Tenure Neighbourhoods in London: Policy Myth Or Effective Device to Alleviate Deprivation?

bs_bs_banner

Volume 37.2 March 2013 451–79 International Journal of Urban and Regional Research DOI:10.1111/j.1468-2427.2012.01145.x Mixed-Tenure Neighbourhoods in London: Policy Myth or Effective Device to Alleviate Deprivation?

SONIA ARBACI and IAN RAE

Abstractijur_1145 451..479 Social and tenure-mixing policies are recurrently deployed as a means to tackle urban deprivation and reduce social inequalities on both sides of the Atlantic, albeit without sound evidence, with policies often resulting in gentrification, social polarization or dispersal towards peripheral areas. This article explores whether housing- tenure diversification is a device for alleviating deprivation in terms of increasing socioeconomic opportunities (production) and access to resources (consumption), and empirically contributes to wider debates on inequality, segregation and neighbourhood effects. Quantitative and qualitative longitudinal analyses examine the extent to which in Greater London there are greater opportunities and access to resources amongst social tenants in mixed-tenure neighbourhoods than amongst those in concentrations of social housing. Changes in deprivation levels according to the Index of Multiple Deprivation (IMD) are contrasted and unpacked, and correlations between tenure, ethnicity, income, location, neighbourhood affordability and deprivation are investigated. The research demonstrates that social tenants do experience differing opportunities and access to resources, but these are not dependent on, or improved by, the level of tenure mix within the neighbourhood. Instead, integration of area-based and people-based policy, as well as decommodified access to welfare services (such as education, training and employment opportunities), is crucial. Furthermore, the IMD as an evidence base for policy formulation is contested and the neighbourhood effects are challenged.

Introduction Concentration of poverty often ‘reﬂects economic inequality, it does not cause it. Forcing neighbourhoods to be mixed in social and economic terms is treating the symptoms of inequality, not the cure’ (Cheshire, 2006: 1241). Notwithstanding this, social and tenure mixing in urban policy are recurrently deployed as a means to alleviate, or prevent, urban deprivation and reduce social inequality on both sides of the Atlantic (Atkinson and Kintrea, 2001; Galster, 2002). This article aims to contribute empirically to the wider theoretical and political debate on social mixing from a redistributive perspective, rather than a relational one (Van Beckhoven and van Kempen, 2003). Emphasis is placed on inequality of opportunities and access to resources within the remit of social mobility and neighbourhood (de-)commodiﬁcation (Musterd et al., 2003; Cheshire, 2009), rather than on interaction,

A version of this article was presented at the XVII ISA World Congress — RC21 in Gothenburg (2010). The authors are grateful to Tim Butler, Paul White, Nick Phelps, Yvonne Rydin and Pablo Mateos, as well as the IJURR reviewers, for their comments and valuable time. We also wish to thank the interviewees.

© 2012 Urban Research Publications Limited. Published by Blackwell Publishing. 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main St, Malden, MA 02148, USA 452 Sonia Arbaci and Ian Rae diversity, aspiration, social cohesion and social capital (Wilson, 1987; Fainstein, 2005; Uitermark et al., 2007). It progresses European debates relating to segregation (Musterd et al., 2006; Journal of Ethnic and Migration Studies, 2010), rather than gentrification (Lees, 2008; Urban Studies, 2008). The former are particularly equipped to question the context-free notion of neighbourhood effects and its instrumental contribution ‘in the broad debate on area-based policy’ (Lupton, 2003) since they also reposition the discussion, and thus the solution, at a structural level and within wider welfare debates where redistributive arrangements and embedded sources of inequalities can be captured (Musterd, 2005; Arbaci, 2007). Within the academic arena, mixed-communities discourse is widely regarded as a hegemonic neoliberal project that marks working-class neighbourhoods and areas with concentrations of social housing as problematic and, as a consequence, justifies marketing places for investment through diversification rhetoric (Hackworth, 2007; Lupton and Tunstall, 2008; Darcy, 2010). It is a ‘faith-based policy because there is scant real evidence that making communities more mixed makes the life chances of the poor any better’ (Cheshire, 2009: 343) and ‘increase[s] the social capital and social cohesion of inner city communities’ (Lees, 2008: 2450). Nonetheless, social mixing is a recurring theme in the political sphere, ranging from US initiatives such as HOPE VI and MTO to housing differentiation programmes in the Netherlands, or Contracts de Ville in France.1 In the UK, it remains at the forefront of spatial planning through mixed-tenure policy under the government’s Sustainable Communities agenda (CLG, 2003), which addresses perceived shortfalls in social cohesion, particularly within large concentrations of social housing (Robson et al., 2000). This approach, which considers clustering of minority ethnic groups in concentrations of social housing as problematic, has been further legitimized in the aftermath of the riots in North-England cities in 2001 and the London bombings of 2005, amid fears that Britain is ‘sleepwalking to segregation’ (see critiques in Finney and Simpson, 2009; Hamnett and Butler, 2010). In the UK, the assumption that diversification of housing tenure can assist in the prevention of a range of socioeconomic problems at the neighbourhood level (JRF, 2005) permeates across scales: in national guidance (CLG, 2006), regional strategies (for example, the London Plan and the Housing Market Renewal Pathfinder) and within local-authority-level policies (Local Development Documents). This assumption is reiterated by the Mixed Community Initiative of 2005, which — in contrast to previous programmes (the late-1990s New Deal for Communities or the 1960s Area-Based Initiative) — promotes an entrepreneurial and instrumental approach to social mixing, furthering the process of re-commodification of public services associated with the devolution process (Raco, 2008; Watt, 2009). Similarly, the continued London Plan designation of ‘Regeneration Areas’ to target (that is, to channel funding and initiatives to) the 20% most deprived wards according to the 2007 Index of Multiple Deprivation (IMD), assumes that deprivation is not as severe in the remaining 80% of wards, because it is not concentrated (Mayor of London, 2009).2 All these planning policies share an area-based approach and prioritize tenure diversification in areas with a concentration of deprivation, often associated with a concentration of social housing. Whilst resonating from the neighbourhood-effects assumption, ‘the discourse framing of poor neighbourhoods provides a framework in which the obvious solution is not to challenge structural inequalities, but to engage in spatial reordering’ and change individual behaviour (Lupton and Tunstall, 2008: 110).

1 For critical reviews, see, respectively, Housing Studies (1999), Simon (2002), Popkin et al. (2004) and Kling et al. (2007). HOPE VI stands for Homeownership and Opportunity for People Everywhere and MTO for Moving to Opportunity for Fair Housing. 2 The Index of Multiple Deprivation (IMD), published by the UK government in 2004 and 2007, identiﬁes the level of deprivation at a range of spatial scales and based on seven individual domains of deprivation: income, employment, health, education, barriers, crime and living environment. Each domain is formulated upon a wide range of indicators.

However, there is little evidence to suggest that such a spatial reordering actually tackles deprivation, particularly amongst social tenants (Friedrichs et al., 2005; Van Gent et al., 2009; Darcy, 2010). Conversely, rather than moving people across tenures, with the aim of increasing educational and labour market attainment, mixed-tenure policies have often driven or led to wider economic transformations of cities, social polarization and state- led gentrification (Allen, 2008; Porter and Shaw, 2009; Urban Studies, 2008). Nonetheless, the extent to which housing-tenure diversification, or mixed-tenure neighbourhood, is a device for alleviating deprivation needs to be tested, as does the question whether social-housing concentration, or concentration of deprivation, constrains residents’ opportunities and access to resources. This article addresses both issues within the London context and focuses on social tenants through comparative qualitative and quantitative analyses. It employs the IMD indicators utilized by the government as evidence for area-based funding initiatives in deprived areas. This article compares urban and suburban neighbourhoods in Greater London, starting with quantitative research in Part 1 to explore the extent to which social tenants in ten mixed-tenure neighbourhoods have greater life chances in terms of opportunities and access to resources than those in ten concentrations of social housing. Changes in deprivation levels (IMD, in 2004 and 2007) are contrasted and the seven individual domains that make up the indices are unpacked, by capturing additional indicators and investigating the correlations between tenure, ethnicity, income, location, neighbourhood affordability and deprivation in differing residential contexts. Part 2 focuses on qualitative analysis and further socioeconomic profiling of two predominantly social- rent neighbourhoods within Heath ward (Barking and Dagenham) and Peckham ward (Southwark) and two mixed-tenure neighbourhoods within Highbury East ward (Islington) and Woolwich Common ward (Greenwich), which appear to be significant in Part 1. A detailed analysis of these four paradigmatic cases reinforces the variance in the relationships between social tenants and the context of their neighbourhood — including ethnicity, housing progression, neighbourhood affordability, training and employment opportunities. It demonstrates that social tenants do experience varying socioeconomic opportunities (production) and access to resources (consumption), but these are not dependent on, or improved by, the level of tenure mix within the neighbourhood. Instead, integration of area-based and people-based policy, as well as decommodified access to the neighbourhood and welfare services, such as education, training and employment opportunities, is crucial. Furthermore, the IMD as an evidence base for policy formulation is contested and the neighbourhood effects are challenged. Thus, the success of mixed-tenure policy is disputed with reference to these paradigmatic cases, which indicate that mixing exacerbates inequality and masks deprivation, while showing that concentration of social housing facilitates opportunities and access to services, thus improving life chances.

