmentalism. As developmental states in other East Asian countries have done in the past, the Chinese state has managed slow change to the market economy.17 Yet, unlike other developmental states which promote universal education and redistribution of wealth through taxation, the Chinese state and the Beijing city government are not inclined to open up education and labor markets to all citizens but rather create new institutions that promote the privileges of party members and state officials. In Istanbul, residential segregation takes place along the lines of religion, ethnicity and class. Tasan-Kok (2012) writes that Muslim and non-Muslim groups have lived separately in some historical times but cohabited in other times, depending on state policies. The religious divide accompanying the ethnic divide is compounded by the rise of socioeconomic classes that Turkey’s opening-up for international capital investment has brought about. Istanbul presents clear-cut housing segregation today. The high-income groups live in gated communities and along the coast which the Muslim elite traditionally occupied, while the low income groups live at the periphery of the city which non-Muslim ethnic groups used to live. Income differentiation is also generated through state clientelism that demands votes in return for state favors. Istanbul has laws and rules that govern housing and land markets but they are neither applied nor enforced effectively. State clientelism, therefore, fills the vacuum. In the formal economy, state clientelism plays the broker role between state agencies and the housing industry, promoting patron-client relations in urban and housing development (Blgur 2010). In the informal economy, state clientelism makes it possible for various groups of people to make profits from the illegal markets. Governments, for instance, allow unskilled migrants to illegally occupy state-owned land in the city, expecting votes in return and simultaneously supporting the informal economy. Illegal land occupation brings the migrants land rents and profits. The migrants also get apartments in housing blocks built on their land when squatter areas are redeveloped. Consequently, some migrants succeed, Tasan-Kok claims, in climbing to the middle-income class and moving into middle-income neighborhoods. Yet, this is a primarily unique benefit and most migrants are left behind in squats. Legal institutions are often weak or incomplete and property rights remain informal in developing countries like Turkey.18 Government typically refrains from using force to evict the invaders. Eventually the squatters’ de facto rights may be made de jure as seen in Latin America (Weimer 2011: 3). This clientelist arrangement between the state and illegal land occupiers represents an institutional complementarity. With the absence of income and wealth redistribution and the lack of effective tax collections from high income classes, Istanbul’s residential segregation that took place around the religious divide between Muslim and non-Muslim population continues to develop into a class division between rich and poor. Residential segregation in Istanbul is one of major factors that produce class inequality and reproduction. 17  Naughton (2007) argues that China has taken time to adopt the market economy over the past three decades rather than introducing the full panoply of market institutions as quickly as possible in as many realms as possible. He also argues that the development of market institutions may not necessarily be any more important or urgent than the development of other types of institutions. 18  This does not apply to African countries like Kenya where the colonial administration in the early 1900s introduced the documentation of land rights (Onoma 2010).

296

Residential Segregation in Comparative Perspective

In Sao Paulo, residential segregation takes place between the ruling elite living in the city center and the working class and poor living in the periphery of the city as Marques et al. (2012) write. Residential segregation originates from the colonial legacy that the ruling colonial power occupied the city center and pushed indigenous people back to the fringe of the city far from the city center. Today, government development policy called the “Brazilian metropolitan model” retains this traditional residential segregation. Residents in the city center can have access not only to high quality public goods and services but also labor markets, benefiting from state and urban policies and economic growth. By contrast, the Brazilian metropolitan model keeps the workingclass and poor from the city and provides them with fewer public services and benefits. There is no policy that alters Sao Paulo’s residential segregation pattern. The recent populist democracy has led to dramatic reduction in inequality trends since around 2000 (Corina 2012) but still has not institutionalized populist promises for the provision of public social services such as education and housing for residents in the periphery. Health care is, however, the exception. Brazil’s health care reform started in the pre-democratic era. Fallerti (2010) argues that the evolution of health care reforms shows that it is possible to break away, in a gradual and incremental manner, from the historical institutional preconditions that preclude universalization of health. Sao Paulo’s recent changes stay within Latin American populism where patron–client based bureaucracies19 prevent political candidates from realizing their promise to “people” to deliver widely demanded food, housing, jobs, social services and income redistribution (Desposato 2006, Domingues 2007, Roberts and Wilson 2009, Hunter 2010). Residential segregation that separates the privileged high social classes from the working class and poor firmly remains and plays one of the major roles in the production and reproduction of class inequality in the city of Sao Paulo and the Sao Paulo metropolitan region. In sum, Beijing, Istanbul and Sao Paulo demonstrate that residential segregation is one of the major factors that produce class inequality as in American metro areas. Yet, there is a major difference between American metro areas and these three cities. No matter how diminished the welfare state is in US metro areas, the formal welfare state still provides segregated black and Hispanic neighborhoods with basic means—formal education, infrastructure, housing, and safer and cleaner environment—which residents can take advantage of for social mobility no matter how limited that social mobility is. On the contrary, the three cities provide absolutely no means for social mobility for the vast majority of underprivileged residents who live in the urban fringe, periphery and squatters.

19  East Asian developmentalism shares similarities and differences with developmental projects in Latin America. But there is the sharpest contrast between Latin American and East Asian developmentalism regarding where developmental state resides in the bureaucracy: Latin American bureaucrats do not have the “embedded autonomy” characterizing East Asian officialdom (Evans 1995). Compared to East Asia, Latin American bureaucracies are bloated, weakly institutionalized, and unstable. Bureaucratic power is distributed through patronage rather than by merit and is more prone to corruption and dependent upon political machines or the military, and subject to immediate dismissal. Bureaucrats in Latin American states are also less insulated from private interests than in East Asia where a credentialed civil service helped create a professional and nationally committed phalanx of officials (Schneider 1999). Also, Wright (2010) claims that authoritarian state-society relations in China can be seen as in Brazil and other developing countries. Yet, developmental economist Dani Rodrik asserts that the most successful societies of the future will leave room for experimentation and allow for further evolution of institutions over time. A global economy that recognizes the need for and value of institutional diversity would foster rather than stifle such experimentation and evolution (Rodrik 2011: 240).

Conclusion

297

If we talk about the degree of residential segregation and class inequality as the outcome of residential inequality, it is utterly impossible to equate American and European segregated neighborhoods and squatters in urban centers and fringe in these developing cities. Branko Milanovic calculates that the poorest American ventile (one twentieth) are better off than more than two-thirds of the world population (2011: 117). The underprivileged poor and privileged rich are thus not simply spatially separated but severely divided on all counts by government policies in Beijing, Istanbul and Sao Paulo. Moderately separate and unequal: Budapest and Paris  Budapest and Paris fall into the ‘moderately separate and unequal’ category. In Budapest, unlike in Beijing, transition from the planned to market economy appeared to be rapid and sudden. But as Kovács (2012) writes, the gradual transition to the market economy had already started toward the end of the Communist era before the Communist regime suddenly collapsed in early 1990s. Since ex-state bureaucrats dominated the transition process in Budapest as in Beijing, new policies and institutions associated with transition primarily benefited ex-state bureaucrats and their networks more than the rest of population. As a result, privileged ex-state bureaucrats have had better access to higher paying jobs and more opportunities to have business contracts with international capital than any other social groups. Similarly, the city’s residential segregation pattern that divides between privileged ex-state officials and non-privileged did not start with the transition process. Nor did it start with the Hungarian communist system. It originated from a bourgeois city Budapest in the pre-Communist era. The city’s residential segregation pattern is, therefore, path dependent as Kovács argues. The bourgeoisie in the nineteenth century and party members and state bureaucrats during the Communist era lived in the same exclusive neighborhoods where today’s privileged ex-state bureaucrats and higher income people live. The force of residential segregation in the city is the communist legacy and emerging market economy which have turned the privileged ex-state bureaucrats into high income capitalists and professionals. Transition economies in Central and Eastern Europe pose challenges as they develop without fully institutionalizing social protection measures. Without such institutions, transition economies may generate more class inequality (Szelenyi and Wilk 2009). Yet, Budapest, which initially aimed to dismantle the redistributive elements of the socialist welfare legacy, has kept the generous pension system that International Financial Institutions—the World Bank in particular—helped Hungary install (Myant and Drahokoupil 2010, 186). The city has also had multiparty elections since the late 1980s unlike Beijing. The city is, however, slow to adopt democratic institutions that govern labor and housing markets and the economy as Beijing, Istanbul and Sao Paulo are (Engerer 2001, Levitsky 2007, Laszlo 2011). Yet, the EU regime may force Budapest to adopt more democratic institutions faster (Aslund 2010) as it has recently intervened in the Hungarian government’s constitutional change that directed the country to the authoritarian regime (Scheppele 2011). Budapest is still an egalitarian city with low inequality measured by Gini coefficient among European cities. But given the absence of policies and institutions that intend to counteract rapidly progressing class and spatial polarization, the city may be moving in the direction of Paris where the role of the welfare state is inadequate to counteract forces of residential segregation and class inequality rather than Copenhagen where class inequality is effectively counteracted by coordinated capitalism and an extensive social welfare state backed by strong tax system. In Paris, the historically developed class divide between bourgeoisie and working classes is the primary force in the city’s residential segregation. The traditional elite lived for centuries in the center and western part of the central city where the rich and high income professionals are

298

Residential Segregation in Comparative Perspective

concentrated today. By contrast, the working class and the poorest lived at the edge of the city and the middle class lived between the two extreme spectrums. These residential arrangements still continue today. The large flow of immigrants over the last three decades has made inroads into traditionally working class areas in urban rings. As Préteceille (2012) writes, the segregation force of the class divide may be compounded by the racial divide today. Immigrants, particularly African and Arab immigrants, are concentrated in the traditional working class areas in the eastern urban rings. Both immigrants as well as residents in some of working class areas in the urban fringe have little access to better education and job markets. Despite the 2005 urban riots in suburban Paris, there has developed no policy or reform to improve conditions of immigrant suburbs at municipal and national levels. As a country with many immigrants but a history of silencing the foreign born, the French political elite are indifferent to segregated conditions of the immigrants. Citizenship and identity remain central to French politics and policy making. French debates over citizenship focus on exclusion at the borders and assimilate (or integrate) as its moderate variant within them (Barrington 2009: 512). The French state empowered itself by providing the principle of direction by government bureaucrats and protecting the interests of the state. But since 1984, governments of the Left and Right have moved to liberalize aspects of economic activity, freeing finance and banking from regulation and privatizing state companies in whole or in parts, and promoting a stock market culture. The French government has also reduced the direct authority of the central state and provided greater autonomy for local and private initiative, transferring fiscal and policy authorities to local and regional agencies. Yet the French state remains an activist state, having adjusted to the new era of innovation based capitalist development. The core of the old system that the state and the large firms were critical actors still remains today (Hancké 2001: 307). Under the old regime, large firms were policy instruments for the state and that today, instead of a state-led path, the French adjustment trajectory was a firm led one. But the gap that emerged was not filled by the market, but by a mode of coordination which included state elites in the state apparatus, large firms, and haute finance, which assured that large firms were able to construct a novel institutional environment for their own adjustment and then induce other relevant actors—the state, labor unions, the workforce, other companies, and the financial world—to act according to their preferences (Hancké 2001: 333, Palier 2005). Although coordination is much weaker compared with Denmark and Germany, coordination continues between employers associations and the state in social policies such as high skill training programs to the young workers (Culpepper 2001). The French welfare state also remains in the form of hybridization between wage earner solidarity and society wide citizenship rather than conversion to a market-led system of health care, pension, education, and family social security (Boyer 2002). Yet, all these state adjustments and reforms are irrelevant to young immigrants who live in segregated neighborhoods in urban rings and are basically unemployed and excluded from the labor market. The Paris residential segregation pattern driven by the force of class and racial divides continues and is clearly one of the major factors leading to class inequality and class reproduction. Separate but equal: Copenhagen  Copenhagen’s residential segregation pattern can be called “separate but equal.” This pattern is supported by the combination of high taxes, the strong welfare state and coordinated labor markets that Copenhagen shares with other cities in the Nordic European region. It’s not one of the major factors that lead to class inequality and class reproduction. The historically developed class divide is the primary force that drives residential segregation in Copenhagen as in Paris. The upper class has preferred the city’s north-east, while the working class the west and south-west over the past 100 years. These socially differentiated residential

Conclusion

299

locations may go back to even the pre-industrial era; the upper classes to the north around the king’s palaces, to the North West the middle classes and skilled workers and to the west and south west a population composition of newly arrived immigrants, unskilled workers and low income class people. Today, the class divide is highly correlated with income, education, labor markets and housing tenure and type. As in Paris, the large flow of immigrants into Copenhagen has now joined the traditionally working-class areas in isolated urban enclaves and suburbs. Some working class areas have higher rates of unemployment and social benefits recipients as well as lower rates of educational attainment than the rest of the city. This is the result of the welfare state. Anderson (2012) refers to this as the paradox of welfarism; the most generous welfare system simultaneously maintains institutionalized segregation by providing immigrants with non-profit tenure housing located in isolated enclaves. Unlike in Paris, there is high quality, free public education for children even in segregated neighborhoods in Copenhagen. The same high quality public education does not, however, seem to be good enough to make up for the effect of residential segregation. Although thanks to the fair public education, children perform better in school regardless of parent’s own education in Denmark than France (Wilkinson and Pickett 2009), Anderson argues that parents’ own education and social position plays a more important role for children’s aspirations and opportunities than the fair education system. The welfare state can keep the city’s poverty rates and income inequality low but is limited to the provision of social skills. It can neither prevent residential segregation nor provide the benefits of a quality education for the children of segregated neighborhoods. This is the same argument that takes place in American metro areas. It has long been argued that residential segregation is the source of the low level of achievement among children in segregated American neighborhoods (Tinker 1976, Gold 2007 and United States Congress Senate Committee 2010, Herbert 2011). Yet, the combination of the Danish style of coordinated capitalism and the welfare state seems to provide the vulnerable groups—including immigrants and their descendants—with basic and equal footing for social mobility and opportunities. The labor market coordinated by schools and employers keeps high rates of the young people in the labor force. Denmark’s coordinated labor market conditions also offer relatively high minimum wages and special programs that combine jobs and educational training for younger people as seen earlier. Employment then allows vulnerable groups to get access to decent housing and housing ownership. Furthermore, Work Sharing Schemes play an important role in maintaining the full employment system in Copenhagen and Denmark.20 The work sharing schemes are the products of institutional complementarity between capital and labor and the state. The 1993 Danish law recognizes a right to work discontinuously and people’s right to a continuous income. The law allows employees to choose a “sabbatical” year, which can be divided into shorter periods, every four or seven years. Unemployed people take the place of those on leave, who receive 70 percent of the unemployment benefit they would get if they lost their jobs (typically, 90 percent of one’s salary). Danish unions have managed to use such statutory individual rights to reduce the working hours of entire company workforces, and thus increase the number of permanent jobs (Skidelsky 2011). Together but unequal: Athens, Hong Kong and Madrid  Residential patterns in Hong Kong and Madrid may be more socially, ethnically and racially integrated than in Copenhagen, Paris and American metro areas but income inequality among residents within the same neighborhoods 20  Work Sharing Schemes are widely used in Denmark and the Netherlands and also have made inroads into Germany and France.

