Social and Ethnic Segregation Manifestations, Understanding, Impacts, Responses Seminar on Mobility, Segregation and Neighbourhood Change Tartu, 14-15 March 2013 Sako Musterd Urban Geography / Urban Studies University of Amsterdam The Netherlands
Four issues 1. Segregation research: a changing focus: from classic description and index construction of segregation to analysis of inequality and concentrations; from a focus on residential segregation to analysis in non-residential domains (work, leisure, public space related). 2. How can we understand the variation of inequalities? 3. What can we say about the effects or impacts of urban inequality? 4. What are the typical policy responses to urban inequality in European cities? 2
1a Social inequality in selected EU cities; segregation indexes as classic tools for describing urban inequality USA Rochester NY MSA poor USA 100 largest cities poor Antwerp 'poor' USA Portland OR MSA poor Rotterdam 5th quintile Amsterdam 5th quintile Copenhagen 10th decile Milan professionals Sheffield unemployed Leeds unemployed Birmingham unemployed Oslo social assistance Amsterdam 5th quintile metro Rotterdam 1st quintile Lille unemployed vs employed Berlin hh income > 3500 Amsterdam 1st quintile Manchester unemployed Manchester income support Milan blue collar workers Birmingham income support Berlin hh income < 900 Bern unemployed Amsterdam 1st quintile metro Copenhagen 1st quintile 0 20 40 60
These overviews provide much information Low levels and lower in EU than in the US State dependent: higher in Belgium, low in Denmark and the Netherlands City dependent: high in Leeds, lower in Manchester Group dependent: higher levels in higher social classes; lower class in EU still in one system with middle class
Social Segregation Patterns (problems with indicator selection, but still information rich) Amsterdam as an example
Social Segregation Pattern in 1960 Social rank (pca on classic indicators)
Social Segregation Pattern in 2007 University education
Social Segregation Pattern in 2007 m 2 Price for Housing
Social Segregation Low Medium High Pattern in 2012 test scores for entry in secondary education 9
Micro-level Social Segregation Pattern 2011 concentrations high (red) and low (blue) real estate values in housing http://regiomonitor-uva.everimap.com/ 10
1b Ethnic inequality: classic segregation measures (D) and recent change (difference between years mentioned) Leeds Bangladeshi 1990-2000 London Bangladeshi 1990-2000 Birmingham Bangladeshi 1990-2000 Barcelona Moroccans 1995-2003 Leeds Pakistani 1990-2000 London Indians 1990-2000 London Black Carribean 1990-2000 Rotterdam Turks 1994-2009 Rotterdam Moroccans 1994-2009 Amsterdam Turks 1994-2009 Amsterdam Moroccans 1994-2009 Birmingham Indians 1990-2000 Amsterdam Surinamese 1994-2009 Barcelona Peruvians 1995-2003 Cologne Turks 1990-2000 Rotterdam Surinamese 1994-2009 Cologne Yugoslavs 1990-2000 Cologne Italians 1990-2000 most recent year difference -40-20 0 20 40 60 80 100 Sources: Arbaci (2007); Friedrichs (1998) ; Malheiros (2002a, b); Martori et al. (2005); Musterd (2005) (1996); Peach (1996); Stillwell & Phillips (2006)
Development of index of dissimmilarity in several European cities 12
Again these overviews provide much information Varied levels, usually lower in EU than in the US State dependent: higher in UK and Belgium, lower in Germany and the Netherlands City dependent: higher in Leeds, lower in Birmingham Group dependent: higher levels for Bangladeshi, lower for Indian origin.