Shifting discourses and approaches This article draws upon and explores two related threads of enquiry that are debated in segregation studies: sources of inequality and critiques of neighbourhood effects. First, the consideration of production of inequality and opportunities for social mobility expands recent efforts to redirect the debate and interpretation of ‘mixed communities’ and ‘concentration of poverty’ to redistributional realms, including (de-)commodiﬁcation of welfare services and neighbourhoods (Musterd et al., 2003; Beaumont, 2006; Cheshire, 2006; Musterd et al., 2006; Cheshire, 2009). This is of paramount importance and calls for a radical shift in both policy and theory. Over the past two decades, under the umbrella of neoliberal agendas and devolution processes, there has been a move from redistributive discourses to relational ones (Van Beckhoven and van Kempen, 2003; Lupton and Tunstall, 2008). ‘Structural problems

[have been] individualized and spatialized’ (Darcy, 2010: 5), and policies have thus moved away from being comprehensive and systemic (redistributive) to focusing on individual behaviour and problematization of poor neighbourhoods (relational). In other words, policy solutions have moved from arranging redistribution of wealth and public services across society to improving relations and interactions between individuals and within selected places. ‘Purposive aggregation of data [for example, IMD] serves to discursively construct places and communities as the source of and the solution to social and economic disadvantage, while binary narratives of community and token participation practices reinforce the perception that it is the poor tenant households who must change their behaviour or their residential location in order to secure the promised benefits’ (Darcy, 2010: 19). In academia, this movement has been mirrored by a similar shift from redistributional to relational perspectives, strongly influenced by USA schools of thought (Wilson, 1987; Briggs, 2005; and the ‘Dispersal Consensus’ — see Imbroscio, 2008). Since concentration of disadvantaged households is regarded as having negative effects, the most important delivery vehicles for social equity are sought in the perceived benefits of diversity (Fainstein, 2005), individual role models adhering to middle-class norms (Atkinson and Kintrea, 2001: 2278; Friedrichs and Blasius, 2003) and social capital through greater interaction and cohesion (Bagguley and Hussain, 2003; Uitermark et al., 2007). The individuality promoted in these relational approaches has warranted a renewed over-emphasis on spatial reordering interventions. These discursive strategies, which focus on individuals and places — rather than on structural arrangements — are historically bound to the devolution process. Given associated financial and fiscal constraints, devolved local governments have further enforced an entrepreneurial approach in the provision and management of public services and social housing (Van Kempen et al., 2005; Lipman, 2008). ‘Local authorities are encouraged to...create opportunities and conditions for private investment — for example, by clearing sites for development, or by selling or giving publicly owned land to developers, in return for contributions to development of public housing, infrastructure or facilities’ (Lupton and Tunstall, 2008: 107). Accordingly, diverse forms and degrees of re-commodification of welfare services and privatization schemes (for example, Right to Buy, Arm’s Length Management Organization), as well as diversification and deconcentration programmes in publicly owned land, have been further justified as the preferred solution (Power, 2007; Raco, 2008; Watt, 2009). The more policies have fostered private provision and welfare services re-commodification, the greater the relational discourse on concentration of deprivation and social housing, framed as a threat resulting from neighbourhood effects, lack of aspirations and lack of individual role models (Musterd and Andersson, 2005; Hackworth, 2007; Lipman, 2008). The policy instrument turns into tenure mixing and area-based programmes on both sides of the Atlantic, based on neighbourhood-effects assumptions. However, two major corpuses of critiques have emerged in transatlantic gentrification and segregation debates. Mixing and deconcentration programmes are contested as promoting polarization and state-led gentrification (Lees, 2008; Urban Studies 2008; Porter and Shaw, 2009; Watt, 2009), whereby the government engages with developers to provide regeneration cooperatively in order to stimulate a larger middle-class homeownership market for new-build developments (Davidson and Lees, 2005). Concurrently, the narrowing focus on individuals and neighbourhoods is essentially diverting attention away from the wider structural causes of poverty and production of inequality, which instead warrants further enquiry (Beaumont, 2006; Musterd et al., 2006; Arbaci, 2007; Cheshire, 2009). This calls for renewed interest in, for instance, the relationship between inequality and access to welfare services, or rather the (re-)commodification of welfare services and their impact at a local level. In the late 1990s, this relationship was sought within macro-scale perspectives, with particular emphasis on housing (Musterd and Ostendorf, 1998). Currently it is explored at a local level, with growing attention to education and training (Butler and Robson, 2003; Burgess

International Journal of Urban and Regional Research 37.2 © 2012 Urban Research Publications Limited. Mixed-tenure neighbourhoods in London 455 et al., 2005; Journal of Education Policy, 2008) and the notion of neighbourhood (re-)commodification (Cheshire, 2009), which considers the degree of neighbourhood affordability — a crucial component of local consumption of wealth and resources. A second thread of enquiry has recently emerged, given the weight of neighbourhood effects in constructing the discourses on mixed community, differentiation and deconcentration policies. It calls for greater empirical and conceptual scrutiny of neighbourhood effects from its discourse approach (see the comprehensive review and critique in Lupton and Tunstall, 2008 and Darcy, 2010) to the scales and dimensions of policy responses (see the multi-level model of neighbourhood effects and methodological debate in Friedrichs et al., 2005: 5–8; also see Lupton, 2003; Cheshire, 2006; 2009). There is a certain ‘circularity’ about the debates — whilst an aggregate of poor people define poor areas, researchers question whether there is an ‘additionality’ about such aggregates that exacerbates residents’ poverty (Atkinson and Kintrea, 2001: 2278). A recurring debate appears to be whether housing should be viewed fundamentally as an end in itself, or as a means to the greater end of reducing social stratification and deprivation (Galster, 2002). Most housing-differentiation policies are based on the assumption that there is a correlation between deprivation/opportunity structures and residential mix (Ostendorf et al., 2001). On the one hand, it is unclear whether mixed- tenure neighbourhoods offer better access to jobs and services, or better opportunities for mutual exchange of various goods and services (Kesteloot et al., 2006). Concentration might play a significant role in both (Kesteloot and Meert, 1999; Power, 2007). On the other hand, it is argued that (poor) neighbourhoods have limited negative effects on individual outcomes, since they do not bear any statistical significance on constraining upward mobility in terms of labour-market prospects, educational achievements and training opportunities (on the Netherlands, see Musterd et al., 2003; on the UK, Burgess et al., 2005; Bolster et al., 2007; Kling et al., 2007; on France, Goux and Maurin, 2007; and on the USA, Lipman, 2008). Darcy (2010: 3) states:

Reviews...intheUKand...intheUSfound inconsistencies in methods and approach and lack of agreement as to what was being measured [as neighbourhood effects]... Lupton (2003) question[s] the extent to which reliable evidence of neighbourhood effects is possible, or even necessary to support policies aimed at income mixing.