300

Residential Segregation in Comparative Perspective

is high. That is, rich and poor are segregated within the same neighborhood. The forces of residential segregation are counteracted by the macro level of economic structures, clientelism that obscures class divide, urban policy and Mediterranean familism. The residential segregation pattern in these cities is therefore not one of the major factors producing class inequality. Class inequality may come from liberal labor markets, schooling, family backgrounds, oligarchical business structures, a long tradition of powerful guilds that control a license to practice in many professions, and clientelism. Clientelism—interest groups within society requesting favors from politicians as clients, often with little regard to a reciprocal contribution to the economy (Manolopoulos 2011: 5)—may be the major factor of inequality in Athens and Madrid, while the liberal labor market may be the major factor of class inequality in Hong Kong. Residential segregation has recently declined in these cities because of the increased influx of new immigrant as in American metros. In Athens, ‘togetherness’ comes from historically developed institutions that regulated labor and housing markets, property right laws, the still fairly strong family-based welfare system and the city’s geography and built environment (Maloutas et al. 2012). Recent immigrants following the traditional immigrant path has narrowed residential segregation rates in the city. Athens’s residential segregation pattern does not seem to lead to one of major factors that produce class inequality. The city’s class inequality has more to do with the political elite that maintains the clientelist democracy21 (Manolopoulos 2011: 82–93). Although the clientelist democracy has helped the public sector expand hugely in the past two decades and contributed to non-polarized labor markets and neighborhoods, it has simultaneously helped increase the city’s class inequality as the political party in power appointed its friends to taxpayer-financed jobs (Dragoumis 2005). Differentiated schools and vertically segregated housing arrangements are also the sources of class inequality. Higher income groups in Athens send children to expensive and better private schools outside their residential areas as in Madrid and Paris. Greek natives live on upper floors, while immigrants occupy lower floors within the same building. In Hong Kong, “togetherness” is, as Yip (2012) writes, a policy consequence of the Hong Kong government’s political will not to make neighborhoods where the poor and disadvantaged are concentrated in slums.22 The Hong Kong government owns all the land in the city and also the public housing it has built. It has used housing and urban planning policies to create socially mixed public housing and made it available to all classes. Public housing accommodates nearly half of the population today. Managerial and professional classes who have gained more income in the past two decades have also found public housing more attractive than housing in the private market sector. As a landowner which leases land to the private sector, the Hong Kong government has developed institutions that guide the private sector housing development. This is the role that the developmental state plays in East Asia. The Hong Kong government emphasizes housing infrastructure development such as roads, community facilities and land formation and reclamation, allowing the developers to bring mixed income households to fill the investment costs. The infrastructure emphasis also comes from the British Colonial legacy that the British Colonial government built infrastructure such as subways and public housing (Applebaum and Henderson 1992). Government thus keeps the growing class inequality from developing into a 21  The clientelist democracy also covers the system of closed shops and powerful guilds that controls who can get a license to practice in many professions and rewards connections over merits (Donadio 2011, Manolopoulos 2011). 22  Hong Kong shows that “together” does not mean that residents with various socioeconomic standings interact and share same values in the same neighborhoods.

Conclusion

301

sharp spatial divide by using its urban planning, housing policy and institutional complementarity with the private sector. In Madrid, ‘togetherness’ comes from the recent decline in residential segregation. Domínguez et al. (2012) attribute the residential segregation decline to two decades of economic growth and the greater influx of immigrants. First, economic growth transformed the city socially and spatially over the two decades of the 1990s and 2000s until the financial crisis hit Spain in 2008. Economic growth based on the real estate boom brought skilled and unskilled immigrants. Highincome immigrants—company directors and managers, professionals and technicians—moved to the city’s upper and middle income neighborhoods, the traditionally reserved upper and middle class areas expanded into neighboring working class areas, on the one hand. On the other, low skilled immigrants joined neighborhoods traditionally reserved for immigrants. In both cases, social mixing increased and led to segregation declines in the city as a whole. Second, stagnation in intergenerational social mobility rates has led to reduced residential segregation. As working class children became professionals with higher education, they tended to live near their parents’ homes. The entry of a great number of educated young people into the labor market and high unemployment rates among the highly educated young led, however, to downward wage pressure for young professionals. The residential choices of the young professionals were, therefore, limited. The result was to encourage the typical Mediterranean residential strategy to settle near the family or to live in parents’ homes, leading to increased social mixing and low residential segregation. Class cleavage is stronger in Athens and Madrid (and in other cities in Southern Europe where money has remained in the hands of an oligarchy of aristocratic and pseudo-aristocratic families) than Copenhagen and other cities in Northern Europe. As Esping-Andersen (1999) writes, the Nordic welfare state’s employment maximization policy was instrumental not only in averting a formidable class division in the occupational hierarchy and core/periphery cleavage that dominates France, Italy and Spain, but also in increasing mobility chances from long-term unemployment and precarious employment to regular positions (Esping-Andersen 1999: 297). In the absence of the Nordic-style welfare state, the Mediterranean super-rich pay ever fewer taxes at home than their counterparts in Northern Europe. The rich in Greece and Spain do not want to invest in their countries.23 Financial deregulation has created ample opportunities for those who invested in US Treasury Bonds and in Fifth Avenue real estate. Some point out that this is related to European integration (Manolopoulos 2011). EU structural funds, later low interest on government borrowing under the EU, and now the transfers that are being paid in various forms to refinance the Greek national debt compensate for Mediterranean countries’ inability to tax their aristocracies. At the same time, they made it unnecessary for governments to invest in more effective tax collection. Northern economies’ aid has filled and still fills the gaps in public coffers caused by national money migration abroad or tax avoidance. It thereby in effect subsidizes a latest social compact under which post-Fascist Mediterranean democracies left pre-democratic quasi-feudal elites alone to allow for national conciliation, instead of expropriating them one way or another in favor of a new entrepreneurial middle class. The place of governments like Greece and Spain within panEuropean institutions shelters them from the contradictions and dynamics visible at the national level (Manolopoulos 2011, Streek 2011). In sum, relatively integrated neighborhoods in Athens, Hong Kong and Madrid cannot be described as segregated and as being one of major factors in producing class inequality. Athens and Madrid now face the consequences of the 2008 global financial crisis. The two decades of the boom were based on state sector expansion as a means to employment for political favorites 23  The exception is Spain’s Basque Region and Catalonia where regionally-invested economies thrive.

302

Residential Segregation in Comparative Perspective

(Manolopoulos 2011: 247) and on the basis of over-borrowing in Athens and the unsustainable housing industry boom which attracted international creditors as well as international immigrants at both ends of the labor market in Madrid. The crisis ended the boom in two cities. Athens and Madrid now face a sovereign debt crisis for different reasons as part of the euro zone.24 But reduced residential segregation pattern will not change the together and unequal pattern. The small decline in segregation rates means that small change has simply been added to the established core pattern, which remains intact. Together and equal: Taipei and Tokyo  The final pattern is one in which residential segregation is absent. To begin with, both Taipei and Tokyo have small racial and ethnic minorities. Taipei has a tiny portion of aborigines, Southeast Asians and immigrants from mainland China. Japan has a restricted international immigration policy and is open to Japanese descendants from Brazil and a few other countries. Tokyo has only 3 percent registered foreigners. Taipei and Tokyo’s togetherness is due to spatial integration across all socioeconomic classes. The two cities can also be described as socially equal in the sense that income inequality among social classes is not great enough to reinforce class cleavages in residential arrangements. Taipei is nested in East Asian regionalism and coordinated capitalism. Wang and Li (2012) write that the city’s social and spatial integration is the result of the Taiwanese developmental state’s economic policy and the Taipei city government’s urban planning. The relocation of the city government office and the city’s CBD from old Taipei city in West to the new Taipei city in East is perhaps the most salient example of the developmental state’s strategy to make the city internationally more competitive in the era of innovation oriented capitalist development. What followed the creation of new Taipei city and other associated urban and infrastructure development is the emergence of a multi-pole metropolis and the population movement seeking both new job opportunities and low cost housing. Simultaneously, the city’s reinvestment in the old city center pushed up prices in land, rent and housing, forcing residents to move to new urban areas. The city’s developmental policy thus triggered the decentralization of population from the city center and dispersed the wealthy that were concentrated in the old city center to other parts of the city and the new Taipei city. The move of the national capital owes much to the particular historical factor that the current ruling political party, Kuomintang, which was transplanted from Mainland China in the 1940s, desired a new capital of its own. Since small businesses that operate relational networks dominate Taipei’s economy, the city’s labor market is less polarized and has a wide strata of middle income jobs. Highly educated people are spread over most parts of the city, and are distributed among all household income groups. The less polarized labor market, the relatively even distribution of the highly educated and the relatively fair distribution of household income among places indicate the low degree of residential segregation. Residential income inequality is quite low. If class inequality exists, it is primarily between the governing political party and its cronies and the rest of the population. The clientelism and corruption seen in Athens, Istanbul and Madrid are rampant in Taipei as well. The ruling political party wishes to maintain its dominance at city and national levels and uses the city’s economic growth policy to maintain its power, benefiting its allied industries and urban developers through economic developmental projects. Yet, this kind of clientelist relation is limited as the government’s 24  When the 2008 crisis happened, Greece had a high level of debt but Spain didn’t. But both countries have faced the debt crisis as the European Central Bank that was supposed to play the role of the central bank in these countries didn’t play that role (De Grauwe 2011).

Conclusion

303

legitimacy lies in its ability to coordinate all stakeholders and players and to promote developmental policy in education, the economy and infrastructure. Governments are judged by successful policy outcomes in generating jobs, income growth, social mobility for the entire population, and social and national cohesion. Policy results weigh over the party and interest group relations. Tokyo is the prime example of residential income inequality not automatically translating into class-based segregation. In the absence of residential segregation in Tokyo, Fujita and Hill (2012a) show that Tokyo’s spatial income inequality over the period between 1970 and 2005 has been cyclical, with ups and downs according to real estate booms and busts which were highly related to the 1990 financial crisis and the 2008 global financial crisis. The four central core wards in the CBD had the highest income and land prices in Tokyo and led the cyclical pattern for the whole Tokyo (23 wards and 27 suburban cities). They represent the concentration of institutional power centers that underpins Tokyo’s functional primacy in Japan’s urban system. The gap between the fortunes of the four core wards and the rest of Tokyo’s districts essentially determines spatial income inequality in metro Tokyo. As Tokyo’s spatial income inequality is exclusively led by booms and busts in the central 4 wards, there is almost no variation in per capita income among communities outside the central four wards. Accordingly, land value and per capita income correlate highly across areas in Tokyo, but occupation and income do not. Tokyo’s neighborhoods are socially integrated in all respects—income, education, occupation. Tokyo’s spatial income inequality is not a function of social class segregation. Tokyo’s political economy helps to explain the absence of residential segregation. Tokyo is nested in Japan’s coordinated capitalism where companies function like communities and relational and cooperative approaches prevail in labor industrial relations, innovation system and industrial network system. Japan has reformed its political economy like other developed countries over the last three decades but retains the continuity of developmental state capitalism as discussed in the chapter on Tokyo. Institutions like the compressed wage system and the bonus-based profit sharing system to all employees continue and militate against occupational segregation and social polarization. The developmental state also plays the role of the counteracting force against the development of class-based residential locations through the centralized form of public education, the centralized tax system that equalizes gaps between high and low tax localities, and fiscal policybased public investment in both central and suburban cities. Implications for Contemporary Urban Theories The 11 city cases show how deeply their residential segregation patterns are woven into the social fabric of the cities. The forces and counteracting forces that shape or dampen residential segregation are embedded in particular institutions—household registration policy, communist legacy, the historical legacy of religious divide, the emerging capitalist economy, the combination of the colonial legacy and the state policy of metropolitan development, the historical legacy of class divide, the combination of the welfare state and coordinated capitalism, Mediterranean familism, urban planning and state housing policies, and the developmental state. These forces are the historical products of political economies nested in various forms of capitalism, embodying accumulated changed values and practices. Changes in housing segregation patterns, therefore, reflect macro-level institutional change—such as labor and housing market changes, changing immigration flows, state and urban policy change—and micro-level change—such as personal values and norms and strategies and choices. But institutional changes, which do not take place in a short period of time, are slow, evolutionary and path dependent as the 11 city study has indicated.

304

Residential Segregation in Comparative Perspective

Many of the 11 city cases often refer to profound social changes they have experienced over the past three decades whether social change refers to industrial structural change from manufacturing to services, the large influx of new immigrant groups, financial sector expansion, technological change, transitions from the planned socialist economy to the capitalist market one, or globalization (opening up the domestic market for international trade and the increased inflow of direct foreign investment). The 11 cities, nonetheless, show that their residential segregation patterns have been stable over the past three decades and have not changed very significantly. This empirical reality throws currently popular urban convergence and metropolitan polarization theories into question. Ungrounded urban convergence  Urban convergence theories—particularly seen in global city (Friedmann 1986, Sassen 1991 and 2002), global city networks (Knox and Taylor 1986, Taylor 1995 and 2004, Taylor et al. 2006 and 2010) and neoliberal urbanization theories (Brenner and Theodore 2002, Peck and Tickell 2002, Tickell and Peck 2003, Peck 2010, Brenner, Peck and Theodore 2010)—claim that globalization produces urban convergence to a common model. Globalization is claimed to produce a convergence in the economic base, spatial organization and social structure of cities, corresponding to the functional role a city plays in the globalization process and the level a city occupies in the international division of labor (Friedmann 1986: 318). Similarly, globalization is contended to produce a world city system that transcends national institutions, politics and culture (Sassen 1991: 86). These theories basically follow neoclassical economic theory which contends that globalization—from trade and global capital mobility—leads poor countries to prosperity through foreign investment and technology transfers and thus reduces national divergences. Neoclassical economic theory makes the direction of market changes unilateral by assuming that markets are rational and self-regulated and unaffected by national diversity, so convergence theories treat changes brought by globalization as universal by assuming that the power of global capital annihilates national divergence, makes state policy making unilaterally turn in a neoliberal direction, and establishes supranational global city network. Neoliberal urbanization theories modify neoclassical economic theory by recognizing that competitive markets are neither inherent nor self-regulating but socially constructed and attended to by the state and by extending the market model beyond the economy to government and society. Yet the recognition of the market as a social creation and the application of neoliberalism to the state are directed toward the single purpose of the neoliberal political project that they claim spreads among nations, weighs into the balance of local political forces, interacts with competing ideologies, and combines with different kinds of governance practices (Tickell and Peck 2003). They also interpret any urban and global change as advancing their claims about the neoliberal turn without any empirical research and substantial data.25 Moreover, they claim that there is a power shift within urban regimes from government regulatory agencies to profit-driven financial institutions (Brenner and Theodore 2002: 367–377; Brenner, Peck and Theodore 2010). The neoliberal turn argument may fit Boyer and Hollingsworth’s contention that globalization expands from national embeddedness to spatial and institutional nestedness in the world economy (1999b). But unlike Boyer and Hollingsworth, neoliberal turn contenders hardly take the logic of institutional change and complementarity into consideration.