Changing Patterns of Segregation Moroccan origin 1973 1981 1990
Moroccan 2011(similar classification relative to average) 15
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Type of index is relevant: D dissimilarity upper half 0,60 0,50 0,40 0,30 xpx* isolation lower half 0,20 0,10 1994-2009 0,00 Surinamese Antillean Turkish Moroccan Surinamese Antillean Turkish Moroccan Amsterdam Rotterdam
Micro-level Ethnic Segregation Pattern 2011 concentrations of Moroccan (2-4sd; >4sd) 17
Here too, a move away from classic measures of segregation to microlevel data used for constructing concentrations and for describing and analysing them 18
Amsterdam, Surinamese origin, 2009, n > 50, > 2sd Concentrations: Surinamese in 2009 Total in the city: 68761 (9.09%) N Surinamese per area: >= 50 Perc. Surinamese in concentrations >= 17.83 110 concentrations with (total): 90087 inhabitants 30227 Surinamese (33.6%) 30227: 43.9% of all Surinamese 562.4 hectares 19
Amsterdam, Surinamese origin, 2009, n > 50, > 4sd Concentrations: Surinamese in 2009 Total in the city: 68761 (9.09%) N Surinamese per area: >= 50 Perc. Surinamese in concentrations >=26.52 61 concentrations with (total): 67539 inhabitants 25427 Surinamese (37.6%) 25427: 36.9% of all Surinamese 419.4 hectares 20
Amsterdam, Surinamese origin, 2009, n > 50, > 50% Concentrations: Surinamese in 2009 Total in the city: 68761 (9.09%) N Surinamese per area: >= 50 Perc. Surinamese in concentrations >=50% 5 concentrations with (total): 3245 inhabitants 1848 Surinamese (56.9%) 1848: 2.7 % of all Surinamese 27.9 hectares 21
Amsterdam, Surinamese origin, 2009, n > 50, > 60% Concentrations: Surinamese in 2009 Total in the city: 68761 (9.09%) N Surinamese per area: >= 50 Perc. Surinamese in concentrations >=60% 2 concentrations with (total): 753 inhabitants 469 Surinamese (62.3%) 469: 0.7% of all Surinamese 6.1 hectares 22
Concentrations (4sd+) of ethnic categories in Amsterdam 2009 1 2 3 4 5 6= (4/5)*100 7=(4/2)*100 Ethnic category Category s city population Percentage of category in city Category s concentration population Total concentration population Percentage of category in concentrations Category s concentration population relative to category s city population Turkish 39654 5.2 16102 68532 23.5 40.6 Moroccan 68099 9.0 32446 91235 35.6 47.6 Surinamese 68761 9.1 23511 61710 38.1 34.2 Antillean 11559 1.5 2010 16895 11.9 17.4
Concentrations (4sd+) of ethnic categories in Amsterdam 2009 1 2 3 4 5 6= (4/5)*100 7=(4/2)*100 Ethnic category Category s city population Percentage of category in city Category s concentration population Total concentration population Percentage of category in concentrations Category s concentration population relative to category s city population Turkish 39654 5.2 16102 68532 23.5 40.6 Moroccan 68099 9.0 32446 91235 35.6 47.6 Surinamese 68761 9.1 23511 61710 38.1 34.2 Antillean 11559 1.5 2010 16895 11.9 17.4
Share of four population categories in the city of Amsterdam, 2000-2011, based on country of origin (1st and 2nd generation) (column 3)
Concentrations (4sd+) of ethnic categories in Amsterdam 2009 1 2 3 4 5 6= (4/5)*100 7=(4/2)*100 Ethnic category Category s city population Percentage of category in city Category s concentration population Total concentration population Percentage of category in concentrations Category s concentration population relative to category s city population Turkish 39654 5.2 16102 68532 23.5 40.6 Moroccan 68099 9.0 32446 91235 35.6 47.6 Surinamese 68761 9.1 23511 61710 38.1 34.2 Antillean 11559 1.5 2010 16895 11.9 17.4
Percentage of each of the population categories relative to total population, in concentrations of these population categories (column 6) 2000-11
Concentrations (4sd+) of ethnic categories in Amsterdam 2009 1 2 3 4 5 6= (4/5)*100 7=(4/2)*100 Ethnic category Category s city population Percentage of category in city Category s concentration population Total concentration population Percentage of category in concentrations Category s concentration population relative to category s city population Turkish 39654 5.2 16102 68532 23.5 40.6 Moroccan 68099 9.0 32446 91235 35.6 47.6 Surinamese 68761 9.1 23511 61710 38.1 34.2 Antillean 11559 1.5 2010 16895 11.9 17.4
Percentage of each of the population categories that lives in concentrations of these population categories (column 7) (2000-11)
2 Structural How to understand urban inequalities? Globalisation Internationalisation Economic restructuring produces inequality polarisation mismatch Pathways Institutional Path dependency (chain migration); urban structure ((un) fit for current urban economies; multi-layered or single-layered) Welfare regime type and change Other Behaviour, culture, tolerance, discrimination, political discourse, choice 30
Context: levels of social inequality, within and between states. Gini-index 2008, black: sharp rise in past decade Lithuania Latvia Portugal Romania Bulgaria United Kingdom Greece Poland Estonia Italy Spain Ireland Germany Belgium France Netherlands Luxembourg Finland Austria Czechia Hungary Denmark Sweden Slovakia 0 5 10 15 20 25 30 35 31 40