Nevertheless, some European scholars view the neighbourhood effects as a Trojan horse to refocus the debate on social mobility, by questioning whether the neighbourhood plays a greater role than the decommodification of welfare services in providing or constraining socioeconomic opportunities and access to resources (Musterd et al., 2003; 2006). In other words, they implicitly question whether area-based policies are more effective than redistributive comprehensive policies in decreasing inequalities and providing more equal opportunities for life chances (Van Gent et al., 2009). This article contributes to this enquiry. There is a clear requirement to set the neighbourhood in its wider context, thereby recognizing that economic forces may exaggerate neighbourhood problems and that public policies beyond the neighbourhood may have more influence on people’s lives than specific area-based initiatives (Atkinson and Kintrea, 2001). Other policy areas, such as labour and education, ultimately influence housing mobility as well as socioeconomic segregation (Musterd et al., 1998; Cheshire, 2006). Therefore, research should not limit its attention to the question of ‘whether’ neighbourhood matters, but begin to tackle the more difficult question of ‘how’ and ‘for whom’ (Ellen and Turner, 1997). The focus on socioeconomic opportunities and access to resources is not new. It follows the welfare debates of the mid- to late 1990s in segregation and urban studies (Musterd and Ostendorf, 1998) that have now revisited past theoretical sources, such as Polanyi’s theories of ‘modes of socioeconomic integration’ and ‘mixed embeddedness’ (Kesteloot and Meert, 1999; Kesteloot et al., 2006), which look at how particular

International Journal of Urban and Regional Research 37.2 © 2012 Urban Research Publications Limited. 456 Sonia Arbaci and Ian Rae conditions of market exchange, reciprocity and redistribution provide opportunities and resources for upward social mobility, particularly in concentrated areas. Overall, both threads of enquiry provide a degree of continuity in their counter- arguments, both redirecting the debate on ‘mixed communities’ and ‘concentration of poverty’ to redistributional realms. The attention to mechanisms and programmes redistributing wealth by providing equality of opportunities and access to resources in turn warrants further understanding of the relationship between production of inequality and re-commodification of welfare pillars (education, health, housing, and so on), given current neoliberal pressures and devolution processes. It is this uneasy relationship that is at the core of the European academic debate on social mixing (and neighbourhood effects), which culminates in the ‘social justice dilemma’ (Lupton and Tunstall, 2008) and the issue of ‘neighbourhood affordability’ (Cheshire, 2009), further contributing to the re-emerging debate on the ‘right to the city’ (City, 2009; Marcuse et al., 2009). In light of the above, this article responds to the need for tenure mixing to be empirically tested within the remit of equality of opportunity and neighbourhood (de-)commodification. It seeks to achieve this by narrowing the focus in order to explore the relationship between diversification of housing tenure and (decreasing) urban deprivation. Such exploratory work is crucial, since the perceived benefits of mixed- tenure neighbourhoods on their residents’ social and economic opportunities form the dominant assumption underlying mixed-community policies and deconcentration policies on both sides of the Atlantic. However, unlike other strands of policy, such an approach is not underpinned by sound evidence, and policymakers have little supporting research upon which to monitor and review such policies.

Methodology Greater London is a mosaic of mixed-tenure neighbourhoods, of socially and ethnically mixed areas, alongside pockets of wealth and poverty (see Figure 1) and spatial concentrations of social housing (see Figure 2). This occurs within both the inner city (inner London boroughs) and the suburbs (outer London boroughs), where concentration of owner-occupation may also be found (Hamnett, 2003; Hamnett and Butler, 2010). Such diversity makes Greater London a crucial case study, yet its fine-grained pattern is complex to unravel and calls for a neighbourhood-level analysis. Testing the relationship between tenure diversification and changes in deprivation requires a longitudinal perspective, which is problematic in nature since it entails several indicators at differing scales being systematically comparable over time (Friedrichs et al., 2005). The 2004 and 2007 Indices of Multiple Deprivation (IMD) (ODPM, 2004; CLG, 2008) are employed here, as they were the first (and latest) successive sets of indices to be set at the same neighbourhood scale that follow the formulation of three layers of Super Output Areas (SOAs) to facilitate further analysis of the 2001 census, in which output areas replaced enumeration districts for the first time (see note in Figure 7). The IMD is therefore the only multidimensional tool available to policymakers and researchers that allows longitudinal comparisons of the same key indicators for the same geographical units, which inform seven distinct domains (see footnote 2). A detailed definition and breakdown of indicators used to inform the IMD are set out in ‘The English Indices of Deprivation 2007’ (CLG, 2008). Moreover, the IMD has been the government’s favoured tool for measuring the comparative success of policy outcomes and identifying neighbourhoods for investment and renewal (for example, the New Deal for Communities; London Regeneration Areas), which consider tenure mixing as a key mechanism to achieve improvement (CLG, 2006; Mayor of London, 2009). However, in this article the IMD is critically deconstructed for the purpose of the research, given its inherent contradictions in the construction of people-based indicators to measure place-based characteristics, with an absence of tenure-neutral measures (Deas

International Journal of Urban and Regional Research 37.2 © 2012 Urban Research Publications Limited. Mixed-tenure neighbourhoods in London 457 et al., 2003; Darcy, 2010). On the one hand, changes in the degree and domain of deprivation are looked at and related to socioeconomic opportunities (production, i.e. employment and training) and access to resources (consumption, i.e. public services, public spaces and amenities, housing affordability), including neighbourhood affordability, as an important component of local consumption of wealth and resources (Cheshire, 2009). On the other hand, the IMD is only used as a starting point to construct the research and identify key case-study neighbourhoods across Greater London, it being the tool the government uses to highlight demographic and socioeconomic patterns at the neighbourhood level (albeit through individual/household data across seven domains). In terms of Lower Super Output Area (LSOA), each neighbourhood averages 1,500 residents and can be consistently selected across inner and outer London boroughs (see Figure 7). Case studies have been identified by isolating the neighbourhoods (at LSOA level) with the greatest tenure mix and those with the highest proportion of social-rent dwellings in the 2001 census. Optimal tenure mix was defined as neighbourhoods with more than 40% social-rented, more than 40% owner-occupied and more than 15% privately rented households. Proportions of at least 40% social-rented and 40% owner- occupied households were used because this provided a substantial portion of each tenure type in the neighbourhood without either being dominant. It was also crucial to recognize the importance of the private-rent sector, the specified 15% reflecting the fact that this is the smallest tenure type of the three throughout Greater London (Hamnett and Butler, 2010). Only ten neighbourhoods reflected these parameters of optimal tenure mix. These were then compared to the ten neighbourhoods with the highest proportion of social-rented dwellings (see Figure 3). In order to analyse whether social tenants in mixed-tenure neighbourhoods have greater opportunities and access to resources than those in concentrations of social housing, relationships between tenure, ethnicity and deprivation are examined first. Whilst it is important not to make a simple correlation between ethnic diversity and socioeconomic deprivation, people from ethnic minority groups often find themselves housed in the social-rented sector in certain parts of a city (Hamnett, 2003; Beaumont, 2006), a pattern that has been increasing in Greater London in the past two decades (Hamnett and Butler, 2010). Debates surrounding the creation of mixed communities have an implicitly strong ethnic component, which has been inherited from neighbourhood-effects theory based on the context of USA ghettos (Wilson, 1987), and more recently has been embedded in the UK as a result of perceived problems of ethnic segregation, whereby tenure mixing has been regarded as a policy solution (Robson et al., 2000; Cantle, 2002; Bagguley and Hussain, 2003). Simultaneously, most European countries have shifted from pluralist models and multicultural policies to new strategies seeking rapid-assimilation models, particularly within perceived segregated residential areas (Grillo, 2007; McGhee, 2008; Prime Minister’s Office, 2011). It is therefore important to consider the ethnic dimension in this research and to what extent the changeable opportunities and resources of different groups are influenced by neighbourhood context. Part 1 of the research interprets all 20 neighbourhoods in relation to the 2004 IMD, other indicators from 2001 to 2003, and changes in deprivation between 2004 and 2007 by unpacking each domain and, where necessary, capturing additional explanatory indicators. The period 2001 to 2007 was a period of economic stability, with consistent employment rates nationally and no economic recession, which reduces the impact of macro-scale externalities in the interpretation of changes. From Part 1, two mixed-tenure neighbourhoods and two predominantly social-rented neighbourhoods are then selected, based on the perceived significance of their diverging rates of deprivation and population composition, including their ethnic components and inner/outer location. The four neighbourhoods are shown to be valuable case studies because they mirror different contexts across the city, which are complementary in their characteristics. The case studies are therefore not completely unique and can be considered alongside other

International Journal of Urban and Regional Research 37.2 © 2012 Urban Research Publications Limited. 458 Sonia Arbaci and Ian Rae neighbourhoods with comparable characteristics for research or policy-formulation purposes. These four neighbourhoods are then explored in more detail in Part 2 using quantitative and qualitative considerations of their wider neighbourhood context. Since Part 1 is a desk-based high-level study, Part 2 required the development of an in-depth knowledge on the ground in order to mitigate against potential ﬂaws in the IMD and thus misinterpretations in Part 1. Semi-structured interviews with two managers of social housing were therefore conducted in each of the four selected neighbourhoods, in addition to interviews with a long-term social tenant with active involvement in the local community, thus allowing feedback from managers as well as general assumptions made about local context to be validated.3 This was combined with additional literature on the local-borough context to develop a more accurate picture. In this sense, the research described in Part 2 plays a vital role in identifying ﬂaws in the assumptions made in Part 1, while also demonstrating the limitations of the IMD and how it can often be an inaccurate measure of relative deprivation and policy success (Darcy, 2010). In light of the above, the research seeks to move away from the government’s dependence on the IMD as a policy-monitoring tool, given the weaknesses of the indices, which are often accused of ignoring neighbourhood processes (change of population, users of local services) and local context, as deprived places are simply viewed as the sum of deprived people within them (Deas et al., 2003; Darcy, 2010), which can often mask deprivation experienced by marginalized residents.