25  Urban and global change includes economic transition from mass production to flexible specialization, national cutbacks in urban expenditures, deregulation of finance, central government retrenchment, the devolution of revenue-raising responsibility to localities, and pressures to innovate and force cities to compete for resources in the private capital markets.

Conclusion

305

Convergence theories invariably look at cities from the perspective of global and financial capital forces and needs, making cities stateless (Therborn 2011) and dissociating them from national and regional configurations in which they are nested (Hill and Fujita 2003). Convergence theories treat changes brought about by globalization as unilateral by assuming that the power of global capital annihilates national divergence, makes state and urban policy-making unilaterally turn to neoliberal direction,26 and establishes supra-national global city networks detached from national and urban institutions. Totally neglecting how institutional change takes place through politics and policies, they make imaginative linkages between cities and global capital. As the 11 city case study shows, the effects of globalization are strongly mediated by domestic institutional arrangements. The stability of their residential segregation patterns means that national and urban policies and politics in response to globalization have not led to the altering of institutions that historically shaped residential segregation patterns. Institutional insights may clarify how convergence occurs. Convergence occurs when rules remain formally the same but are interpreted and enacted in new ways. The gap between the rules and their instantiation is not driven by neglect in the face of changed setting; instead, the gap is produced by actors who actively exploit the inherent ambiguities of the institutions. Through deployment, they convert the institution to new goals, functions or purposes. The old institutionalism in sociology is replete with examples of institutional innovators working with existing materials to craft solutions to new problems. New elites come to power and orchestrate the shift from within. However, in some cases even those who are disadvantaged by an institution can get traction out of convergence strategies. Yet differences between existing institutional rules and the prevailing political context affect the likelihood of specific style of change (Mahoney and Thelen 2010: 17– 18). Convergence may eventually lead to institutional change like the Civil Rights Act which came into being after years of civil rights movements by strong civil rights activists. In the institutional view, the effect of globalization on cities is not convergence but layering of new elements onto the old rules. Some comparative political scientists do not attribute any changes in policy making structures to globalization. In fact, some totally reject globalization as a pressure on institutional change (Iverson 2001) and argue that pressures on institutional change have been generated primarily within domestic politics (Pierson 2001). The direct influence of global capital is, therefore, totally out of question. Control over the economy always falls upon state politics but not on international capital (Clark 2005). Whereas there are some institutional changes, they are not strong enough to alter institutions that support policymaking structures. In privatization, deregulation and decentralization which neoliberal urbanists take as the neoliberal political project, for instance, they ignore politics that has taken place among policy makers and actors with various interests and intentions. State and urban policy makers’ intention to make deregulation, for example, may not be neoliberal in intention but presume other aims (Weiss 2010, Pickvance 2012). In East Asian and Scandinavian regionalism, privatization, deregulation and decentralization have mainly been used as policy adjustments to 26  The neoliberal state provides the requisite governance institutions and regulatory policies–legal system, sound money, public order and national defense–to secure private property and economic competition; but the state refrains from direct intervention in markets to achieve substantive economic goals because public involvement is believed to distort prices and benefit powerful interest groups (Harvey 2005: 2). But Brenner and Theodore (2002), Peck and Tickell (2002), Tickell and Peck (2003) and Peck (2010) claim that neoliberals seek to undo the class compromise between business and labor, embodied in collective bargaining and the welfare state, which underpinned Keynesian liberalism in the West from the 1930s to the 1970s. Finance capital, particularly institutional investors, is the leading force in the neoliberal socio-economic alliance and political regime.

306

Residential Segregation in Comparative Perspective

move in the direction of various stages of the developmental state and social democracy but not to neoliberalism. Even in Greece where entry to the European Monetary Union was said to be a turn to neoliberalism, Athens proves to have been deeply embedded in Mediterranean clientelist capitalism and class structure. Convergence theories lack comparative institutional perspective. Institutions developed around capital and labor in various forms of capitalism, for instance, neither fall into the question of state and market or the question of the strong and weak state or market. State and urban policies in various forms of capitalism tend to be inclined to make policies and laws that preserve their forms of capitalism. Institutional complementarity is evidenced in state- market relations in Athens, Beijing, Budapest, Copenhagen, Hong Kong, Istanbul, Paris, Sao Paulo and Tokyo. Convergence theories are so one-dimensionally focused upon global capital power that they do not even admit that global capital depends upon institutions—including markets which are socially constructed in various forms of capitalism which then make up the world economy. Global capital that is not rooted in urban and national institutions cannot influence politics and policy change in cities. The effects of globalization as social change require institutional changes which in turn depend upon politics and policies that are nested in local institutions. The 11 city study shows that divergent residential segregation patterns based on various forms of capitalism remain the same over the past three decades. This demonstrates that national structures and institutions in which residential segregation patterns are embedded have not changed significantly in the past three decades. The structures and institutions have constantly evolved but have not accumulated changed values and practices radical enough to convert the fundamentals of national structures. There are of course some changes in residential segregation: The increased social mix by class and racial and ethnic groups in the city center in Copenhagen and Paris, expansion of the wealthy residential areas in Madrid and Paris, decline of residential segregation with the increased flows of immigrants in Athens and Madrid. However, these changes are not altering the already established core patterns of residential segregation in these cities as seen earlier. Politics and policies around residential segregation in these cities suggest that there exist no actors (policy makers, social movements or other players) who could fight the political context and institutional arrangements in order to change the patterns. These changes are, therefore, seen as adding or layering new elements to the established patterns. Each new element may be a small change in itself, yet these small changes can accumulate, leading to a big change over the long run. Nonetheless, three decades are perhaps too short to observe institutional change. Douglass North writes that after 1000 years of civilization, despite the implications of neoclassical international trade models that would suggest convergence, there is enormous contrast between economies (North 1990: 92). Hall and Soskice (2001b: 60) also observe that world trade has been increasing for 50 years without enforcing convergence and because of comparative institutional advantage, nations often prosper, not only by becoming more similar, but by building on their institutional differences. Moreover, Rodrik (2011) contends that developing countries like China, Brazil and Turkey keep their distinct institutions and make institutional innovations to embrace and respond to global pressures, and that those markets and governments are complementary, not substitutes, and that markets work best not where states are weakest, but where they are strong. Unexamined metropolitan polarization  Polarization theories—seen in dual city (Mollenkopf and Castells 1992) and divided city (Fainstein et al. 1995) theses—claim that cities experience social polarization as a consequence of industrial structural change. Polarization is also claimed as a consequence of globalization (Sassen 1991, van Kempen 1994). In fact, the cities have gone through industrial, global and multicultural changes but have not developed polarization in labor

Conclusion

307

markets and residential segregation over the past three decades. Residential segregation among the 11 cities has been stable as seen earlier, reflecting both historical institutionalism that does not change the historically established institutions like the forces driving residential segregation. The forces are the products of multilayered institutional changes and practices and conflicts and struggles at both macro and micro levels. The effects of industrial structural change at the macro level do not reflect upon residential segregation patterns directly. Many of the 11 cities have experienced deindustrialization but have not developed polarized industrial structures. Various institutional configurations of industrial structural change have kept the 11 cities from developing a more polarized labor force. Copenhagen, Hong Kong, Paris, and Tokyo have shrunken their manufacturing and been transformed to the service economy. But, simultaneously, new forms of flexible specialization based on regional cluster economies rather than globalization have also emerged and counteracted the decline of manufacturing (Piore and Sabel 1984, Sabel 1989, Krugman 1991, Sabel and Zeitlin 1997, Porter 1998, Zeitlin 2007, Bellandi and Caloffi 2008, Herrigel and Zeitlin 2010, Fujita and Hill 2012b). Consequently, the cities belonging to the various forms of capitalism have made policy adjustments to industrial structural change over an extended period of time. Copenhagen, Hong Kong, and Paris have made up for manufacturing job losses with service jobs of various skill levels, leading to non-polarized labor markets. Also, not all cities have undergone deindustrialization. Tokyo has constantly upgraded its manufacturing industrial structure to involve more research and development, while the growth of the service economy is highly linked to the manufacturing sector. Moreover, in none of the 11 cities has the service industry replaced manufacturing in economic importance. Manufacturing industry is still considered to be important as research and development and intellectual property rights are strongly linked to manufacturing. This is true even in the US where the post-industrial society was believed to be a highly likely promise since Daniel Bell wrote “The Coming of PostIndustrial Society” in 1972. The US has, however, remained a manufacturing power specialized in higher valued items in machinery and chemistry (The Institute of Supply Management 2012) as other industrialized nations have. As manufacturing has helped to keep a weak economic recovery from turning into a new recession, the emphasis on manufacturing is arising in post 2008 financial crisis America (Meyerson 2011, Obama 2011, Norris 2012). Finally, the manufacturing sector decline has also not led to the loss of labor unions. Unionization rates remain high in Denmark and Scandinavian countries. Some of the 11 cities have experienced increased international immigration flows but have not seen an impact great enough to alter the established immigrant labor patterns that serve all strata of labor markets. None of the 11 cities has developed a polarized labor force with high income professionals linked to the global economy and low income unskilled workers and immigrants mainly serving globally linked managers and professionals. The structural changes surely influence the residential segregation patterns. But changes in residential segregation at the neighborhood level need to be filtered through micro-level institutions as well. Institutional changes are slow and evolutionary as they are the result of the accumulation of changed values and practices. The 11 city study demonstrates that the basic residential segregation patterns historically shaped over the past several decades or even past few centuries are remarkably stable. The effect of structural changes on residential segregation simply adds to established spatial patterns as seen in the case of Athens, Beijing, Budapest, Copenhagen, Hong Kong, Istanbul, Paris, Madrid and Sao Paulo. Residential segregation is deeply embedded in local institutions that have evolved and layered upon through social, cultural and political struggles. For instance, the residential patterns of natives and immigrants does not change once both groups establish it. Studying five groups of whites,

308

Residential Segregation in Comparative Perspective

blacks, Puerto Ricans, Koreans and Eastern Europeans in three communities in Philadelphia, Goode and Schneider (1994) offer insights that initial establishment involves so many factors within the already historically and racially divided American society. They argue that norms, ideologies, prejudice, racial and ethnic conflicts, family networks, institutions, and historical legacy play out and that the life choices, strategies, and perceptions of individuals are profoundly affected by both historical and present-day economic structures and power dynamics (Goode and Schneider 1994: 242–3). Social and economic structural changes also need to be seen as complementary to changes at neighborhood levels. Many of the 11 cities have gone through the decline of manufacturing industry and rise of a service economy but have not altered their overall segregation patterns. The pattern of residential segregation built upon the already established race relations and built environment does not change drastically. As Athens, Beijing, Budapest, Copenhagen, Hong Kong, Istanbul, Madrid, Paris and Sao Paulo indicate, the established residential segregation patterns have remained unchanged, simply adding and expanding the edge of the established areas. As immigrants with Muslim backgrounds from the Middle East and Africa have moved into the most disadvantaged working class neighborhoods in suburban Paris, Paris’s residential segregation pattern has been compounded by class and race. But the recent increase in residential segregation rates in Paris does not change the established segregation pattern as immigrants moved to the already segregated traditional working class suburbs. Social change in the 11 cities has been filtered by institutional configurations at national and local levels and coalesced into a national culture. The segregation study shows that 11 cities have evolved but remain with their own forms of capitalism. Metropolitan polarization along the lines of space, class, race and ethnicity has not also been seen to be worsened in US metro areas over the past three decades, either. To be sure, income polarization has grown dramatically in the US since 1980 (Weinberg 1996, Jones and Weinberg 2000, Piketty and Saez 2003, Krugman 2007, EPI 2009, CBO 2011, IRS 2012). The Congressional Budget Office (2011) reports that from 1979 and 2007, income after transfers and federal taxes for households at the higher end of the income scale rose much more rapidly than income for households in the middle and at the lower end of the income scale. In particular, for the 1 percent of the population with the highest income that the Occupy Wall Street movement in the urban America popularized the long-neglected inequality growth (Byrne 2012), average real-after tax household income grew by 275 percent (CBO 2011). For others in the 20 percent of population with the highest income (those in the 81st through 99th percentiles), average real after-tax household income grew by 65 percent over that period, much faster than it did for the remaining 80 percent of the population.27 IRS income tax return reports also show that income shares of the top 1 percent grew much larger than those of the top 5 percent and the top 10 percent which hardly changed between 1986 and 2008 (IRS 2012). CBO and IRS clearly show that the income inequality growth is about the top 1 percent and everyone else. Piketty and Saez (2003) attribute the cause of the income polarization growth to American politics and policies (especially tax cuts on the rich) but not to skill-based technological change, globalization and educational levels.28 Similarly, Krugman (2006, 2007 and 2011a) argues that highly educated workers have done better than those with 27  For the 60 percent of the population in the middle of the income scale (the 21st through 89th percentiles), the growth in average real after-tax household income was just under 40 percent. For the 20 percent of the population with the lowest income, average real after-tax household income was 18 percent higher in 2007 than it had been in 1979 (CBO 2011). 28  Comparative studies of top incomes show that countries like France, Germany and Japan have gone through the same technological and globalization forces but haven’t experienced such a dramatic increase in income gap as the US in the past three decades (Atkinson et al. 2010, 2011). And Atkinson et al. suggest that

Conclusion

309

less education, but America’s income polarization growth is not about the educated versus the uneducated. Krugman attributes the cause of the income polarization growth to the rise of narrow oligarchy that market forces and politics and policies have helped concentrate income and wealth in small elite over the past three decades. Many others also support the argument that the growth of income polarization is caused by American politics (lobbying by special interest groups), policies (deregulation, tax cuts on the rich), and institutional changes (shrunken welfare state and weakened labor unions) (McCarthy et al. 2008, EPI 2009, Hacker and Pierson 2010, Noah 2010, Gilligan 2011, CBO 2011). Then, how has the growth of income polarization at the national level been translated into metropolitan and neighborhood levels? Income sorting has no doubt important implications for characteristics of neighborhoods and schools (Watson 2009). Some argue that neighbors’ incomes and other characteristics are the market-driven outcome of individual choices and that prices set in the housing market determine what household units and neighborhoods households can afford (Hardman and Ioannides 2004). Yet, following two studies show otherwise. Reardon and Bischoff’s study (2011), which examined residential segregation by family income at the metropolitan level between 1970 and 2010, shows that the dramatic growth of income polarization at the national level has certainly reflected on growth in the metropolitan residential segregation. And it shows a continuously shrinking middle income class and increasing residential segregation by class over time. But it hardly reveals a dramatic change in residential segregation. Also, John Logan, who rejects market driven individual choices for housing and neighborhoods, finds that there are little relationships between income differences and neighborhood inequality across racial groups in metropolitan America between 1990 and 2010 as seen earlier. Logan’s findings (2011a) reveal that blacks’ neighborhoods are separate and unequal not because blacks cannot afford homes in better neighborhoods, but because even when they achieve higher incomes they are unable to translate them into residential quality. Logan concludes that racial bias polarizes neighborhoods before income inequality does at neighborhood levels. The two studies by Logan and Reardon and Bischoff thus hardly indicate that America’s dramatic growth of national income inequality has been translated into residential segregation by race and ethnicity at the metropolitan level. Neighborhood inequality based on the separate and unequal divide in American metro areas was already deep enough 30 years ago. As seen earlier, residential segregation by race and ethnicity, which had been declining since 1990, further declined in US metro areas and regions by 2010 (US 2010 census: discover America in a new century 2011). These empirical evidences are the most powerful facts, indicating that income polarization at the national level has not been translated into spatial polarization in metropolitan America along the lines of race and ethnicity over the past three decades. Consequently, the empirical studies on income inequality growth and residential segregation in the US do not support metropolitan polarization theories in two accounts. First, there is no empirical evidence that the dramatic growth of income polarization is caused by skill based technological change or globalization forces, or education to which technological change and globalization ultimately come down to for the measurement of wage and income differences. Second, there is no empirical study that can demonstrate how the income polarization growth caused by industrial, technological and global forces is translated into city and neighborhood levels. Metropolitan polarization theories are not examined after all. They are constructed on assumption that income polarization at the national level may take place in the same way as in cities and neighborhoods. institutional arrangements such as high income taxes and social norms in these countries have kept the top one percent from growing