Levels of social inequality, 2008 32
Polarisation or integrated social classes share of total income per income decile, The Netherlands, 2010 33
Welfare states and levels of social inequality in European cities residual extensive universal hybrid Source: Domburg 2005 Segregation index low ses 34
3 Segregation and concentration receive much attention because of expected negative effects on participation 35
Urban inequalities, effects on participation; theory Mechanisms Socialisation, role models, peer groups Stigmatisation Social networks Opportunity for encountering others In Europe, hypotheses Lower levels of segregation small effects Perhaps even positive effects 36
Effects of urban inequalities, methods Qualitative in-depth research Quantitative large datasets, longitudinal Both rich in terms of variables Detailed geo-coding available Quantitative research using techniques that make efforts dealing with selection effects Quantitative research enables detection of non-linearities, thresholds, etc. 37
Examples of results of some own (co-researched) large-scale longitudinal, individual level studies in Sweden and The Netherlands on the impact of neighbourhood composition (various sizes and compositions (ethnic, social, tenure)) on social outcomes (mostly income) 38
A selection of (mostly recent) neighbourhood effect studies (co) authored Musterd, S. (1996) Ruimtelijke segregatie en sociale effecten. Inaugurele rede. Assen: Van Gorcum. Ostendorf, W., S. Musterd & S. de Vos (2001) Social mix and the neighbourhood-effect: policy-ambition and empirical support. Housing Studies, vol. 16, no. 3, pp. 371-380. Musterd, S., W. Ostendorf & S. de Vos (2003) Environmental Effects and Social Mobility. Housing Studies. Vol. 18 6. pp. 877-892. Andersson, R., Musterd, S., Galster, G. and Kauppinen, T. (2007) What Mix Matters? Exploring the Relationships between Individual s Incomes and Different Measures of their Neighbourhood Context. Housing Studies, 22 (5), pp. 637-660. Musterd, S., Andersson, R., Galster, G. and Kauppinen, T. (2008) Are Immigrants Earnings Influenced by the Characteristics of their Neighbours? Environment and Planning A. 40, pp. 785-805. Galster, G., Kauppinen, T., Musterd, S. and Andersson, R. (2008). Does Neighborhood Income Mix Affect Earnings of Adults? A New Approach using Evidence from Sweden. Journal of Urban Economics. 63, pp. 858-870 Musterd, S. (2008) Residents Views on Social Mix: Social Mix, Social Networks and Stigmatisation in Post-war Housing Estates. Urban Studies, Volume 45, No. 4, pp. pp. 897-915. Galster, G., R. Andersson & S. Musterd (2010) Who Is Affected by Neighbourhood Income Mix? Gender, Age, Family, Employment and Income Differences. Urban Studies. 47(14), pp. 2915-2944. Andersson R. & S. Musterd (2010) What Scale Matters? Exploring the Relationships between Individuals Social Position, Neighbourhood Context and the Scale of Neighbourhood. Geografiska Annaler: Series B, Human Geography 92 (1), pp. 23-43. Sykes, B. & S. Musterd (2011) Examining Neighbourhood and School Effects Simultaneously: What Does the Dutch Evidence Show? Urban Studies, 48(7), pp. 1307-1331. Musterd, S., Vos, S. de, Das, M. & Latten, J.J. (2012). Neighbourhood Composition and Economic Prospects: A Longitudinal Study in the Netherlands. Tijdschrift voor Economische en Sociale Geografie, 103(1), 85-100. Musterd, S., G. Galster & R. Andersson (2012) Temporal dimensions and measurement of neighbourhood effects. Environment and Planning A 44(3) 605 627. Andersson, R., S. Musterd & G.Galster (2013) Neighbourhood Ethnic Composition Effects On Immigrant Incomes. Accepted by Journal of Ethnic and Migration Studies. 39
Sweden 1.Concentrations of low income households have negative effects on individual s social mobility. 2.Effects are generally stronger in metropolitan than in non-metropolitan areas and stronger for males than for females. 3.Own group ethnic concentrations can initially pay dividends for immigrants, but these benefits turn into disadvantages over time, after approx. two years. 4.The impact of other immigrants turned out positive only if unemployment levels are very low. 40
Sweden continued 5.For males who are not employed full time, or have a low income, middle income neighbours have a positive marginal impact (relative to either high- or low-income neighbours). 6.Increases in middle-income neighbours will have a negative effect on high-income males if they substitute for high-income neighbours. 7.Smallest scale areas (10,000 m 2 ) have strongest effects. 8.Recent, and continued cumulative exposure yields stronger associations than lagged, temporary ones. 9.There is distinct time decay (but some persistence) in the potential effects after exposure ceases. 41