Part 1 — Setting the neighbourhoods: relationship between tenure, ethnicity and deprivation As outlined in the previous section, mixed-tenure neighbourhoods and socially and ethnically mixed areas in Greater London sit alongside pockets of wealth and poverty (see Figure 1) and concentrations of social housing (see Figure 2). Fine-grain and complex patterns exist across both inner and outer boroughs. Correlations between tenure, ethnicity and deprivation are examined at a neighbourhood scale (LSOA, each averaging 1,500 residents), which facilitates analysis of the question whether social tenants in mixed-tenure neighbourhoods have greater opportunities and access to resources than tenants in concentrations of social housing. The ten neighbourhoods with the highest proportion of social-rent dwellings (see Figure 3, right) and the ten neighbourhoods with the greatest tenure mix (see Figure 3, left) are isolated and contrasted. Of the ten neighbourhoods with the highest proportion of social-rent dwellings (see Figure 3, right), six are located in Southwark (inner London). These feature a signiﬁcant non-white population, whilst the neighbourhood with the largest white population (87%) is located in Barking and Dagenham (outer London). This suggests a direct relationship between minority ethnic dimension and social-rent stock in inner-London neighbourhoods (Hamnett and Butler, 2010). Of the ten neighbourhoods with the greatest tenure mix (see Figure 3, left), the majority are located in inner London. These feature an ethnically diverse population, except in the City of London, Camden and Islington, which have a signiﬁcant white

3 A total of 12 semi-structured interviews with key social-housing managers in selected neighbourhoods were undertaken during the summer of 2008. Two managers for those agencies shown in italics were interviewed, as well as one active long-term tenant for each neighbourhood: (1) Barking and Dagenham (10, Heath ward): London Borough of Barking and Dagenham; (2) Southwark (9, Peckham ward): Hyde Housing Association, Family Housing Association, Southwark and London Diocesan, New World Housing Association, Presentation Housing Association; (3) Islington (17, Highbury East ward): Family Mosaic Housing, London Borough of Islington; (4) Greenwich (19, Woolwich Common ward): London and Quadrant Housing Trust, Hyde Housing Association.

Figure 1 Distribution of deprivation, Greater London Authority 2007 IMD, by wards (source: CLG, 2008)

Figure 2 Distribution of social-rent stock, 2001, by wards (sources: 2001 census, as quoted in Mayor of London, 2008) population (78%), despite the presence of social-rent stock (above 40%) and a diverse population across the wider borough. Therefore, there is no obvious link between ethnicity and tenure, but there is a spatial relationship between tenure mix and proximity to inner London. As previously discussed, there is an assumption that concentration of social housing is correlated to concentration of poverty, and that such aggregation of deprivation constrains opportunities for life chances. This assumption can be examined by comparing changes in the Index of Multiple Deprivation (IMD) for 2004 to 2007 within

02UbnRsac ulctosLimited. Publications Research Urban 2012 © nentoa ora fUbnadRgoa eerh37.2 Research Regional and Urban of Journal International oe oiieﬁue niaea nraei ervto.Ec egbuho ubrcrepnst nLO oe(e iue7) Figure (see code 2008) LSOA CLG, ( an 2004; (%) to ODPM, population corresponds non-white number and neighbourhood (%) Each stock deprivation. in 3 increase Figure an indicate ﬁgures Positive Note: ie-euenihorod n oilrn egbuhoscmae:dge fdpiain(M cr 04ad20) oilhousing social 2007), and 2004 score (IMD deprivation of degree compared: neighbourhoods social-rent and neighbourhoods Mixed-tenure 15 5 35 55 3 60 4 45 0 0 0 0 IMD score Degree of Deprivation ( IMD scores, 0 - 100 ) Nw 20 MIXED TENURENEIGHBOURHOODS Lb 4 0 19 Nw 18 Br Is CL 14 17 16 4 sources 1 Gw 15 Cm 13 Ls 12 rp opldadcluae yato,bsdo aafo N,20:Tbe V9adUV63; and UV09 Tables 2001: ONS, from data on based author, by calculated and compiled graph : 4 2 Hc 11 43 8 BD 7 10 SOCIAL-RENT NEIGHBOURHOODS Ls Sw 6 9 Gw Sw 88 7 8 Sw 5 Sw 8 9 4 social-rent stock Sw Gw 3 2 9 0 Sw 1 (% oftotalhousingstock) neighbourhoods: changes inIMD % ofnon-whitepopulation Hc Ls Lb Br BD Nw Is Cm CL Gw Sw Dagenham OuterLondon InnerLondon ------Barking & Barnet City Greenwich Hackney Camden Southwark Newham Islington Lambeth Lewisham 20 20 20 0 -30% 30 -50% 50 -65% 65 -75% of 07 04 07 London : %

: 6 oi raiadInRae Ian and Arbaci Sonia 460 Mixed-tenure neighbourhoods in London 461 the 20 neighbourhoods. A reduction in IMD is assumed to suggest an increase in (equality of) opportunities and resources, since it should measure an improvement in the domain of employment, education, health, and so forth.

IMD in social-rent neighbourhoods When comparing the degree of deprivation and changes therein between 2004 and 2007 (Figure 3, right, and Table 1), the neighbourhoods with large concentrations of social- rent dwellings score a high IMD. These areas appear to be amongst the most deprived neighbourhoods, not only in London, but in the country. However, there has been no uniform pattern of decline between 2004 and 2007, thus indicating possible opportunities for improvement in social-housing neighbourhoods. In fact, a mixed picture is emerging. Many of the neighbourhoods deemed to have suffered from multiple deprivation in the 2004 indices have become progressively more socioeconomically marginalized in the 3 years that followed. For example, Lewisham (6) experienced a signiﬁcant increase in IMD by 18%. The neighbourhoods in Barking and Dagenham (10), Greenwich (2 and 8) and Southwark (3) all experienced an increase in scores for the overall IMD, with Barking and Dagenham (10) now the most deprived of all 20 neighbourhoods selected for this study (see Figure 4). These also have the largest white population of all 20 neighbourhoods (see Figure 7). This case warrants further analysis, especially as these are among the poorest parts of Greater London and have suffered as a result of the decline of their traditional manufacturing base (Goodwin, 2008). Conversely, ﬁve of the ten social-rent neighbourhoods actually experienced reductions in IMD scores (between 2004 and 2007). These are all located in Southwark, and in some cases (1 and 9) the IMD scores dropped by more than 10%, so that they became less deprived than several of the mixed-tenure neighbourhoods studied. Interestingly, these Southwark neighbourhoods have the largest non-white population of all 20 neighbourhoods. This suggests not only that there is no correlation between ethnicity and IMD changes, but also that other policy areas, regeneration programmes or recent urban processes (see Part 2) may have triggered the socioeconomic progress of these neighbourhoods, indicating improved opportunities for production (for example, employment) and access to resources, irrespective of tenure mix. This is consistent with several of the discourses supported by Ostendorf et al. (2001), Musterd (2002), Cheshire (2006) and Kesteloot et al. (2006).