310

Residential Segregation in Comparative Perspective

In particular, global city theories are totally fallacious by claiming that globalization polarize metropolitan labor markets into the two distinct classes of highly paid managers and professionals connecting to the global economy and low waged and unskilled workers and immigrants mainly serving global capitalist classes and that labor market polarization leads to metropolitan spatial polarization along the lines of class, race and ethnicity. The importance of the institutional context is apparent for residential segregation study and for urban theory. Any social change like the effects of globalization and neoliberal influences is filtered through institutional configurations at the local, national, and regional levels. Policies and politics and institutional change play a vital role in social and spatial change. And institutional change occurs on the basis of an accumulation of changed values and practices which political and economic actors like individuals and organizations generate by interacting and reacting to the external pressures within an institutionally constrained framework. Ultimately, the effects of globalization and industrial and technological change on the 11 cities are hardly described as convergence or polarization. The 11 cities’ divergent residential segregation patterns have been stable over the past three decades and reveal neither urban convergence nor metropolitan polarization. Research Limits and Future Research The 11 city study has aimed to widen the scope of research on context-based residential segregation to many more parts of the world. What is meant by context is the historical comparative framework which requires institutional analyses based on varieties of capitalism and regionalism. It is also hoped that the future study of residential segregation can overcome limits and shortcomings the 11 city study has faced. It is, first of all, desirable to explore more institutional analyses of forces and factors that led to the shaping of residential segregation. There is no extensive study about how markets are socially created in the developing urban world. Nor are there studies about how institutions are made, work and change in the areas of labor markets, the welfare state, and property rights law in different cities nested in various capitalist forms, regionalism and transitional economies. Second, more longitudinal analyses are needed for the study of residential segregation. It takes at least a few generations for residential segregation to take place effectively although any change in residential segregation patterns constantly takes place around the already established patterns as the 11 city study confirms. Unfortunately, data to support longitudinal analyses are unavailable for most city cases. Some studies are anecdotal and not substantiated with longitudinal statistical data. The future study can trace back to history in a more systematic way. Lastly, the lack of comparable data for the 11 cities has led to the use of any data available which has led to problems of compatibility. Some cities used the census data by individuals and families. Although all data have their own limits and shortcomings, income statistics by family present a problem unless family income is divided by family members. Family size varies and may distort income distribution. Focusing on consumption statistics like disposable income poses another problem. The rich may not spend as much as the poor as the rich may save much of their income. Fiscal data are not always accurate as the poorest that do not pay taxes are not included in the data. Fiscal data can only be used where income distribution is fairly egalitarian. The use of higher education may not be accurate to measure income distribution among places as higher education holders do not always have high incomes. In the US, for instance, the wages of college graduates have declined in the past three decades (Autor et al. 2003 and 2008). Many explain the cause as the result of changing rules and policies (Davidson 2011) and of technological

Conclusion

311

change that computers have taken over professional skills like legal assistance, design and software programming that college graduates used to take (Goldin and Katz 2009, Krugman 2011b.) Nonetheless, we hope this study stimulates more institutional context-based residential segregation research. References Alba, R. and Nee, V. 2005. Remaking American Mainstream: Assimilation and Contemporary Immigration. Cambridge, MA: Harvard University Press. Albert, M. 1993. Capitalism vs. Capitalism. New York: Four Wall Eight Windows Press. Amsden, A. 1992. Asia’s Next Giant: South Korea and Late Industrialization. Oxford: Oxford University Press. Anderson, H.T. 2012. The Solidity of Urban Socio-spatial Structures in Copenhagen. This volume. Aoki, M. 2001. Toward a Comparative Institutional Analysis. Cambridge, MA: The MIT Press. Applebaum, R.P. and Henderson, J. eds. 1992. States and Development in the Asian Pacific Rim. London: Sage. Arthur, W.B. 1994. Increasing Returns and Path Dependence in the Economy. Ann Arbor: University of Michigan. Aslund, A. 2010. The Last Shall be the First: The East European Financial Crisis. Washington, DC: Peterson Institute for International Economics. Atkinson, A.B. and Piketty, T. eds. 2010. Top Incomes: A Global Perspective. Oxford, New York: Oxford University Press. Atkinson, A.B., Piketty, T., and Saez, E. 2011. “Top Incomes in the Long Run of History.” Journal of Economic Literature 49(1), 3–71. Autor, D.H., Katz, L.F. and Kearney, M.S. 2008. “Trends in US Wage Inequality: Revising Revisions”. Review of Economics and Statistics 90(2), 300–23. Autor, D.H., Levy, F. and Murnane, R.J. 2003. “The Sill Content of Recent Technological Change: An Empirical Explanation.” Quarterly Journal of Economics 18(4), 1279–333. Ballis Lal, B. 1990. The Romance of Culture in an Urban Civilization: Robert E. Park on Race and Ethnic Relations in Cities. London, New York: Routledge. Barrington, L. 2009. Comparative Politics and Choices. New York: Wadsworth Publishing. Bell, D. A. 2012. What China Can Teach Europe. New York Times, January 8. Page SR5. Bellandi, M. and Caloffi, A. 2008. District Internationalization and Trans-Local Development, Entrepreneurship and Regional Development, 20(6), 517-532. Berger, S. and Dore, R. eds. 1996. National Diversity and Global Capitalism. Ithaca: Cornell University Press. Bjornberg, U. 2010. “Immigrants and the Growth of the City History and Present (1958–2008).” In (Re) searching Gothenburg: Essays on A Changing City, ed. H. Holgersson, C. Thorn, H. Thorn and M. Wahlstrom. Gothenburg: Glanta Production, 199–204. Blgur, B. 2010. The Mobilization of Political Islam in Turkey. Cambridge: Cambridge University Press. Body-Gendrot, S. 2010. “Police, Marginality, Racial Logics and Discrimination in the Banlieues of France.” Ethnic and Racial Studies 33(4), 656–74. Boyer, R. 1997. “French Statism at the Crossroads.” In Political Economy of Modern Capitalism: Mapping Convergence and Diversity, ed. C. Crouch and W. Streek. London: Sage, 71–101.

312

Residential Segregation in Comparative Perspective

Boyer, R. 2002. Is There a Welfare State Crisis? A Comparative Study of French Social Policy. International Labor Office, SES Series, May. http://www.ilo.org/ses. [accessed 25 November 2011]. Boyer, R. and Hollingsworth, R. 1997a. “The Variety of Institutional Arrangements and Their Complementary in Modern Economics.” In Contemporary Capitalism: The Embeddedness of Institutions, eds R. Hollingsworth and R. Boyer. Cambridge, UK: Cambridge University Press, 49–54. Boyer, R. and Hollingsworth, R. 1997b. “From National Embeddness to Spatial and Institutional Nestedness.” In Contemporary Capitalism: The Embeddedness of Institutions, eds J.R. Hollingsworth and R. Boyer. Cambridge: Cambridge University Press, 433–84. Brenner, N., Peck, J., and Theodore, N. 2010. ‘After Neoliberalization?’ Globalizations 7(3): 327–45. Brenner, N. and Theodore, N. 2002. “Cities and the Geographies of Actually Existing Neoliberalism.” Antipode 34(3), 349–79. Byrne, J. ed. 2012. The Occupy Handbook. NY, Boston and London: Back Bay Books. Bullard, R.D., ed. 2007. Growing Smarter: Achieving Livable Communities, Environmental Justice and Regional Equity. Cambridge, MA: MIT Press. Campbell, J.L. 2004. Institutional Change and Globalization. Princeton: Princeton University Press. Clark, W.R. 2005. Capitalism, Not Globalism: Capital Mobility, Central Bank Independence and the Political Control of the Economy. Ann Arbor: University of Michigan Press. Congressional Budget Office (CBO) 2011. Trends in the Distribution of Household Income between 1979 and 2007. October. Washington, DC: CBO. Corina, G. A. 2012. Inequality Trends and Their Determinants: Latin America Over 1990–2010. United Nations University World Institute for Development Economic Research Working Paper, No. 2012/09. http://www.wider.unu.edu/publications/working-papers/2012/en_GB/ wp201209/_files/87048562096144508/default/wp2012-009.pdf [accessed 4 April 2012]. Crouch, C. 2005. “Models of Capitalism.” New Political Economy 10, 439–56. Crouch, C. 2010. “Complementarity.” In The Oxford Handbook of Comparative Institutional Analysis, eds G. Morgan, J.L. Campbell, C. Crouch, et al. Oxford: Oxford University Press, 117–38. Culpepper, P.D. 2001. Employers, Public Policy, and the Politics of Decentralized Cooperation in Germany and France. In P.A. Hall and D. Soskice, eds Varieties of Capitalism. Oxford: Oxford University Press, 275-306. Dancygier, R. 2010. Immigration and Conflict in Europe. Cambridge: Cambridge University Press. Davidson, A. 2011. The Dwindling Power of a College Degree. New York Times, November 23. Appeared on page MM16, under the title of “Why Did the Rules Change,” Sunday Magazine, November 27. De Grauwe, P. 2011. The Governance of a Fragile Eurozone. Department of Economics, University of Leuven. Discussion Paper, April. http://www.econ.kuleuven.be/ew/academic/intecon/ Degrauwe/PDG-papers/Discussion_papers/Governance-fragile-eurozone_s.pdf. [accessed 25 November 2011]. Deeg, R. 2010. Institutional Change in Financial Systems. In The Oxford Handbook of Comparative Institutional Analysis, edited by G. Morgan, J.L. Campbell, C. Crouch, O.K. Pederson, and R. Whitley. Oxford: Oxford University Press, 309-333. Desposato, S.W. 2006. “How Informal Electoral Institutions Shape the Brazilian Legislative Area.” In Informal Institutions and Democracy: Lessons from Latin America, eds G. Helmke and S. Levitsky. Baltimore: The Johns Hopkins University Press, 56–78.

Conclusion

313

Domingues, J.M. 2007. Latin America and Contemporary Modernity: A Sociological Interpretation. New York: Routledge. Domínguez, M., Leal, J. and Goytre, M.E. 2012. The Limits of Segregation as an Expression of Socioeconomic Inequality: The Madrid Case. This volume. Donadio, R. 2011. “Europeans Doubt Greece’s Ability to Stick to Budget.” New York Times June 19. Page A6. Dore, R. 2000. Stockmarket Capitalism: Welfare Capitalism: Japan and Germany versus the Anglo-Saxon. Oxford: Oxford University Press. Dore, R., Lazonick, W. and O’Sullivan, M. 1999. “Varieties of Capitalism in theTwentieth Century”. Oxford Review of Economic Policy 15(4), 102–20. Dragoumis, M. 2005. Greece on the Couch. Athens: Athens News Publication. Drake, S. C. and Cayton, H.R. 1945. Blackmetropolis: A Study of Negro Life in Northman City. Chicago: University of Chicago Press. Dreier, P., Mollenkoph, J. and Swanstrom, J. 2004. Place Matters: Metropolitics for the Twentyfirst Century. Kansas: University Press of Kansas. Second Edition and Revised. Economic Policy Institute 2009. The State of Working America Data on Wages and Compensation Trends by Education. Washington, DC: EPI. Engelen, E. and Konings, M. 2010. Financial Capitalism Resurgent: Comparative Institutionalism and the Challenge of Financialization. In The Oxford Handbook of Comparative Institutional Analysis, eds G. Morgan, J.L. Campbell, C. Crouch, O.K. Pederson, and R. Whitley. Oxford: Oxford University Press, 601-623. Engerer, H. 2001. Privatization and Its Limits in Central and Eastern Europe: Property Rights in Transition. London: Palgrave Macmillan. Esping-Andersen, G. 1990. The Three Worlds of Welfare Capitalism. Princeton, NJ: Princeton University Press. Esping-Andersen, G. 1999. “Politics Without Class: Postindustrial Cleavages in Europe and America.” In Continuity and Change in Contemporary Capitalism, eds H. Kitshelt, P. Lange, G. Marks, et al. Cambridge: Cambridge University Press, 293–316. Estevez-Abe, M., Iversen, T. and Soskice, D. 2001. “Social Protection and the Formation of Skills: A Reinterpretation of the Welfare State.” In Varieties of Capitalism, eds P.A. Hall and D. Soskice. Cambridge: Cambridge University Press, 145–212. Evans, P. 1995. Embedded Autonomy: States and Industrial Transformation. Princeton: Princeton University Press. Fainstein, S., Gordon, I. and Harloe, M. eds. 1995. Divided Cities: New York and London in the Contemporary World. Oxford: Blackwell. Fallerti, T.G. 2010. “Infiltrating the State: The Evolution of Health Care Reform in Brazil, 1964– 1988.” In Explaining Institutional Change: Ambiguity, Agency and Power, eds J. Mahoney and K. Thelen. New York: Cambridge University Press, 38–62. Forman, R.E. 1972. Black Ghettos, White Ghettos and Slums. 1972. NY: Prentice Hall. Friedmann, J. 1986. “The World City Hypothesis.” In World Cities in a World System, eds P. Knox and P. Taylor. NY: Cambridge University Press, 317–31. Originally published in Development and Change 17: 69–84. Fujita, K. and Hill, R.C. 2012a. “Residential Income Inequality in Tokyo and Why It Does Not Translate Into Class-Based Segregation.” This volume. Fujita, K. and Hill, R.C. 2012b. Industry Clusters and Transnational Networks: Japan’s New Directions in Regional Policy. In Locating Neoliberalism in East Asia: Neoliberalizing Space in Developmental States, edited by B. Park, R.C. Hill and A. Saito. Oxford: Wiley-Blackwell,