The Netherlands 10.Higher median neighbourhood income is positively related to individual income prospects. 11.Income mix in the neighbourhood likely has a small positive effect. 12.A high share of rented dwellings relates to higher income gain than a high share of owner-occupied dwellings (possibly because the rented sector functions as a springboard rather than a trap). 42
4 Typical policy responses to urban inequalities in European cities Fear for lack of integration and assumed relation with spatial inequalities Call for forced interventions to obtain reduction of segregation Policies aimed at creating balanced communities and social mix Instruments: urban restructuring and mixed tenure, mixed qualities and mixed house prices, gentrification 43
Theoretical considerations; Empirical findings Macro and managerial; social mix policy 1.prevents the development of uncontrollable neighbourhoods and rising segregation 2.is a conspiracy against the poor : it produces gentrification, favours the affluent and displaces the poor 44
L. Lees (2008) [i]t is ironic that a process that results in segregation and polarization gentrification is being promoted via social mix policies as the positive solution to segregation the rhetoric of social mixing tends to conceal the inequalities of fortune and economic circumstance that are produced through the process of gentrification 45
Theoretical considerations; Empirical findings Macro and managerial; social mix policy 1.prevents the development of uncontrollable neighbourhoods and rising segregation 2.is a conspiracy against the poor : it produces gentrification, favours the affluent and displaces the poor 3.helps to improve local conditions and service levels 4.creates negative spill-over effects (displacement) 5.reduces stigmatisation through social mix 6.increases stigmatisation through selection of areas 7.reduces the accumulation of neighbourhood problems 8.improves neighbourhood liveability 46
Theoretical considerations; Empirical findings Micro and individual; social mix policy 1.helps the poor to realize social mobility through socialisation, role models, peers, and better social networks through weak ties 2.stimulates economic, social and cultural participation and integration 3.destroys individual s local social networks; spatial proximity is no guarantee for social integration 4.frustrates individual s behaviour especially of those who can afford to sort into relatively homogeneous neighbourhoods 5.neglects metropolitan and state level impacts on social compositions 47
Comments on predominant policy interventions Social mix may indeed help social prospects Social mix may be good for some, but harming others. Segregation and social mix are often weakly defined, if at all, and mostly used as metaphores, which blurs discussions. Fear often drives policies, which is a bad thing. Social mix does not work when micro-level social distances and social inequality are too large. Social mix may create injustice effects (due to reduction of affordable housing). 48
Conclusions and discussion Differences between cities relate to urban histories, welfare regimes,and opportunities to adapt to structural economic transformation and globalisation. These differences produce different levels of social/ethnic (spatial) inequality. Which results in governmental intervention, strongest in universal welfare states Caution is required: Interventions are good for some but bad for others Don t pathologise the poor or social housing; Lupton and Tunstall (2008, p. 114): [i]t is a short step from there to the assumption that the solution is to reduce the proportion of poor people, rather than to address structural inequalities or inject additional resources to provide the services that people need 49
Conclusions and discussion Continued Caution is required: Interventions are good for some but bad for others Don t pathologise the poor or social housing Spatial sorting is a strong process, difficult to engineer Interventions may add to stigma Interventions may produce negative externalities (demolition, spillovers and fear) Moderate segregation associates with universal welfare states; however, almost all of these states appear to be heading towards more neo-liberal models Consider the potential of structural and institutional factors and interventions at various scales aimed at services delivery (schools) and enhanced individual labour market participation and social mobility 50
Conclusions and discussion Continued Avoid cynical attitudes and neglect of potential impact of concentrated poverty; this may produce more segregation and marginalisation of the poor Recognise the complexity and reciprocal relations and (social) ties between people, place, space and scale Address more comprehensive measures of individual and neighbourhood well-being Avoid classifying poor neighbourhoods as a problem and middle-class as a solution 51
Sako Musterd Urban Geography /Urban Studies University of Amsterdam 52