IMD in mixed-tenure neighbourhoods Similarly, the ten mixed-tenure neighbourhoods show a varied pattern of change between the 2004 and 2007 IMD scores (see Figure 3, left, and Table 2). Six of the ten neighbourhoods experienced a reduction in deprivation scores, indicating potential improvement, in particular the City of London (16), Camden (13), Hackney (11) and Islington (17), which are all characterized by a large white population (see Figure 7), even though this is not the case for these boroughs as a whole. Simultaneously, four of the ten neighbourhoods showed an increase in deprivation score, the most severe case being Greenwich (15), one of the most ethnically and demographically diverse neighbourhoods, with a 20% increase. Greenwich not only had the sharpest increase in deprivation of all 20 neighbourhoods, but is also now rated more deprived than several neighbourhoods with large concentrations of social housing. This warrants further analysis in the remainder of the article. As already argued by van Beckhoven and van Kempen (2003), mixed-tenure neighbourhoods are not always successful and need to be interpreted within the speciﬁcities of their city, regional and national contexts (Musterd, 2005; Kesteloot et al., 2006).

International Journal of Urban and Regional Research 37.2 © 2012 Urban Research Publications Limited. 02UbnRsac ulctosLimited. Publications 37.2 Research Research Urban Regional 2012 and © Urban of Journal International Table 1 Social-rent neighbourhoods: 2004 and 2007 deprivation comparison Rae Ian and Arbaci Sonia 462

IMD Income Employment Health LSOA Score 2004 Score 2007 Change Change (%) Score 2004 Score 2007 Change Change (%) Score 2004 Score 2007 Change Change (%) Score 2004 Score 2007 Change Change (%) Southwark (1) 42.07 36.81 -5.26 -12.50 0.30 0.33 0.03 10.00 0.16 0.14 -0.02 -12.50 0.45 0.36 -0.09 -20.00 Greenwich (2) 55.88 56.77 0.89 1.59 0.50 0.50 0.00 0.00 0.20 0.20 0.00 0.00 0.70 0.71 0.01 1.43 Southwark (3) 42.7 44.01 1.31 3.07 0.33 0.39 0.06 18.18 0.15 0.19 0.04 26.67 0.54 0.60 0.06 11.11 Southwark (4) 47.68 43.54 -4.14 -8.68 0.35 0.32 -0.03 -8.57 0.19 0.16 -0.03 -15.79 0.62 0.56 -0.06 -9.68 Southwark (5) 37.22 35.96 -1.26 -3.39 0.31 0.34 0.03 9.68 0.17 0.15 -0.02 -11.76 0.28 0.18 -0.10 -35.71 Lewisham (6) 35.81 42.12 6.31 17.62 0.32 0.38 0.06 18.75 0.15 0.17 0.02 13.33 0.25 0.54 0.29 116.00 Southwark (7) 45.01 43.15 -1.86 -4.13 0.34 0.37 0.03 8.82 0.17 0.17 0.00 0.00 0.84 0.6 -0.24 -28.57 Greenwich (8) 49.18 53.9 4.72 9.60 0.44 0.47 0.03 6.82 0.19 0.21 0.02 10.53 0.83 1.06 0.23 27.71 Southwark (9) 38.71 32.57 -6.14 -15.86 0.33 0.30 -0.03 -9.09 0.13 0.11 -0.02 -15.38 0.51 0.32 -0.19 -37.25 Barking and Dagenham (10) 53.75 58.28 4.53 8.43 0.47 0.52 0.05 10.64 0.2 0.21 0.01 5.00 1.11 1.12 0.01 0.90

Education Barriers Crime Environment LSOA Score 2004 Score 2007 Change Change (%) Score 2004 Score 2007 Change Change (%) Score 2004 Score 2007 Change Change (%) Score 2004 Score 2007 Change Change (%) Southwark (1) 20.31 21.06 0.75 3.69 47.55 40.32 -7.23 -15.21 1.3 0.67 -0.63 -48.46 35.93 39.59 3.66 10.19 Greenwich (2) 37.97 36.84 -1.13 -2.98 43.6 49.3 5.7 13.07 1.35 1.03 -0.32 -23.70 37.2 46.64 9.44 25.38 Southwark (3) 23.65 22.31 -1.34 -5.67 48.43 43.53 -4.9 -10.12 0.81 0.41 -0.4 -49.38 41.77 35.74 -6.03 -14.44 Southwark (4) 23.83 27.82 3.99 16.74 46.36 41.24 -5.12 -11.04 1.49 1.61 0.12 8.05 37.03 41.38 4.35 11.75 Southwark (5) 22.02 18.78 -3.24 -14.71 47.44 44.02 -3.42 -7.21 -1.16 -0.39 0.77 66.38 39.28 41.91 2.63 6.70 Lewisham (6) 26.57 27.55 0.98 3.69 40.86 41.92 1.06 2.59 -0.05 0.46 0.51 1,020.00 33.66 36.7 3.04 9.03 Southwark (7) 23.12 30.66 7.54 32.61 50.93 46.19 -4.74 -9.31 0.53 0.33 -0.2 -37.74 43.02 38.78 -4.24 -9.86 Greenwich (8) 33.79 31.32 -2.47 -7.31 36.86 45.21 8.35 22.65 0.98 0.84 -0.14 -14.29 29.42 40.27 10.85 36.88 Southwark (9) 25.45 25.67 0.22 0.86 48.18 42.53 -5.65 -11.73 0.46 -0.04 -0.5 -108.70 32.47 35.54 3.07 9.45 Barking & Dagenham (10) 55.55 60.59 5.04 9.07 32.64 28.48 -4.16 -12.75 0.93 1.27 0.34 36.56 30.43 35.59 5.16 16.96 Sources: Calculated and compiled by author, based on data from ODPM (2004) and CLG (2008) 02UbnRsac ulctosLimited. Publications 37.2 Research Research Urban Regional 2012 and © Urban of Journal International Table 2 Mixed-tenure neighbourhoods: 2004 and 2007 deprivation comparison 463 London in neighbourhoods Mixed-tenure

IMD Income Employment Health LSOA Score 2004 Score 2007 Change Change (%) Score 2004 Score 2007 Change Change (%) Score 2004 Score 2007 Change Change (%) Score 2004 Score 2007 Change Change (%) Hackney (11) 50.67 45.62 -5.05 -9.97 0.39 0.35 -0.04 -10.26 0.21 0.19 -0.02 -9.52 1.01 1.00 -0.01 -0.99 Lewisham (12) 31.74 35.22 3.48 10.96 0.22 0.27 0.05 22.73 0.12 0.11 -0.01 -8.33 0.57 0.72 0.15 26.32 Camden (13) 36.92 32.20 -4.72 -12.78 0.22 0.26 0.04 18.18 0.13 0.13 0 0.00 1.17 0.87 -0.30 -25.64 Barnet (14) 32.26 34.04 1.78 5.52 0.3 0.29 -0.01 -3.33 0.16 0.16 0 0.00 0.29 0.3 0.01 3.45 Greenwich (15) 36.80 44.19 7.39 20.08 0.24 0.33 0.09 37.50 0.15 0.17 0.02 13.33 0.73 1.12 0.39 53.42 City of London (16) 17.39 14.38 -3.01 -17.31 0.09 0.09 0 0.00 0.08 0.08 0 0.00 -0.11 -0.31 -0.2 181.82 Islington (17) 40.23 34.41 -5.82 -14.47 0.18 0.17 -0.01 -5.56 0.14 0.13 -0.01 -7.14 0.83 0.71 -0.12 -14.46 Newham (18) 39.58 36.20 -3.38 -8.54 0.33 0.29 -0.04 -12.12 0.16 0.12 -0.04 -25.00 1.05 0.92 -0.13 -12.38 Lambeth (19) 29.20 28.09 -1.11 -3.80 0.21 0.20 -0.01 -4.76 0.10 0.09 -0.01 -10.00 0.50 0.49 -0.01 -2.00 Newham (20) 35.95 37.73 1.78 4.95 0.26 0.29 0.03 11.54 0.14 0.12 -0.02 -14.29 0.85 0.99 0.14 16.47