314

Residential Segregation in Comparative Perspective

27-58.Gal, J. 2010. Exploring the Extended Family of Mediterranean Welfare States or: Did Beveridge and Bismarck Take a Mediterranean Cruise Together? In Children, Gender and Families in Mediterranean Welfare Sates, eds M. Aizenstadt and J. Gal. NY: Springer. Gans, H. 1962. Urban Villages. NY: Free Press. Gilligan, J. 2011. Why Some Politicians Are More Dangerous Than Others. Cambridge: Polity. Ginther, D., Haveman, R. and Wolf, B. 2000. “Neighborhood Attributes as Determinants of Children’s Outcomes: How Robust are the Relationships?” Journal of Human Resources 35(4), 603–42. Gold, B.A. 2007. Still Separate and Unequal: Segregation and the Failure of Urban School Reform. NYC: Teachers College Press. Goldin, C. and Katz, L. 2009. The Race between Education and Technology. Cambridge, MA: Belknap Press. Goldstone, L. 2011. Inherently Unequal: The Betrayal of Equal Rights by the Supreme Court, 1865–1903. NY: Walker & Co. Goode, J. and Schneider, J.A. 1994. Reshaping Ethnic and Racial Relations in Philadelphia Immigrants in a Divided City. Philadelphia: Temple University Press. Hacker, A. 2003. Two Nations: Black and White and Separate, Hostile, and Unequal. NY: Scribner. Hacker, J. and Pierson, P. 2010. Winner-Take All Politics: How Washington Made the Rich Richer— And Turned Its Back on the Middle Class. New York: Simon and Shuster. Haggard, S. and Kaufman, R. 2008. Development, Democracy, and Welfare States: Latin America, East Asia and Eastern Europe. Princeton: Princeton University Pres. Hall, P.A. 1999. “The Political Economy of Europe in an Era of Interdependence.” In Continuity and Change in Contemporary Capitalism, eds H. Kitschelt, et al. Cambridge: Cambridge University Press, 136–63. Hall, P.A. 2010. “Historical Institutions in Rationalist and Sociological Perspective.” In Explaining Institutional Change: Ambiguity, Agency and Power, eds J. Mahoney and K. Thelen. Cambridge: Cambridge University Press, 204–24. Hall, P.A. and Gingerich, G.W. 2009. “Varieties of Capitalism and Institutional Complementarities in the Political Economy: An Empirical Analysis.” In Debating Varieties of Capitalism: A Reader, ed. B. Hancke. Oxford: Oxford University Press, 135–79. Hall, P.A. and Soskice, D. eds. 2001a. Varieties of Capitalism: The Institutional Foundations of Comparative Advantage. Oxford, New York: Oxford University Press. Hall, P.A. and Soskice, D. 2001b. “An Introduction to Varieties of Capitalism.” In Varieties of Capitalism: The Institutions of Comparative Advantage, eds P.A. Hall and D. Soskice. Oxford, New York: Oxford University Press, 1–68. Hall, P.A. and Thelen, K. 2009. “Institutional Change in Varieties of Capitalism.” Social and Economic Review 7(1), 7–34. Hamnett, C. 1996. “Social Polarization, Economic Restructuring and Welfare State Regimes.” Urban Studies (Edinburgh, Scotland) 33(8), 1407–30. Hancké, B. 2001. “Revising the French Model: Coordination and Restructuring in French Industry.” In Varieties of Capitalism: The Institutional Foundations of Comparative Advantage, eds P.E. Hall and D. Soskice. Oxford: Oxford University Press, 307–33. Hancké, B., ed. 2009. Debating Varities of Capitalism: A Reader. Oxford: Oxford University Press. Hardman, A. and Ioannides, Y.M. 2004. “Neighborhood Income Distribution: Economic Segregation and Mixing in US Urban Neighborhoods.” Journal of Housing Economics 13, 368–82.

Conclusion

315

Hartman, C. and Squire, G. eds. 2009. The Integration Debate: Competing Futures for American Cities. New York: Rutledge. Harvey, D. 2005. A Brief History of Neoliberalism. Oxford: Oxford University Press. Hatch, W. and Yamamura, K. 1996. Asia in Japan’s Embrace: Building a Regional Production Alliance. Cambridge: Cambridge University Press. Henderson, J. and Applebaum, R.P. 1992. “Situating the State in the East Asian Development Process.” In States and Development in the Asian Pacific Rim., eds R.P. Applebaum and J. Henderson. London: Sage. Herbert, B. 2011. Separate and Unequal. New York Times, March 22, page 27. Héritier, A. and Rhodes, M. eds. 2011. New Modes of Governance in Europe: Governing in the Shadow of Hierarchy. London: Palgrave Macmillan. Herrigel, G. and Zeitlin, J. 2010. Inter-Firm Relations in Global Manufacturing: Disintegrated Production and Its Globalization. In The Oxford Handbook of Comparative Institutional Analysis, edited by G. Morgan, J.L. Campbell, C. Crouch, O.K. Pederson, and R. Whitley. Oxford: Oxford University Press, 527-561. Hill, R. and Fujita, K. 2003. “The Nested City: Introduction.” Urban Studies (Edinburgh, Scotland) 40(2), 207–17. Hollingsworth, J.R. and Boyer, R. eds. 1997a. Contemporary Capitalism: The Embeddedness of Institutions. Cambridge: Cambridge University Press. Hollingsworth, J.R. and Boyer, R. 1997b. “Coordination of Economic Actors and Social Systems of Production.” In Contemporary Capitalism: The Embeddness of Institutions, eds J.R. Hollingsworth and R. Boyer. Cambridge: Cambridge University Press, 1–48. Hsing, Y. 2010. The Great Urban Transformation: Politics of Land and Property in China. Oxford: Oxford University Press. Hughes, E.C. and Hughes, H.M. 1952. Where Peoples Meet: Racial and Ethnic Frontiers. Chicago: University of Chicago Press. Hunter, W. 2010. The Transformation of the Workers’ Party in Brazil, 1989–2009. Cambridge: Cambridge University Press. Iceland, J., Weinberg, D. and Stainmetz, S. 2002. Racial and Ethnic Residential Segregation in the United States: 1980–2000. Census 2000 Special Reports. US Census Bureau. Series CENSR-3. Washington, DC: U.S. Government Printing Office. Institute of Supply Management. 2012. December 2011 Manufacturing ISM Report on Business. January 3. http://www.ism.ws/ISMReport/MfgROB.cfm. Internal Revenue Sservice (IRS). 2012. SOI Tax Stats-Individual Income Tax Rates and Tax Shares. The US government. http://www.irs.gov/taxstats/indtaxstats/article/0,,id=129270,00.html [accessed 10 January 2012]. Iverson, T. 2001. “The Dynamics of Welfare State Expansion: Trade Openness, De-industrialization, and Partisan Politics.” In The New Politics of the Welfare State, ed. P. Pierson. Oxford: Oxford University Press, 45–79. Iverson, T., Pontussen, J. and Soskice, D. eds. 2000. Unions, Employers and Central Banks. Cambridge: Cambridge University Press. Jessop, B. ed. 1991. The Politics of Flexibility: Restructuring State and Industry in Britain, Germany and Scandinavia. Aldershot: Edward Elgar. Johnson, C. 1982. MITI and the Japanese Miracle: The Growth of Industrial Policy, 1925–1975. Stanford, CA: Stanford University Press. Jones, A.F. and Weinberg, D.H. 2000. The Changing Shape of the Nation’s Income Distribution: 1947–1998. US Census Bureau Current Population Report, June: 60–204.

316

Residential Segregation in Comparative Perspective

Kawachi, I. and Berkman, L. 2003. Neighborhoods and Health. Oxford: Oxford University Press. King, D. 1995. Separate and Unequal: Black Americans and the US Federal Government. NY: Oxford University Press. Kitschelt, H., Lange, P. Marks, G. eds. 1999. Continuity and Change in Contemporary Capitalism. Cambridge: Cambridge University Press. Knox, P. and Taylor, P. eds. 1986. World Cities in a World System. NY: Cambridge University Press. Kovács, Z. 2012. “Residential Segregation in Budapest: Before and After Transition.” This volume. Kozol, J. 2005. The Shame of the Nation: The Restoration of Apartheid Schooling in America. NY: Random House. Krugman, P. 1991. Trade and Geography. Cambridge, MA: MIT Press.Krugman, P. 2006. Graduate versus Oligarchs. New York Times. February 27. http://select.nytimes.com/2006/02/27/ opinion/27krugman.html [accessed 20 January 2010]. Krugman, P. 2007. The Conscience of a Liberal. New York: W.W. Norton. Krugman, P. 2011a. Oligarchy, American Style. New York Times. November 4, A31. Krugman, P. 2011b. Degrees and Dollars. New York Times. March 6. Kume, I. and Thelen, K. 1999. “The Rise of Non-market Training Regimes: Germany and Japan Compared.” Journal of Japanese Studies 25, 33–64.Laszlo, U. 2011. “Privatization as Institutional Change in Hungary.” In The Political Economy of Property Rights: Institutional Change and Credibility in the Reform of Centrally Planned Economies, ed. D.I. Weimer. Cambridge: Cambridge University Press, 239–55. Levitsky, S. 2007. “From Populism to Clientelism? The Transformation of Labor-based Party Linkages in Latin America.” In Patrons, Clients and Policies: Patterns of Democratic Accountability and Political Competition, eds H. Kitschelt and S.I. Wilson. Cambridge: Cambridge University Press, 206–26. Logan, J.R. 2011a. Separate and Unequal: The Neighbourhood Gap for Blacks, Hispanics and Asians in Metropolitan America. US 2010: Discover America in a New Century. http://www. s4.brown.edu/us2010/Data/Report/report0727.pdf [accessed 25 July 2011]. Logan, J.R. 2011b. Whose Schools are Failing? US 2010: Discover American in a New Century. http://www.s4.brown.edu/us2010/Data/Report/report5.pdf [accessed 15 July, 2011]. Logan, J.R. and Li, L. 2012. “The Impact of Housing Tenure on Residential Segregation in Beijing, China.” This volume. Logan, J.R. and Stults, B.J. 2011. The Persistence of Segregation in the Metropolis: New Findings from the 2010 Census. http://www.s4.brown.edu/us2010/Data/Report/report2.pdf [accessed 21 June 2011]. Logan, J.R., Stults, B.J. and Farley, R. 2004. “Segregation of Minorities in the Metropolis: Two Decades of Change.” Demography 41(1), 1–22. Logan, J.R. and Zhang, C. 2010. “Global Neighborhoods: New Pathways to Diversity and Separation”. American Journal of Sociology 115, 1069–109. Logan, J.R. and Zhang, C. 2011. Global Neighborhoods: New Evidence from Census 2010. US 2010 Census: Discover America in a New Century. http://www.s4.brown.edu/us2010/Data/ Report/globalfinal2.pdf [accessed 4 November 2011]. Loriaux, M. 1999. “The French Developmental State as Myth and Moral Ambition.” In The Developmental State, ed. M. Woo-Cumings. Ithaca: Cornell University Press, 235–75. Lynn, M. 2011. Bust: Greece, the Euro and the Sovereign Debt Crisis. New York: Bloomberg Press. Mahoney, J. 2000. “Path Dependence in Historical Sociology”. Theory and Society 29, 507–48.

Conclusion

317

Mahoney, J. and Thelen, K. eds. 2010. Explaining Institutional Change: Ambiguity, Agency and Power. New York: Cambridge University Press. Maloutas, T, Arapoglou, V, Kandylis, G. and Sayas, J. 2012. “Polarization and De-segregation in Athens.” This volume. Manolopoulos, J. 2011. Greece’s ‘Odious’ Debt: The Looting of the Hellenic Republic by the Euro, the Political Elite and the International Community. Athens: Anthem Press. Marlier, E., Atkinson, A.B., Cantillon, B., et al. 2009. The EU and Social Inclusion: Facing the Challenge. London: Policy Press. Marques, E., Bichir, R. and Scalon, C. 2012. “Residential Segregation and Social Structure in São Paulo: Continuity and Change since the 1990s.” This volume. Martin, M.E. 2006. Residential Segregation Patterns of Latinos in the United States, 1990–2000. New York: Routledge. Massey, D. and Denton, N. 1993. American Apartheit: Separation and the Making of the Underclass. Cambridge, MA: Harvard University Press. Matthews, F.H. 1976. Quest for American Sociology: Robert E. Park and the Chicago School. Toronto: University of Toronto Press. Matthews, F.H. 1989. “Social Scientists and the Cultural Concept, 1930–1950: The Conflict between Processual and Structural Approaches.” Sociological Theory 7, 87–101. Mauro, F.D., Dees, S. and McKibbin, W.J. eds. 2011. Globalization, Regionalism, and Economic Interdependence. Cambridge: Cambridge University Press. McCarthy, N., Poole, K. and Rosenthal, H. 2008. Polarized America: The Dance of Ideology and Unequal Riches. Cambridge, MA: MIT Press. McKenzie, R.D. 1924. “The Ecological Approach to the State of Human Community.” American Journal of Sociology 30, 287–301. Meyerson, H. 2011. The Fallacy of Post-industrial Prosperity. Washington Post. September 8. Milanovic, B. 2011. The Haves and Haves-Nots: A Brief History and Idiosyncratic History of Global Inequality. New York: Basic Books. Mollenkopf, J. and Castells, M. eds. 1992. Dual City: The Restructuring of New York. New York: Russell Sage Foundation. Molotch, H. and Noren, N. eds. 2010. Toilet: Public Restrooms and the Poitics of Sharing. NY: NYU Press. Morgan, G., Campbell, J.L., Crouch, C., Pederson, O.K., and Whitley, R. eds. 2010. The Oxford Handbook of Comparative Institutional Analysis. Oxford: Oxford University Press. Morgan, G., Whitley, R. and Moen, E. eds. 2005. Changing Capitalism? Institutionalism, Institutional Change and Systems of Economic Organizations. Oxford: Oxford University Press. Muller, T. and Smets, P. 2009. “Welcome to the Neighborhood: Social Contacts between Iraqis and Natives.” Local Environment: The International Journal of Justice and Sustainability 14(5), 403–15. Musterd, S. 2005. “Social and Ethnic Segregation in Europe: Levels, Causes and Effects.” Journal of Urban Affairs 37(3), 331–48. Musterd, S., Marie, A. and Kestloot, C. eds. 2006. Neighborhoods of Poverty: Urban Social Exclusion and Integration in Europe. London: Palgrave Macmillan. Musterd, S. and Ostendorf, W.J.M. eds. 1998. Urban Segregation and the Welfare State: Inequality and Exclusion in Western Cities. London: Routledge. Myant, M. and Drahokoupil, J. 2010. Transition Economies in Russia, Eastern Europe and Central Asia. London: Wiley.