Education Barriers Crime Environment LSOA Score 2004 Score 2007 Change Change (%) Score 2004 Score 2007 Change Change (%) Score 2004 Score 2007 Change Change (%) Score 2004 Score 2007 Change Change (%) Hackney (11) 24.72 13.03 -11.69 -47.29 38.52 42.72 4.2 10.90 1.03 0.66 -0.37 -35.92 54.45 57.13 2.68 4.92 Lewisham (12) 20.28 18.43 -1.85 -9.12 39.60 39.98 0.38 0.96 0.84 0.86 0.02 2.38 26.27 45.38 19.11 72.74 Camden (13) 13.29 7.57 -5.72 -43.04 38.98 26.62 -12.36 -31.71 1.11 0.92 -0.19 -17.12 41.00 40.41 -0.59 -1.44 Barnet (14) 12.85 6.78 -6.07 -47.24 25.59 42.90 17.31 67.64 0.44 0.37 -0.07 -15.91 36.18 40.21 4.03 11.14 Greenwich (15) 27.41 25.73 -1.68 -6.13 35.23 40.27 5.04 14.31 1.11 1.25 0.14 12.61 32.37 30.37 -2.00 -6.18 City of London (16) 5.73 6.75 1.02 17.80 36.80 33.36 -3.44 -9.35 -2.15 -1.36 0.79 -36.74 42.24 28.31 -13.93 -32.98 Islington (17) 12.59 9.55 -3.04 -24.15 45.49 35.70 -9.79 -21.52 1.28 1.17 -0.11 -8.59 66.27 63.15 -3.12 -4.71 Newham (18) 21.57 18.79 -2.78 -12.89 35.95 48.33 12.38 34.44 0.68 0.53 -0.15 -22.06 27.92 27.10 -0.82 -2.94 Lambeth (19) 13.10 10.40 -2.70 -20.61 34.71 38.39 3.68 10.60 0.48 0.51 0.03 6.25 49.89 49.19 -0.70 -1.40 Newham (20) 27.03 15.55 -11.48 -42.47 39.13 50.09 10.96 28.01 -0.03 -0.14 -0.11 366.67 42.16 48.66 6.50 15.42 Sources: Calculated and compiled by author, based on data from ODPM (2004) and CLG (2008) 464 Sonia Arbaci and Ian Rae

Overall, both social-rent and mixed-tenure neighbourhoods do not show a correlation between deprivation and tenure in terms of either scores or changes. However, these findings are not representative for the whole of Greater London, as the further residualization of the social-rent stock has led to an increasing concentration of vulnerable groups within this tenure type (Hamnett, 2003; Watt, 2009). Within mixed- tenure neighbourhoods, ethnicity seems to be associated with both the extent and change of deprivation. However, these initial findings can be misleading unless it is known which of the seven domains affect the changes in deprivation and which processes (for example, gentrification, re-commodification of social housing) and conditions (for example, training, regeneration programmes) respectively constrain or enhance opportunities and access to resources. The former are explored in the remainder of this section by means of additional supporting information, whilst the latter are examined in Part 2.

Unpacking the seven domains of the IMD Income domain Deprivation scores (see Table 1) in the income domain are predictably high and stable for social-rent housing, since residents are disproportionally dependent on income support beneﬁts, and working and child tax credit (CLG, 2008). These households are typically already well embraced by the welfare state (Kesteloot et al., 2006), as the high number of Income Support claimants in social-rent housing shows (see Figure 4, right) compared to the number in mixed-tenure neighbourhoods (see Figure 6, left). In the latter

Income Support Income Support 300 300

MIXED-TENURE NEIGHBOURHOODS SOCIAL-RENT NEIGHBOURHOODS 250 250

200 200

average social-rent neighb. 150 150

average mixed-tenure neighb.

100 100

50 50

0 0 1 2 8 3 9 6 7 5 4 17 16 19 12 18 20 14 13 15 11 10 ark ich ich ark ark ark ark ark ham ich w net w w eth ag. w w w w w den r s ham ham ham b w kney D a s m c lington w w ondon B a a outh e e ewi am reen reen L outh outh outh outh outh Is C S H L ewi L N N reen S S S S S G G L G ity of arking & C B

Note: The vertical axis refers to the total number of claimants of income support in the selected neighbourhood. Income support includes social security beneﬁt for adults aged 18 or over who are working less than 16 hours a week, or have a partner working less than 24 hours a week and who have less money coming in than the law says they need to live on. In general, income support is now only available to people who are not required to be available for work, such as pensioners, lone parents and sick and disabled people. Figure 4 Mixed-tenure neighbourhoods and social-rent neighbourhoods compared: income support 2006 (source: graph compiled by author, based on data from ONS, 2008)

JSA Claimants JSA Claimants 120 120

MIXED-TENURE NEIGHBOURHOODS SOCIAL-RENT NEIGHBOURHOODS 100 100

80 80

60 60

40 40

20 20 Greenwich 15 Hackney 11 Greenwich 2 Greenwich 8 City of London 16 Barnet 14 Southwark 4 Southwark 5 Camden 13 Islington 17 Lewisham 6 Barking & Dag. 10 0 0 01 01 02 02 03 03 06 06 05 05 04 04 ' ' ' ' ' ' ' ' 20 20 20 20

Note: Job Seekers Allowance (JSA) is a beneﬁt for people of working age who are out of work or work less than 16 hours a week on average. As ﬁgures exclude the long-term economically inactive, they allow consideration of changing opportunities Figure 5 Mixed-tenure neighbourhoods and social-rent neighbourhoods compared: Job Seekers Allowance (JSA) claimants, August 2001 to 2006 (source: graph compiled by author, based on data from ONS, 2008)

neighbourhoods, the more afﬂuent owner-occupiers are likely to mask the presence of income support claimants. Not surprisingly, Barking and Dagenham (10), with their high concentrations of social housing and Income Support claimants, is the most multiply deprived neighbourhood in this study. Interestingly, despite their mixed-tenure composition, four neighbourhoods experienced an increase in income deprivation scores, Greenwich (15) having the greatest increase within the income domain for the neighbourhoods studied.

Employment domain This domain is crucial to shedding light on potential opportunities (production of wealth and resources), as it entails investigating whether large concentrations of social housing hamper access to employment. Employment is dependent on structural macro-economic factors (Cheshire, 2006) and indicators used to formulate this domain draw upon people- based policies, which arguably influence this domain more than area-based policies (Musterd et al., 1998). Interestingly, during the period of economic stability from 2001 to 2006, the change in unemployment benefit claimants (Job Seekers Allowance or JSA) in selected social-rent neighbourhoods showed no major increase and actually improved in many of the neighbourhoods studied (see Table 1 and Figure 5, right). This is particularly significant in Southwark (1, 4, 5 and 9), where residents managed to gain access to employment initiatives despite the lack of tenure mix. Therefore, concentrations of social housing do not always constrain access to employment. Not surprisingly, given the presence of highly skilled residents, mixed-tenure neighbourhoods experienced more stability in JSA claimants and a reduction in unemployment (see Figure 5, left), except Greenwich (15), a fact that again warrants further investigation.

Education, skills and training domain Of all seven domains, the education, skills and training scores illustrate the clearest contrasts between social-rent and mixed-tenure neighbourhoods (see Table 1 and Table 2). Amongst the social-rent neighbourhoods studied, the majority experienced an increase in deprivation scores in this domain, particularly Barking and Dagenham (10) which are now amongst the 5% most deprived neighbourhoods in the country for education, skills and training, with a score more than double that of fellow social-rent neighbourhoods. By contrast, all mixed-tenure neighbourhoods show significant improvement in education, skills and training deprivation scores, with the exception of the City of London (16). However, the indicator provides a false picture, since its spatial coverage does not necessarily relate to pupil performance based on attainment at local schools.4 Children living in a mixed-tenure neighbourhood do not necessarily attend local schools (Cheshire, 2006), which are also assumed to be mixed. In London, school pupils are generally more segregated at school than in their neighbourhoods (Burgess et al., 2005), such ‘micro-public’ spaces playing a key role in shaping patterns of division (Amin, 2002). These scores could thus be inflated by children of owner-occupied households attending private schools, or high-performing state schools outside the immediate neighbourhood (Butler and Hamnett, 2011), a practice that leads to higher school segregation for pupils residing in social housing. Therefore, population change and gentrification processes could have a significant impact on the performance of neighbourhoods within this domain.