318

Residential Segregation in Comparative Perspective

Naughton, B. 2007. The Chinese Economy: Transition and Growth. Cambridge, MA: The MIT Press. Noah, T. 2010. The United States of Inequality: Why We Can’t Ignore Growing Income Inequality. Slate, September 17. http://www.slate.com/id/2266025/ [accessed 10 January 2011]. Norris, F. 2012. Manufacturing is Surprising Bright Spot in U.S. Economy. January 6. New York Times, Page B1. North, D.C. 1990. Institutions, Institutional Change and Economic Performance. Cambridge, New York: Cambridge University Press. Obama, B. 2011. Transcript: Obama’s Speech to Congress on Jobs. New York Times, September 8. http://www.nytimes.com/2011/09/09/us/politics/09text-obama-jobs- speech.html?pagewanted =6&adxnnl=1&ref=politics&adxnnlx=1315580403- Os6JdAd/Ax40g2j3UwNszA [accessed 9 September 2011]. OECD Program for International Student Assessment (PISA) 2010. PISA 2009 Results: What Students Know and Can Do. http://www.oecd.org/edu/pisa/2009 [accessed 25 March 2011]. Onoma, A.K. 2010. The Politics of Property Rights Institutions in Africa. Cambridge: Cambridge University Press. Palier, B. 2005. “Ambiguous Agreement, Cumulative Change: French Social Policy in the 1990s.” In Beyond Continuity: Insitutional Change in Advanced Political Economies, eds W. Streek and K. Thelen. Oxford: Oxford University Press, 127–44. Park, R.E. 1915. “The City: Suggestions for the Investigation of Behaviors in the City.” American Journal of Sociology 20, 579–83. Park, R.E. Burgess, E. and McKenzie, R.D. 1925. The City. Chicago: University of Chicago Press. Payne, C.M. 2008. So Much Reform, So Little Change: The Persistence of Failures in Urban Schools. Cambridge, MA: Harvard Education Press. Peck, J. 2010. Constructions of Neoliberal Reason. Oxford: Oxford University Press. Peck, J. and Tickell, A. 2002. “Neoliberalizing Space.” Antipode 34(3), 380–404. Pederson, O.K. 2006. Denmark: An Ongoing Experiment. In National Identity and the Variety of Capitalism, edited by J.L. Cambpell, J.A. Hall, and O.K. Pedersen. Montreall: McGill & Queens University Press, 453–70. Pempel, T.J. 1999. “The Developmental Regime in a Changing World Economy.” In The Developmental State, ed. M. Woo-Cumings. Ithaca: Cornell University Press, 137–81. Philpott, T.L. 1991. The Slum and the Ghetto: Immigrants, Blacks, and Reforms in Chicago, 1880– 1991. New York: Wadsworth Pub. Co. Pichett, W.E. 2003. Brownville, Brooklyn: Blacks, Jews, and the Changing Face of the Ghetto. Chicago: University of Chicago Press. Pickvance, C. 2012. “The Limits of Neoliberalism: Is the Concept of Neoliberalism Helpful in the Study of Urban Policy?” In Neoliberal Urbanism and Its Contestations, eds J. Kunkel and M. Mayer. London: Palgrave, 46–62. Pierson, P., ed. 2001. The New Politics of the Welfare States. Oxford: Oxford University Press. Pierson, P. 2004. Politics in Time: History, Institutions and Social Analysis. Princeton: Princeton University Press. Piketty, T. and Saez, E. 2003. Income Inequality in the US: 1913–1998. Quarterly Journal of Economics 118(1), 1–39. Updated to 2008 (July 2010). See Saez’s web site for the updated study at http://elsa.berkeley.edu/~saez/ [accessed 2 October 2010]. Piore, M.J. and Sabel, C. 1984. The Second Industrial Divide: Prospect for Prosperity. NY: Basic Books.

Conclusion

319

Polanyi, K. 1943. The Great Transformation: The Political and Economic Origin of Our Time. Boston: Beacon Press. Polikoff, A. 2007. Waiting for Gautreaux: A Story of Segregation, Housing and the Black Ghetto. Chicago: Northwestern University Press. Porter, M.F. 1998. Comparative Advantage of Nations. NY: Free Press. Préteceille, E. 2012. “Segregation, Social Mix and Public Policies in Paris.” This volume. Rajan, R.G. 2010. Fault Lines: How Hidden Fractures Still Threaten the World Economy. Princeton, Oxford: Princeton University Press. Reardon, S.F. and Bischoff, K. 2011. Growth in the Residential Segregation of Families by Income, 1970-2009. US2010: discover aerica in a new century. Brown University. http://www.s4.brown. edu/us2010/Data/Report/report111111.pdf [accessed 24 December 2011]. Rhodes, M. 1997. Southern European Welfare States: Identity, Problems and Prospects for Reform. In Southern European Welfare States: Between Crisis and Reform, ed M. Rhodes: NY: Routledge, 1-22. Roberts, B.R. and Wilson, R.H. 2009. Urban Segregation and Governance in the Americas. New York: Palgrave Macmillan. Rodrik, D. 2011. The Globalization Paradox: Democracy and the Future of the World Economy. New York: W.W. Norton. Sabel, C. 1989. “Flexible Specialization and the Re-Emergence of Regional Economics.” In Reversing Industrial Decline: Industrial Structure and Policy in Britain and in her Competitors, eds P. Hirst and J. Zeitlin. Oxford: Berg, 17–70. Sabel, C. and Zeitlin, J. 1997. World of Possibilities: Flexibility and Mass Production in Western Industrialization. Cambridge: Cambridge University Press. Sassen, S. 1991. The Global City. Princeton: Princeton University Press. Sassen, S. 2002. “Introduction: Locating Cities on Global Circuits.” In Global Networks, Linked Cities, ed. S. Sassen. New York. London: Routledge, 1–37. Scheffer, P. 2011. Immigrant Nations. London: Polity. Scheppele, K.M. 2011. Hungary’s constitutional revolution. In Paul Krugman’s Hungary’s constitutional revolution. New York Times. December 19. http://krugman.blogs.nytimes. com/2011/12/19/hungarys-constitutional-revolution/#more-27489 [accessed 19 December 2011]. Schnell, I. and Ostendorf, W.J.M. eds 2002. Studies in Segregation and Desegregation. Surrey: Ashgate. Schneider, B.R. 1999. “The Desarrollista State in Brazil and Mexico.” In The Developmental State, ed. M. Woo-Cumings. Ithaca: Cornell University Press, 276–305. Skidelsky, R. 2011. Lumpy Labor, Project Syndicate, May 19. http://www.project-syndicate.org/ commentary/skidelsky41/English [accessed 19 May 2011]. Soskice, D. 1999. “Divergent Production Regimes: Coordinated and Uncoordinated Market Economies in the 1980s and 1990s.” In Continuity and Change in Contemporary Capitalism, eds H. Kitschelt, P. Lange, G. Marks, et al. Cambridge, New York: Cambridge University Press, 101–34. Stallings, B., ed. 1995. Global Change, Regional Response. Cambridge: Cambridge University Press. Streek, W. 2009a. Institutional Change in Advanced Political Economies. In Debaring Varieties of Capitalism: A Reader, ed B. Hancke. Oxford University Press, 95–133. Streek, W. 2009b. Re-Forming Capitalism: Institutional Change in the German Political Economy. Oxford: Oxford University Press.

320

Residential Segregation in Comparative Perspective

Streek, W. 2010. “Institutions in History: Bringing Capitalism Back in.” In The Oxford Handbook of Comparative Institutional Analysis, eds G. Morgan, J.L. Campbell, C. Crouch, et al. Oxford: Oxford University Press, 659–86. Streek, W. 2011. A Plan for Growth. Thecurrentmoment, September 7. http://thecurrentmoment. wordpress.com/author/thecurrentmoment/ [accessed 7 September 2011]. Streek, W. and Thelen, K. 2005. “Introduction: Institutional Change in Advanced Political Economies.” In Beyond Continuity: Institutional Change in Advanced Political Economies, eds W. Streek and K. Thelen. Oxford: Oxford University Press, 1–39. Street, P. 2005. Still Separate, Unequal: Race, Place, Policy and the State of Black Chicago. Chicago: Chicago Urban League. Szelenyi, I. and Wilk, K. 2009. “Institutional Transformation in European Post-Communist Regimes.” In The Oxford Handbook of Comparative Institutional Analysis, eds G. Morgan, J.L. Campbell, C. Crouch, et al. Oxford: Oxford University Press, 565–86. Taeuber, K.E. and Taeuber, A.F. 1969. Negroes in Cities: Residential Segregation and Neighborhood Change. NYC: Atheneum. Tasan-Kok, T. 2012. “Changing Dynamics of Residential Segregation in Istanbul.” This volume. Taylor, P.J. 1995. “World Cities and Territorial States: The Rise and Fall of Their Maturity.” In World Cities in a World System, eds P. Knox and P.J. Taylor. Cambridge: Cambridge University Press, 48–62. Taylor, P.J. 2004. World City Network: A Global Urban Analyses. London: Routledge. Taylor, P.J., Derudder, B. Saey, P. and Witlox, F. eds. 2006. Cities in Globalization: Practices, Policies and Theories. London and New York: Routledge. Taylor, P.J., Ni, P., Drudder, B., Huang, J. and Hoyler, M. eds. 2010. Global Urban Analysis: A Survey of Cities in Globalization. London: Earthscan Publications. Thelen, K. 2004. How Institutions Evolve: The Political Economy of Skills in Germany, Britain, the United States and Japan. Cambridge: Cambridge University Press. Thelen, K. and Streek, W. 2009. “Beyond Comparative Statics: Historical Institutional Approaches to Stability and Change in the Political Economy of Labor.” In The Oxford Handbook of Comparative Institutional Analysis, eds G. Morgan, J.L. Campbell, C. Crouch, et al. Oxford: Oxford University Press, 41–62. Therborn, G. 2011. “End of a Paradigm. The Current Crisis and the Idea of Stateless Cities.” Environment & Planning A 43(2), 272–85. Thomas, E.I. and Znaniecki, F. 1918. The Polish Peasant in Europe and America: Monograph of an Immigrant Group. Chicago: University of Chicago Press. Tickell, A. and Peck, J. 2003. “Making Global Rules: Globalization or Neoliberalization?” In Remaking the Global Economy: Economic-Geographical Perspective, ed. H.W. Yeung. London: Sage, 163–81Tinker, H. 1976. Separate and Unequal. NYC: C Hurst & Co Publishers. Tsai, K.S. 2007. Capitalism without Democracy: The Private Sector in Contemporary China. Ithaca: Cornell University Press. Turner, A.T., Popkin, S. and Rawlings, L.A. 2008. Public Housing and the Legacy of Segregation. Washington, DC: Urban Institute Press. United Nations 2009. Human Development Report 2009. http://hdr.undp.org/en/media/ HDR_2009_EN_Complete.pdf [accessed 20 February 2011]. United States Congress Senate Committee 2010. America’s Schools: Providing Equal Opportunity Or Still Separate and Unequal? Washington, DC: BiblioGov.

Conclusion

321

US 2010 Census: Discover America in a New Century 2011. Spatial Structure in the Social Sciences, Brown University. http://www.s4.brown.edu/us2010/segregation2010/Default.aspx [accessed 25 June 2011]. van Kempen, R. 1994. “The Dual City and the Poor: Social Polarization, Social Segregation and Life Chances.” Urban Studies (Edinburgh, Scotland) 31(7), 995–1015. Vogel, S. 2006. Japan Remodeled: How Government and Industry are Reforming Japanese Capitalism. Ithaca: Cornell University Press. Wade, R. 1990. Governing the Market: Economic Theory and the Role of Government in East Asia Industrialization. Princeton, NJ: Princeton University Press. Wade, R. 1996. “Globalization and its Limits: Reports of the Death of the National Economy are Greatly Exaggerated.” In National Diversity and Global Capitalism, eds S. Berger and R. Dore. Ithaca: Cornell University Press, 60–88. Walter, C.E. and Howie, F.J.T. 2011. Red Capitalism: The Fragile Financial Foundation of China’s Extraordinary Rise. Oxford, New York: Wiley. Wang, C.H. and Li, C.H. 2012. “A Portrait of Residential Differentiation in Taipei City (1980– 2010).” This volume. Watson, T. 2009. “Inequality and the Measurement of Residential Segregation by Income in American Neighborhoods.” Review of Income and Wealth Series 55(3), September, 820–44. Weimer, D.L. 2011. The Political Economy of Property Rights: Institutional Change and Credibility in the Reform of Centrally Planned Economies. Cambridge: Cambridge University Press. Weinberg, D.H. 1996. A Brief Look at Postwar US Income Inequality. US Census Bureau Current Population Report, June, 60–191. Weinberg, D.H. 2011. US Neighborhood Income Inequality in the 2005–2009 Period. American Community Survey Reports, October. Weiss, L. 2010. “The State in the Economy: Neoliberal or Neoactive?” In The Oxford Handbook of Comparative Institutional Analysis, eds G. Morgan, J.L. Campbell, C. Crouch, O.K. Pederson, and R. Whitley, eds Oxford: Oxford University Press, 83–210. Whyte, M. 2010a. Myth of the Social Volcano: Perceptions of Inequality and Distributive Injustice in Contemporary China. Stanford: Stanford University Press. Whyte, M., ed. 2010b. One Country, Two Societies: Rural-Urban Inequality in Contemporary China. Cambridge, MA: Harvard University Press. Wilkinson, R. and Pickett, K. 2009. The Spirit Level: Why More Equal Societies Almost Always Do Better? London: Allen Lane. Wilson, W.J. 1987. The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy. Chicago: University of Chicago Press. Wilson, W.J. 2009. More than Race: Being Black and Poor In the Inner City. New York: W.W. Norton. Wirth, L. 1928. The Ghetto. Chicago: The University of Chicago Press. Wirth, L. 1938. “Urbanism as a Way of Urban Life.” American Journal of Sociology 44, 1–44. Woo-Cumings, M., ed. 1999. The Developmental State. Ithaca: Cornell University Press. Wood, S. 2001. “Labor Market Regimes under Threat? Sources of Continuity in Germany, Britain and Sweden.” In The New Politics of the Welfare State, ed. P. Pierson. Oxford: Oxford University Press, 368–409. Wright, T. 2010. Accepting Authoritarian State-Society Relations in China’s Reform. Stanford: Stanford University Press. Yip, N.M. 2012. “Residential Segregation in an Unequal City: Why There Are No Urban Ghettos in Hong Kong?” This volume.

322

Residential Segregation in Comparative Perspective

Zeitlin, J. 2007. Industrial Districts and Regional Custers. In The Oxford Handbook of Business History, ed G. Jones and J. Zeitlin, Oxford: Oxford University Press, 219-243. Zysman, J. 1996. “The Myth of a “Global” Economy: Enduring National Foundations and Emerging Regional Realities.” New Political Economy 1(2), 157–84.