Barriers to housing and services domain The barriers to housing and services domain shows an inverse pattern to the education domain. The majority of the social-rent neighbourhoods, including Barking and Dagenham (10), showed a reduction of scores in this domain, with Southwark neighbourhoods again displaying particularly good progress (see Table 1). In addition to distance to key services, scores in this domain are also based on ‘rate of acceptance under the homelessness provisions of the 1996 Housing Act’ and ‘difﬁculty of access to owner-occupation’ (CLG, 2008), which is more effectively managed in areas with large quantities of social housing stock. Thus, these positive changes can be misleading, as some of the indicators used to construct this domain are by default biased towards this tenure. By contrast, the mixed-tenure neighbourhoods are not able to offer the same levels of social and intermediate affordable housing (thus reducing opportunities for homelessness provision and access to owner-occupation). This has been exacerbated by the surge in house prices in London since the late 1990s (Hamnett, 2009), for example, in Islington.5 Indeed, the majority of mixed-tenure neighbourhoods showed an increase in deprivation scores for this domain (Table 2). Such mixed-tenure neighbourhoods are more likely to be targeted by (and indeed marketed to) high-earning professionals as potential up-and-coming places to buy property, especially given their close proximity to Central London, for example, Hackney and Newham. This might be the result of a relative lack of availability of social housing and difﬁculty in accessing owner- occupation, particularly for middle-income groups in inner-London boroughs. It is therefore debatable whether some of the mixed-tenure neighbourhoods allow residents

4 Indicators are also based on snapshots of test scores for different years, which were obviously obtained by different pupils, for example, one set of 14-year-olds for the 2004 data and another group of 14-year-olds for the 2007 data. 5 ‘Intermediate affordable housing’ refers to housing at intermediate prices (i.e. low-cost homes for sales) and intermediate rents (that is, above social rents but below market rents). It also includes equity sharing, such as shared ownership (for example, HomeBuy schemes), in which the resident buys part of the property and rents part of it from a housing association.

International Journal of Urban and Regional Research 37.2 © 2012 Urban Research Publications Limited. Mixed-tenure neighbourhoods in London 467 to progress in their ‘housing careers’, as local planning policies seek to achieve (Greenwich, 2006; Barking and Dagenham, 2010; Islington, 2011; Southwark, 2011). The living environment domain Although, on the whole, actual scores within the living environment domain are broadly similar in the social-rent and mixed-tenure neighbourhoods, there are clear differences in rates of change, with an increase in deprivation in most social-rent neighbourhoods and a decrease in most mixed-tenure neighbourhoods scores (see Tables 1 and 2). Scores consider the ‘outdoors’ living environment, including air quality and road traffic accidents, which are tenure neutral (CLG, 2008), yet are simultaneously influenced by the ‘indoors’ living environment, namely ‘social and private housing in poor condition’, which are not tenure neutral. The improvement in housing conditions in some mixed-tenure neighbourhoods is associated with the increasing gentrification process and gradual privatization of social-housing stock (Beaumont, 2006). Also, residents of social housing in mixed-tenure neighbourhoods may observe the high living standards of their home-owning neighbours and demand higher standards for their social-rent dwellings, whilst larger clusters of social housing are more difficult to manage and have become increasingly residualized and of a poorer quality as a result of long-term systemic underfunding of council housing (Van Kempen et al., 2005; Beaumont, 2006; Watt, 2009). Health deprivation and disability domain Variations between neighbourhoods in this domain can be seen to readily reflect ‘neighbourhood effects’, contributory indicators including ‘years of potential life lost’ and the ‘proportion of adults under 60 suffering from mood or anxiety disorders based on prescribing’. The mixed-tenure neighbourhoods showed far more stability within this domain, yet improvements were not as marked as in some of the social-rent neighbourhoods in Southwark (see Tables 1 and 2). Interestingly, the mixed-tenure neighbourhood within Greenwich (15) again experienced a dramatic increase in score in this domain and now matched that of social-rent neighbourhoods in Barking and Dagenham (10), which thus became the two most deprived neighbourhoods in this domain. Crime domain There does not appear to be any correlation between crime and tenure in the neighbourhoods studied (see Tables 1 and 2). This correlation has played a major role in American ghetto debates, whereas European research has been focused along income lines (Ellen and Turner, 1997; Musterd, 2005). Moreover, many of the mixed-tenure neighbourhoods located in inner London boroughs have higher scores in the crime domain than the outer boroughs, owing to their central location.

Preliminary ﬁndings Overall, it has been shown that the individual domains of deprivation have a complex and inconsistent relationship with residential tenure. Moreover, caution is required when considering ﬁndings, given the inherent limitations of indicators, and the inability to capture changes in population, users of local services (for example, schools) and local urban processes. In terms of the ‘income’, ‘employment’ and ‘health’ domains, many of the social-rent neighbourhoods became progressively more deprived between 2004 and 2007, Southwark neighbourhoods being the exception, which suggests improvement in employment opportunities and access to resources and services despite the lack of tenure mix. Amongst the mixed-tenure neighbourhoods, Greenwich (15) has surprisingly become more deprived, showing an inverse pattern to Southwark. Although both these neighbourhoods have a strong minority ethnic dimension, the patterns of change are

International Journal of Urban and Regional Research 37.2 © 2012 Urban Research Publications Limited. 468 Sonia Arbaci and Ian Rae diverging. The predominantly white neighbourhoods in Barking and Dagenham (10) and Islington (17) show a correlation in the degree of concentration of social housing and deprivation scores. However, the alleged relationship between ethnicity, tenure and deprivation is not straightforward and warrants further investigation. In terms of the ‘education, skills and training’, ‘barriers to housing and services’ and ‘living environment’ domains, most of the mixed-tenure neighbourhoods displayed a distorted picture, possibly as a result of a recent influx of middle-income residents, which has been driven by gentrification processes and middle-class school strategies, for instance in Islington (Cheshire, 2006; 2009). Social-rent neighbourhoods displayed patterns of decline in the ‘education’ domain, although providing greater rates of progress within the ‘barriers to housing and services’ domain. This was the case in, for instance, Barking and Dagenham (10) and Southwark (9), which offer social tenants a greater range of house sizes and types than mixed neighbourhoods. Again, in both cases further investigation of the apparent changes is required, given the limitation of some of the indicators. Overall, the white social housing neighbourhood in Barking and Dagenham (10) and the ethnically and tenure-mixed neighbourhood in Greenwich (15) emerge as being comparatively significant, given sharp increases in deprivation. Moreover, the predominantly black neighbourhood of Southwark (9) stands out as an important case study, since positive performance occurs within a concentration of social housing, whilst the predominantly white neighbourhood of Islington (17) with its ongoing gentrification processes highlights additional elements in the relationship between decreasing deprivation and mixed tenure. A more in-depth contextual analysis of these four case studies in Part 2 will enable us to explore socioeconomic opportunities (production) and access to resources (consumption) for social tenants in both types of neighbourhoods, thus allowing us to empirically contribute to the theoretical debates on neighbourhood effects and tenure mixing.

Part 2: Exploring opportunities and access to resources: a tale of four neighbourhoods Based on Part 1, we analyse four key case studies to investigate whether neighbourhoods with large concentrations of social stock provide a framework that limits socioeconomic opportunities and access to resources to social tenants in comparison to mixed-tenure neighbourhoods. These four case studies provide a range of neighbourhood contexts in terms of tenure, ethnic and socioeconomic composition, and geographical location within Greater London (see Figure 6 and Table 3). Of the social-rent neighbourhoods studied, white social housing (10) within Heath ward in Barking and Dagenham had an increasingly high multiple deprivation score and uniquely is almost entirely white. Conversely, black social housing (9) within Peckham ward in Southwark appeared to be progressing both socially and economically (IMD 2004 to 2007), despite its homogeneous tenure make-up. It is interesting to compare local circumstances to those in the selected Barking and Dagenham neighbourhood, which may also give an indication of reasons for progress seen in other Southwark neighbourhoods, which challenge neighbourhood-effects discourses. In terms of the mixed-tenure neighbourhoods it is necessary to further explore why the mixed ethnicity and tenure neighbourhood (15) within Woolwich Common in Greenwich has become significantly more deprived despite its tenure mix, whilst neighbourhoods such as the white leafy mixed neighbourhood (17) within Highbury East in Islington continues to become increasingly affluent. Both cases challenge the alleged benefits of mixed-tenure neighbourhoods. Findings from interviews (see footnote 3) and additional quantitative analysis will consider whether differences in opportunities and access to resources, including neighbourhood affordability, are specific to a particular context and population groups.