Index age 5, 13, 24, 42, 57–8, 72–3, 80, 82–3, 95, 137, 139, 180, 187, 209–10, 221–2, 225, 232 Amsterdam 288 area effect 10, 13, 19, 20 see also neighborhood effect Athens 16, 22, 26–30, 129, 138, 144, 257–71, 273–9, 285, 289, 291–4, 299–302, 306–308 banks 41, 58, 168, 277 banlieue 24, 153–4, 156, 164, 174 Beijing 21–2, 29, 69–80, 83–4, 285, 289, 291–7, 306–8 Berlin 212, 220 Budapest 25, 29, 197–212, 285, 289, 292, 294, 297, 306–8 capitalism coordinated 289–90, 293, 297, 299, 302–3 varieties of 10, 12, 289, 310 capitalist globalization 1–2, 14, 28–9 see also globalization society 10–11, 13, 57, 212, 274, 288 caste 294–295 Chicago 4, 5, 15–17, 24, 163, 167, 173, 180, 194–5, 200, 217, 278, 285 Chicago School 4, 15–17, 167, 180, 194–5, 217, 285, 289 China 22, 69, 70–71, 73–9, 81, 83, 95, 114–15, 120–21, 123, 130, 294–6, 302, 306 cities European 2, 5, 14, 16–17, 19–20, 22, 37, 90, 193–5, 197–8, 212 global see global city; see also world city Indian 8 world, see world city; see also global city city center 25, 27, 80, 82–3, 91, 122, 170, 172, 180, 194, 198, 200, 207–8, 227, 229, 231, 234, 245, 247, 258, 262–3, 271–4, 277, 296, 302, 306 clientelism 2, 19, 26–7, 29, 37, 59, 276, 279, 289, 292–3, 295, 300, 302, 306, see also populism colonial 22, 89, 91–2, 113, 292, 295–6, 300, 303 communism 197, 200–02, 204, 206, 209, 210

Chinese 123, 294 Hungarian 297 communist legacy 292, 297, 303 party 123, 170, 200, 292, 294 context 1–7, 9–10, 13, 15–18, 20–21, 27–8, 30, 37, 90, 100, 129–31, 136, 139, 157, 186, 230, 238–9, 257–9, 278–9, 285, 287, 291, 305–6, 310–11 contextual 1–3, 6, 9–12, 14–19, 21, 27–30, 130, 259, 279, 287–88, 290 convergence 28, 304–6, 310 Copenhagen 22, 24, 28–9, 177–80, 182, 186–9, 191, 193–5, 285, 289–94, 297–9, 301, 306–8 crisis 20, 25, 49, 59, 96, 172, 191, 199–200, 219–20, 222, 234, 247, 258, 260, 277, 279, 301–3, 307 culture 20, 37, 63–4, 120–21, 125, 174, 180, 194, 237, 287, 289–90, 298, 304, 308 cultural industries 122, 174, 204 deindustrialization 16, 24, 37, 96, 138, 178, 187, 307 Denmark 180, 182, 187, 190, 195, 283, 298–9, 307 deprivation 11, 22, 24, 26–7, 30, 97–8, 136, 234 see also poverty deregulation 1, 48, 126, 301, 304–5, 309 see also regulation developmental state East Asian 2, 12, 29, 128, 130, 291, 295, 300, 305 Hong Kong 289, 293, 300 Japanese 21, 62, 63, 293, 303 Latin American 296 Taiwanese 22, 111–12, 119, 121, 123, 125, 128, 130–32, 293, 302 discrimination 5, 10–11, 13, 16, 21, 24, 84, 120–21, 129, 162, 168–9, 171, 278–9 see also inequality ethno-racial 11–12, 14–16, 18, 22, 28, 37, 112, 129, 132, 162–3, 168–9, 173, 274 dissimilarity index 7, 15, 22, 73, 89, 97–9, 113, 116, 120, 125, 139–40, 149, 162–3, 180, 189, 224, 228, 267–71, 279, 286 distance see social distance; spatial distance

324

Residential Segregation in Comparative Perspective

divergence income 96 national 289, 304–5 racial 129 diversity 162 contextual 10, 14, 19 cultural 237 ethnic 11, 26, 266, 286 housing 13, 28–9 institutional 296 multi-group 89 national 289, 304 residential strategies 225 socioeconomic 28, 234 divided city 27, 90–91, 173, 267, 306 see also dual city dual city 23, 90, 156–7, 174, 260, 306 see also divided city economic restructuring 1, 3, 18, 24, 28, 178, 204–5, 259, 279 education 6, 20–21, 25, 37, 50–51, 57–62, 64, 69–74, 76, 79–81, 83–4, 90–91, 95–7, 100, 112–13, 115–16, 119–20, 125, 129, 131–2, 136–7, 139, 145–9, 157, 171, 177–8, 180, 182, 186–7, 194–5, 201, 206, 210, 212, 218, 222–4, 231, 238, 242–3, 246, 248, 250–51, 258, 261, 268, 278, 285, 286, 290, 294–6, 298–9, 301, 303, 308–10 see also schools embeddedness 2, 3, 9, 14, 16, 59, 182, 193, 259, 289–91, 296, 303–4, 306–7 see also nestedness enclaves 22, 25, 37, 69–72, 89, 92, 102, 137, 195, 205, 270–71, 299 European Socio-economic Classes (ESeC) 6, 261 exchange 3, 10, 26, 41, 58, 120, 126, 191, 239–40, 244, 249 exclusion see social exclusion factorial ecology 8, 69 familism 291, 293, 300, 303 family businesses 260 centered welfare system 2, 19, 27, 29, 258, 276, 300 income 135, 145, 148, 309–10 networks 3, 14, 19, 234, 246, 276, 278, 287, 291, 308 size 8, 180, 310 solidarity 26–7, 276 status 8, 90, 92 strategies 278

favela 137, 143, 152 filtering-down 25, 209 finance 42, 50, 52–3, 57–8, 119, 122, 127, 138, 174, 222, 249–50, 259, 289 , 298, 304, 305 financial see market, financial flexibility 234, 237, 249, 260, 275, 304, 307 Fordism 5, 257, 260, 279. France 6–7, 19, 20, 90–91, 155–6, 159, 161–2, 165–6, 168–72, 299, 301, 308 gated communities 23, 25–6, 29–30, 69–70, 72, 137–8, 140, 142, 150, 211–12, 218, 238–42, 248, 251–3, 295 gecekondu 238, 240–42, 246, 248 gender 20, 59, 80, 155, 221, 258 gentrification 1, 9, 16–17, 19, 25–6, 30, 38, 102, 128, 166, 170–71, 177, 195, 209, 212, 225, 229, 241–2, 244–5, 273–4, 277–9 Germany 7, 91, 187, 271, 285, 289, 298–9, 308 ghetto 5, 6, 15–16, 18, 20, 28, 89, 90–91, 99, 101–2, 153, 156, 161, 165, 173–4, 193, 195, 247, 260, 270, 286, 288 see also ghettoization hyper- 5 ghettoization 177, 212 see also ghetto and stigmatization global city 1–2, 11, 128, 130, 138, 160–61, 165–6, 260, 279, 304–5, 310 economy 37, 91, 94, 237–8, 296, 307, 310 globalization 1, 2, 11, 13, 14, 23–5, 28–9, 37, 50, 130, 136, 166, 219, 239, 241, 259–61, 279, 288, 304–10 governance 72, 91, 124, 289, 290, 304–5 government 12, 38–9, 42–4, 47–50, 59–64, 71–3, 75–7, 81–2, 92, 100, 102, 111, 113–14, 119–28, 130–32, 137–138, 170–72, 177–80, 186, 193, 195, 198, 209, 221, 229, 139, 241, 243–4, 247–52, 261, 290, 294–8, 300–4, 306 local government, 38, 60, 62–3, 76, 120, 123–4, 138, 177–8, 186, 193, 209, 229, 244, 248, 250–51, 261 Greece 187, 190, 257–60, 266, 271, 279, 301–2, 306 homeownership 13, 22, 25, 27, 44–5, 71, 92, 97–9, 102, 128, 172, 187, 191, 194, 222, 232–3, 252, 259, 275, 277 see also owner occupation and tenure Hong Kong 22, 29–30, 52–53, 89–108, 119, 123, 128, 144, 285, 289–91, 293–4, 299–301, 306–8

Index households 10, 13, 18, 24, 27, 29, 43, 59, 80, 89, 92–102, 116, 118, 120, 123, 127–8, 131, 148, 154–5, 165–6, 169–70, 187, 194, 204, 206, 209–10, 212, 218, 222, 225, 229–34, 244, 252, 272, 274, 276, 278, 286, 288, 300, 308–9 housing 2–3, 10–11, 13–15, 17–19, 21–7, 29–30, 40–41, 47–8, 59, 62, 69–71, 73, 76–84, 89– 94, 96–7, 99–103, 122–3, 125–32, 137–8, 154, 156, 165–72, 177–80, 182, 186–7, 190–95, 197–203, 205–6, 208–12, 217–19, 222–9, 231–4, 237–42, 245–52, 257–9, 270, 273–7, 279, 286–97, 299–303, 309 classes 206 demand 2, 13, 123, 178, 202, 234, 276 illegal see illegal settlements market 3, 10, 13, 15, 17–19, 24–7, 29–30, 62, 71, 74, 84, 102, 122–3, 127–9, 186, 191–4, 200, 203, 206, 209, 211–12, 222, 225–9, 234, 257, 270, 274–6, 287–8, 294, 297, 300, 303, 309 parental 13, 63, 222–3, 234, 276, 301 policy 3, 19, 22–4, 62, 64, 89, 99–101, 111–12, 122–3, 125–8, 130–32, 177, 179, 193, 201–2, 209–11, 277, 286, 293, 301, 303 prices 14, 23, 25, 40–41, 47–8, 53, 62, 80–81, 84, 111, 116–20, 123, 126–7, 131, 165–6, 180, 218, 222, 226, 277, 302 public see public housing rental, private 26, 71, 76–8, 83–4, 92, 96, 99, 168, 191, 194, 201, 212, 232–3, 272, 276 see also rent and tenure rental, public 71, 76–9, 80, 82, 84, 92–3, 97, 99, 101–2, 177, 191, 194, 201, 203, 274 see also rent and tenure supply 2, 22, 26, 169–72, 191, 199, 201, 219, 225, 233, 241, 245, 273–4, 276, 279 tenure see tenure human ecology 4, 207 Hungary 187, 198, 199–205, 207–8, 210–12, 297 illegal settlements/housing 23, 26, 136, 138, 168, 237–40, 246–9, 275–6, 295 immigration/immigrants 5, 15, 24–30, 37, 69–71, 73–5, 78, 80, 82–4, 95, 102, 120–21, 125, 128–9, 132, 156, 161, 164, 167–73, 178, 182, 186–7, 190–91, 193–4, 203, 219, 222, 224, 228, 231–4, 237–40, 242, 244–8, 253, 257–8, 260–62, 266–7, 269–74, 276–7, 279, 285, 287–8, 294–5, 298–302, 304, 306–8, 310 see also migrants

325

income 5–6, 12–14, 21–6, 28, 37–64, 70, 76, 83–4, 89–102, 113, 116, 118–19, 123, 127–9, 131–2, 135, 139, 145–50, 157, 165–6, 177–8, 180–82, 186, 190–95, 197, 201–2, 204–6, 210–11, 218, 220–26, 229–30, 232, 234, 238–48, 250–53, 285–8, 291–2, 294–7, 299–303, 307–10 polarization 95, 308–9 index 8, 135, 139, 141, 145, 147–8, 195, 204, 222–3 autocorrelation 7–9, 159 dissimilarity see dissimilarity index exposure 5, 7–8 isolation 7–8 Moran’s I see Moran’s I segregation see segregation index industrial 5, 15–17, 22, 27, 30, 41–2, 54, 59–60, 63, 71, 94, 111–13, 119, 121–2, 124–8, 130–32, 136, 145, 159–60, 166, 170, 177–9, 182, 187, 197, 199, 204, 219–20, 223, 228, 244–5, 247, 257–8, 260, 273, 275, 290–91, 299, 303–4, 306–7, 309–10 decline 17, 24, 37, 96, 138, 140, 142, 178, 187, 219, 258, 304, 307 post see post-industrial industry 1, 23, 50, 52–3, 57–8, 60, 63, 71, 94, 119, 121–2, 127, 131, 138, 159, 166, 170, 174, 178, 182, 187, 197, 200, 204, 221, 257–60, 263, 268–70, 274, 279, 295, 302, 308 inequality 2, 6, 10–7, 19, 21–6, 28, 30, 37–64, 89–102, 129, 136–8, 146–50, 154, 165–6, 170, 173–4, 194–5, 197, 200–05, 212, 217–24, 227–9, 233–4, 237–9, 244, 248, 250, 253, 274, 278–9, 285–7, 289, 292–303, 308–9 see also discrimination and social inequality infrastructure 23, 59, 100, 102, 136, 154, 166, 178, 194, 204, 210, 219, 227, 237, 249–50, 258, 272, 277, 296, 300, 302–3 inner city 15, 18, 37, 71, 75, 102, 157, 177, 182, 186–7, 191, 194 institutional 2, 3, 11, 13, 42,43, 56, 138, 169, 194, 195, 243, 245, 259, 285, 286, 287, 288, 289, 290, 291, 292, 295, 296, 297, 298, 299, 301, 303, 304, 305, 306, 307, 308, 309, 310, 311 change 286, 291, 295, 303–7, 309–10 complementarity 290–91, 294–5, 299, 301, 306 nestedness see nestedness institutions 3, 5, 11–3, 37, 41–3, 48, 58, 59–63, 64, 84, 89, 102, 130, 138, 156, 166, 167, 168–9, 174, 178–80, 194, 195, 243, 245, 259, 261, 285, 286, 287–301, 303–11 invasion 28, 242, 247, 257

326

Residential Segregation in Comparative Perspective

ISCO 6, 224 Istanbul 26, 28, 30, 237–53, 259, 285, 289, 292–7, 302, 306–8 Japan 37–8, 41–4, 48–50, 56–64, 113, 120, 123, 289, 302–3, 308 Keynesianism 305 labor casualization 155 see also precariousness force 50–51, 59, 94, 138, 210, 220, 237, 246, 249, 258, 299, 307 market 2, 5, 10–14, 21, 26, 27, 29, 59, 96, 102, 129, 135, 138, 139, 156, 166–8, 177, 182, 187, 189, 191, 194, 195, 204, 209, 222, 249, 257, 261, 270, 274, 287–91, 293–6, 298–302, 307, 310 land market 136–8, 295 prices 41, 44–50, 52–3, 62, 71, 172, 303 landlords 96, 168, 200 landowners 21, 258–9, 275, 277, 300 liberal 21, 157–8, 178, 217, 240, 248, 289–90, 293, 300, 305 liberalism 17–8, 21, 25, 123, 130, 200–1, 206, 212, 239–41, 248–9, 259, 298, 305 see also neoliberalism London 2, 17, 50, 128, 138, 144, 153, 166, 219 Madrid 16, 22, 25, 27, 28, 30, 138, 217–34, 285, 289, 291, 292, 293, 294, 299–302, 306, 307, 308 managers 6, 13, 21–4, 44, 50–54, 57–9, 62, 94, 96–8, 100–1, 189, 194, 197, 205, 219–21, 225–31, 234, 245–6, 250, 261–5, 268, 269, 270, 274, 300–301, 307, 310 marginalization 16, 26, 186, 187, 191, 195, 209, 253, 262 see also social exclusion marginalized neighborhoods/spaces 26, 37, 191, 161, 193, 209, 253, 262, 270 see also social exclusion market 2, 3, 5, 10–19, 21–2, 24, 25–7, 29, 30, 48, 50, 58–60, 62–3, 69–71, 74, 76, 84, 92–3, 95–6, 100, 102, 122–3, 126–30, 135–9, 145, 156, 166–70, 172–3, 177, 182, 186–7, 189, 191, 193–5, 197, 200, 203–6, 208–9, 211–12, 222, 225, 229–30, 234, 238–41, 243, 247–50, 253, 257, 259–61, 268, 270, 272, 275–6, 286–91, 293–307, 309–10 dynamics 127,136, 253

financial 21–3, 44–9, 58–9, 94, 101–2, 121, 126–7, 136, 165–6, 204–5, 219, 234, 237, 239, 246, 248–50, 253, 258–9, 274, 278, 292, 297–8, 301, 303–5, 307 forces/mechanisms 5, 10, 11, 27, 212 housing see housing market labor see labor market land see land market Marxist 154 media 20, 23, 42, 50, 94, 121, 153, 155–6, 158–9, 161, 171, 173–4, 186, 195, 237, 259, 277 Mediterranean familism see familism, see also family centered welfare system middle class 17–18, 20, 23, 27, 29–30, 42, 89, 99, 100–102, 123, 128, 131, 137, 149, 154, 157, 161, 165–7, 169–71, 173, 177–8, 182, 191, 193–5, 200, 210, 223, 225, 227, 231, 258, 260, 272–3, 277, 298–9, 301 migrants 25–7, 70–75, 78, 80, 82–4, 190, 203, 240, 247–8, 258, 273–4, 276, 294–5 see also immigration/immigrants migrant enclaves, see enclaves Moran’s I 9, 139–40, 143, 159 mortgage 25, 44, 59, 118, 126–8, 172, 222, 232–3, 251 neighborhood 4, 5, 7, 10, 13–14, 19–20, 23, 27, 37–8, 50–51, 54, 56–7, 61–2, 64, 69, 70–72, 76, 79–80, 83–4, 90–91, 101–2, 139, 150, 154–7, 159–69, 171–4, 177, 182, 186, 191, 193–5, 199–200, 206–12, 217–8, 223–5, 233, 238, 240–53, 261, 266, 271, 273–74, 278–9, 285–8, 290, 292, 294–301, 303, 307–9 effect 10, 14, 19, 172, 218, 224, 278 see also area effect marginalized see marginalized neighborhoods/ spaces stigmatized see stigmatization neoliberalism 1–2, 19, 26, 43, 64, 165, 172, 204–5, 239, 241, 259–60, 304–6, 310 see also liberalism nestedness 1–2, 16–17, 64, 111, 128, 130–1, 288–9, 302–6, 310 see also embeddedness New York 5, 24, 50, 163, 166, 173, 286, 287 Nordic 166, 182, 195, 289, 298, 301 social democratic capitalism 289 see also varieties of capitalism occupation 1–2, 5–6, 16, 21–8, 37–8, 40, 43, 50–52, 54–7, 64, 80, 89, 94, 96, 119, 135, 139, 141, 148, 157, 187–9, 219–20, 222, 224–7, 234,

Index 238, 242, 249–50, 258, 262–3, 265, 267–71, 274–6, 278–9, 301, 303 occupational categories 6, 16, 21, 24–7, 50–51, 188–9, 220, 222, 224–7, 234, 262–3, 267–9, 271, 275–6, 278–9 occupational status 5, 135, 139, 141, 148, 224 owner-occupation 22, 26, 101, 179, 272 see also homeownership and tenure Paris 2, 15, 16, 23, 24, 28, 29, 128, 138, 142, 144, 153–60, 162–74, 219, 22, 285, 288–91, 293, 294, 297–300, 306–8 parents 26, 162, 164, 167, 195, 222–3, 234, 276, 299, 301 home see housing, parental lone/single 96, 99–101, 190, 225 path dependency 2, 14, 22, 25, 28, 132, 212, 237–9, 253, 259, 291–2, 297, 303 periphery 23, 25–8, 41, 70–71, 78, 80, 83, 170, 178, 198, 200, 209–10, 227, 231–3, 236, 242, 246, 250, 257–8, 262, 266, 271–2, 275–6, 295–6, 301 polarization 1, 2, 8, 11, 19, 21–3, 26–8, 38, 94–6, 119, 128–9, 131, 160, 166, 169, 191, 203–5, 210, 221, 234, 257, 260, 274, 279, 297, 300, 303–4, 306–10 metropolitan 304, 306, 308–10 social 1–2, 19, 26, 28, 38, 119, 128–9, 131, 204, 210, 257, 279, 303, 306 spatial 1, 21, 166, 257, 309–10 thesis 1–2, 11, 27, 128–9, 131, 138 policy 2–3, 5, 9, 17–8, 20, 42, 49–50, 59–60, 62–4, 84, 90–91, 99–102, 111–12, 119–28, 130–32, 137, 149, 153–5, 170–73, 178, 187, 194–5, 201–2, 204, 209, 211–2, 217, 237, 247, 261, 277, 288, 290–94, 296, 298, 300–307 housing see housing policy local 2, 22, 123, 126, 138, 173, 304 urban 9, 16, 24, 60, 154–5, 170, 182, 217, 274, 277, 290, 293, 296, 300, 303, 305–6 populism 27, 247, 292, 296 see also clientelism Latin American 292 post-industrial 1, 17, 219, 273, 307 poverty 11, 13, 19–20, 24, 26, 37, 91, 101, 135–6, 138, 144, 150, 156, 169, 171, 174, 187, 190, 205, 209, 249, 278, 286, 288, 294, 299 precariousness 25, 155,159–60, 164, 225, 234 see also labor casualization privatization 25, 84, 130, 204, 206, 209, 212, 239, 241, 248–9, 259, 298, 305

327

professionals 6, 13, 16, 22–5, 52, 70, 73, 75, 96, 98, 100–101, 119, 122, 125, 131, 139, 140–42, 144–6, 154, 157–9, 168, 201, 208, 219–22, 225–6, 228–30, 234, 245, 261, 263, 268–70, 274, 294, 297, 301, 307, 310 professionalization 96, 119, 128, 260 public housing 13, 18–9, 21–2, 25, 29–30, 62, 70–71, 76, 89, 92–4, 97, 99–103, 122–3, 127–8, 156, 166, 168–72, 179, 191–2, 197, 200, 206, 211, 217, 259, 289, 291, 300 see also tenure regionalism 288, 291–2, 310 East Asian 291, 302, 305 Mediterranean 291 Nordic 291, 305 regional cluster economies 307 regulation 1, 3, 5, 9, 11–13, 15, 19, 21, 23, 25, 29, 60, 124, 126–7, 136–8, 173, 195, 238–9, 241, 243–4, 248, 258–9, 277, 289–91, 298, 300, 304–5 see also deregulation religion 237–8, 240, 242–3, 245–6, 292, 294–5, 303 rent 13, 30, 43–45, 47–48, 52, 64, 74, 76, 77–78, 93, 118, 127–8, 131, 169, 177, 191, 193–4, 199–201, 208–9, 233, 247, 252, 259, 275–6, 295, 302 see also housing, rental; tenure rented housing see housing, rental; tenure representation 15, 19, 153, 155–6 residential patterns 37, 70, 74, 84, 111, 113, 180, 189, 289, 299, 307 see also spatial patterns segregation 1, 3–5, 10–3, 18, 22–3, 27–8, 30, 37–8, 42, 61, 64, 69–71, 73, 79, 84, 89–90, 101–2, 111, 121, 125, 129, 131–2, 135, 138–9, 149, 167–8, 180, 197–8, 201–3, 206, 209, 212, 224, 228, 234, 238–40, 242, 250, 252–3, 257, 261, 267, 276, 285–311 Roma 205, 209, 250, 262, Rome 5 routine occupations 261–2, 265, 270 rural 25–6, 63, 71, 73–6, 78, 91, 140, 145–7, 178, 200, 210, 219–20, 237, 240, 242, 244–5, 247–8, 250, 257–8, 271, 274–6, 294–5 São Paulo 23, 28, 30, 135–44, 146–50, 285, 289, 292–4, 296–7, 306–8, schools 4, 13, 17–8, 20, 27, 30, 58, 61–2, 82, 156, 158, 167, 168–9, 171, 173, 194–5, 229, 278, 285–86, 287, 290, 299–300, 309 see also education segregation see also residential segregation

328

Residential Segregation in Comparative Perspective

desegregation 5, 15, 26–7, 30, 159, 161, 227–8, 234, 269, 273, 275 ethnic/ethnoracial 1, 5–6, 8–10, 12, 14, 16, 22, 24, 27, 84, 153, 161–4, 168–9, 173, 238, 243–4, 266, 269 hyper 5, 8 increasing 1, 2, 19, 29, 138, 161, 170, 211, 217, 224, 228, 234 index 4–8, 11, 14, 16–17, 19, 22, 26, 69, 79, 89, 97–9, 155, 157–8, 162–3, 167–8, 182, 189, 200, 227–9, 232, 234, 279 measures 5–6, 7, 8, 15, 294 social/class 5–6, 10, 16, 22, 24, 27, 37–8, 40, 43, 56–7, 60, 63–4, 69, 135, 144, 170, 195, 200, 224, 243–4, 262–3, 267, 303 vertical see vertical segregation service jobs 50, 84, 249, 307 industries 50, 119, 274, 307 services 1, 3, 7, 10, 13, 17, 20–23, 25, 30, 37, 50, 53, 57, 59–60, 63–4, 70, 82, 94, 119, 135–7, 145, 150, 166, 169–71, 173–4, 187, 201, 204–5, 207, 210–12, 219–21, 226–32, 234, 237, 245–7, 249–50, 258–60, 263, 266, 268, 270, 276, 278, 285, 287, 292, 296, 304 shifting and sorting 10–11, 13, 16–17, 21, 25–6, 29, 275–7 social area analysis, 5, 8 categories 6–7, 23–4, 27, 154–5, 161, 165–7, 170, 198, 218, 220–26, 229, 232, 234, 271 class 5, 40, 56–7, 70–71, 135, 139–40, 144, 146–8, 150, 154–5, 224, 238, 296, 302–3 differentiation 177, 182, 194, 197, 201, 205, 226, 272 distance 4, 11, 15, 23, 26–7, 30, 150, 160, 223, 276, 279 division 2, 178, 226 exclusion 90, 102, 156, 224 groups 5–7, 17–18, 22, 26, 29–30, 135–7, 139, 144, 147, 149, 161, 165–7, 173–4, 190, 194, 218, 223–4, 226, 230, 237–8, 240, 243–4, 248, 250, 253, 278–9, 297 hierarchy 1, 6–8, 15, 24, 225, 274, 278 housing 13, 18, 23–5, 29, 128, 166–72, 191–2, 200, 217, 259, 276 see also housing, rental public inequality 6, 10, 16, 21, 24–5, 30, 64, 99, 129, 137, 173, 197, 200–202, 212, 217, 223–4, 278 see also inequality

mix 13, 17, 19, 21, 27, 29, 54, 90–91, 99–100, 102, 137, 153, 161, 166–7, 170–73, 199, 224, 229, 234, 273, 288, 301, 306 mobility 5, 15, 19–20, 25–7, 37, 42, 129, 169, 195, 222, 230, 242, 247–8, 253, 258, 261, 274–9, 286, 288, 294, 296, 299, 301, 303 movements 127, 137, 153–4, 173, 306 patterns 180, 219, 230 status 50, 70, 169, 198–200, 203, 210, 225, 239–40, 248, 250 structure 11, 16, 21–2, 24, 135, 138–9, 141, 144–5, 149, 153–5, 161, 174, 178, 195, 203, 212, 218–20, 253, 258, 262, 274, 304 socialism 18–9, 25, 71, 170, 197–8, 200 socialist 10, 18, 21, 25, 69–71, 84, 197–8, 201, 203, 205–7, 212, 266, 297, 304 Spain 163, 187, 190, 222, 224, 232–3, 301–2 spatial autocorrelation 9, 159 see also index; Moran’s I distance 11, 15, 28–30, 159, 223, 229, 234, 268–9, 279 division 1–3, 15–16, 24, 89, 101, 226, 229 patterns 22, 71, 80, 91, 139, 144, 177, 197, 206, 239, 242, 244, 247, 307 see also segregation patterns units 4–5, 7–9, 15, 21, 38, 50–51, 148, 219 sprawl 16, 219, 257 squatter 26, 238–42, 246–50, 252–3, 295–7 State local see local government welfare see welfare state stigmatization 20, 89–90, 159, 169, 288 suburb 21, 24, 27, 38–41, 43–51, 60–2, 70–78, 83–4, 154, 156–9, 166, 170–71, 177–80, 182, 186–7, 191, 193–5, 198, 200–201, 203, 205, 207, 209–10, 212, 258, 262, 266, 276, 279, 288, 298–9, 303, 308 suburbanization 15–6, 71, 178, 180, 200, 203, 210, 212, 260, 262, 271, 273, 275–6 Taipei 22–3, 111–32, 144, 285, 289, 291, 293–4, 302 Taiwan, 111, 113–18, 120–32 tax 21, 37–8, 43–4, 54, 59–64, 124, 126, 128, 170, 177, 233, 240, 244, 259, 275, 295, 297–8, 300–301, 303, 308–10 tenure 11, 13, 21–2, 64, 69, 71, 73, 76–84, 90, 94, 97–9, 101, 138, 177, 180, 182, 186, 191, 193–5, 203, 299 see also homeownership, rented housing and public housing theory 9–10, 111, 128–31, 163, 289–90, 304 see also urban theory

Index Tokyo 2, 21, 28–9, 37–64, 285, 289–91, 293–4, 302–3, 306–7 transition 2, 69, 71, 124, 137, 197, 203–4, 206, 209, 239–40, 248, 260, 271, 289, 294, 297, 304, 310 Turkey 26, 162, 237–9, 242, 246–7, 253, 266, 295, 306 UK 91, 99, 101, 190, 271, 288–9 unemployment 25, 60, 63–4, 96, 120, 138, 144, 150, 155–6, 160, 164–5, 171, 178, 182, 187, 191, 204–5, 218, 222, 248, 268, 285, 290, 299, 301 unskilled workers 96, 101, 178, 187, 220–21, 225–6, 228, 268–9, 290, 299, 307, 310 upper class, 126, 142, 157, 159–61, 165–7, 172–4, 177–8, 182, 200, 221, 298–9 urban development 2, 59–60, 111, 122–3, 125–6, 129, 131, 154, 173, 197–9, 212, 241, 248–9, 251, 277, 302 policy 9, 16, 24, 60, 154–5, 170, 217, 274, 277, 290, 293, 296, 300, 303, 305–6 theory 1, 218, 285, 310

329

urbanization 91, 127, 158, 178, 197, 200, 212, 257, 276, 279, 304 US 1–3, 5, 7–8, 13–6, 18–20, 28, 30, 57, 60, 101, 153–4, 156, 158, 161–3, 166, 168, 172–3, 190, 285–9, 294, 296, 301, 307–10 varieties of capitalism 10, 12, 289, 310 vertical segregation, 27, 224, 272–4, 276, 279, 300 welfare regimes 12–13, 19–20, 24, 30, 90, 138, 195 state, 2, 12–14, 18–19, 21, 24, 26, 29, 37, 57, 63, 91, 129, 186, 189–90, 195, 212, 217–18, 246, 257–8, 287–9, 291–4, 296–9, 301, 303, 305, 309–10 white flight 16, 25, 161, 169 working class 15–18, 20, 22–3, 26–7, 29, 128, 136, 154, 156, 159–61, 163, 166–74, 177–9, 182, 193–5, 197, 200–201, 209, 227, 230, 232–4, 258, 260, 262, 264, 266, 268, 271, 273–6, 278–9, 296–9, 301, 308 world city 128, 237, 304 see also global city zones of affluence 70