Social-rent neighbourhoods Mixed-tenure neighbourhoods

Southwark (9) Islington (17) Inner London - Peckham ward Inner London - Highbury East ward Black population (50%) White population (80%)

<< – crime –108.70% << – education –24.15% IMD IMD –15.86% – health –37.25% –14.47% – barriers –21.52%

–employment –15.38% – health –14.46%

(IMD 2004-07 change) – barriers –11.73% – crime –8.59% Decreasing deprivation Decreasing deprivation

– income –9.09% – employment –7.14%

– income –5.56%

Barking and Dagenham (10) Greenwich (19) Outer London - Heath ward Outer London - Woolwich Common ward White population (90%) Ethnically mixed population (50%)

>> + crime 36.56% >> + health 53.42% IMD IMD 8.43% + environment 16.96% 20.08% + income 37.50%

+ income 10.64% + barriers 14.31% (IMD 2004-07 change) Increasing deprivation + education 9.07% + employment 13.33%

+ employment 5.00% + crime 12.61%

Figure 6 Selected neighbourhoods, geography and composition (sources: calculated and compiled by author, based on data from ODPM, 2004; CLG, 2008)

This will also shed light on the extent to which housing tenure diversiﬁcation is a device for alleviating deprivation, as suggested by the tenure-mix policy approach in the UK. Further reﬂection on the relevance and robustness of the IMD as evidence for area-based policies will also be considered. Each of the four case studies will now be examined in turn.

White social housing in Barking and Dagenham (10) This neighbourhood is a typical development of the 1960s, with high-rise social housing and associated public infrastructure (school, civic centre, leisure centre) still owned and managed by the local authority. Low turnover rates could indicate an element of stability amongst residents or an element of entrapment. However, interviews suggest that residents view the area as their permanent home and do not see the neighbourhood as a ‘stepping stone’to more desirable accommodation, thus strengthening local cultural ties (Goodwin, 2008). Relatively weak public transport connections and spatial isolation have limited the influx of new populations. This stability has enabled shops and services to remain affordable and accessible to the local population, although many now appear outdated. High deprivation scores in the neighbourhood are a reflection of the residents’ low skills base, for example, there remains a dependence on the declining manufacturing sector for employment, which accounts for 11% of working-age residents’ jobs, double that of the other neighbourhoods studied (see Table 3). Over 40% of residents have no qualifications and 25% of households depend on income support (for example, lone parents) — by far the highest proportions of all the neighbourhoods studied (Table 3).

Table 3 Selected neighbourhoods: characteristics by tenure, ethnicity and employment (% 2001), and IMD (2004–07)

Barking & Data Dagenham (10) Southwark (9) Islington (17) Greenwich (19) (1) Ward Heath Peckham Highbury East Woolwich Common Outer London Inner London Inner London Outer London (2) Social housing % 87.40 87.57 40.94 41.11 (3) Ethnic group % White 86.75 26.64 78.69 47.93 Black 8.25 55.82 11.59 17.54 Indian 0.72 0.58 2.00 22.72 Pakistani 0.18 0.32 0.34 3.04 Bangladeshi 0.78 0.58 0.97 0.83 (4) Employment % Manufacturing 10.87 5.26 8.42 10.20 Construction 7.39 2.74 3.37 8.70 Finance 5.00 3.16 8.16 4.01 Retail 17.83 14.95 7.64 13.71 Education 14.13 7.37 7.64 8.19 Health & social work 8.91 15.16 7.38 12.88 Business activities 11.09 20.21 29.27 13.55 (5) IMD score 2004 53.75 38.71 40.23 36.80 2007 58.28 32.57 34.41 44.19 (5) IMD English rank* 2004 1,648 5,011 4,525 5,619 2007 1,082 7,149 6,504 3,547 *Where 32,482 is best and 1 is worst Sources: Table compiled by author, based on data from (1) LSAO code in Figure 7; (2) ONS (2001: Table UV63); (3) ONS (2001: Table UV09); (4) ONS (2001: Table KS11A, % people aged 16–74 in employment/working in); and (5) ODPM (2004); CLG (2008)

Despite these shortfalls in skills, education and income, and a lack of employment opportunities, there have not been any major employment or training initiatives, nor any investment in new infrastructure in the area. The close proximity of two relatively highly regarded secondary schools adds to the stability of the neighbourhood and may lead to educational improvements in the coming years, although this may deﬂect attention away from the need for increased public intervention in the area (Barking and Dagenham, 2010). Publicized white working-class dissatisfaction and increased support for the far-right British National Party (BNP) in this part of Outer East London results from the area’s relative economic isolation and perceived lack of, or unfair allocation of, resources (for example, health services, education, social housing and jobs) (Amin, 2002; Goodwin, 2008). Whilst the greatest part of Barking and Dagenham has been witnessing the largest increase in ethnic minorities in London (Hamnett and Butler, 2010), this neighbourhood does stand out as relatively isolated and socially homogeneous, with social housing often passed down from generation to generation within the same family.

International Journal of Urban and Regional Research 37.2 © 2012 Urban Research Publications Limited. Mixed-tenure neighbourhoods in London 471 short London Number code Population Borough Neighb. LSOA* Population Non-white (%)

Southwark 1 018C 1563 61.93 Greenwich 2 024C 1538 39.40 Southwark 3 019B 1446 73.86 Southwark 4 022A 1401 60.10 Southwark 5 015C 1538 60.53 Lewisham 6 001A 1516 68.14 Southwark 7 019E 1336 71.26 Greenwich 8 005D 1494 46.32 Southwark 9 019D 1539 73.36 Barking and 10 006C 1661 13.25 Social-rent neighbourhoods Dagenham

Hackney 11 018C 1524 47.18 Lewisham 12 024C 1494 55.42 Camden 13 019B 1569 20.01 Barnet 14 022A 1546 30.01 Greenwich 15 015C 1448 52.07 City of 16 001A 1535 12.96 London Islington 17 019E 1450 21.31 Newham 18 005D 1566 63.73 Lambeth 19 019D 1458 27.71 Mixed-tenure neighbourhoods Newham 20 006C 1584 50.95

Note: *These are the LSOA codes for Super Output Areas, which are a set of geographies developed after the 2001 census. The aim was to produce a set of areas of consistent size, whose boundaries would not change (unlike electoral wards). They are an aggregation of adjacent Output Areas with similar social characteristics. The Ofﬁce for National Statistics (ONS) coding system is a hierarchical code used in the UK for tabulating census and other statistical data Figure 7 Super Output Areas, neighbourhood numbers by population, 2001 (source: ONS, 2001: Tables UV09 and UV63)

Interviewees acknowledge the lack of training and employment interventions, suggesting that the design and management of social housing is key to improving residents’ quality of life, and that tenure mix is of no relevance to increased opportunity or access to resources. However, overall it can be seen that whilst access to, and affordability of, resources have remained stable at a level set in the 1960s (for example, schools, shops and services), socioeconomic opportunities have continued to decrease (for example, training and employment). ‘Focus groups similarly found that a shortage of available housing combined with rapid demographic changes have fuelled the perception amongst residents that some areas of the borough are being “taken over” by minority ethnic groups’ (Cruddas et al., 2005: 15). Local resources, especially housing, were perceived to be unfairly distributed in favour of newly arrived groups at the expense of long-term residents (Goodwin, 2008). In the case of this neighbourhood, the relative isolation has been key to historical availability of council housing and retention of neighbourhood affordability.

This is a representative case for the lack of public intervention, given the plentiful supply of public services nearby, although these are outdated and were originally provided alongside the 1960s council housing. The area’s associated limited residential mobility and its cultural legacy, gained through its historical working-class identity, has also triggered feelings of resentment and entrapment (John et al., 2006; Goodwin, 2008).

Black social housing in Southwark (9) In contrast to Barking and Dagenham, this neighbourhood has a predominantly non- white population (74%), with more than 50% of all residents ‘Black or Black British’, half of which are first-generation immigrants born in Africa (see Table 3). However, despite the high volume of social housing and the relatively significant enclave of black African residents, there were significant reductions in deprivation scores between 2004 and 2007, which suggests potential increases in socioeconomic opportunities (for example, through training programmes) and access to resources (for example, affordable consumption and services). In fact, interviewees attribute these improvements to a substantial re-housing programme delivered by the ‘Peckham Partnership’ (the local strategic partnership), which is considered to have made a huge impact on the quality of life of the residents, with far less antisocial behaviour and more contentment amongst residents, as well as improved standards of housing and public spaces. Completed in 2004, the partnership involved the demolition of the neighbourhood’s old 1960s high-rise local-authority stock, which was replaced by new well-designed low-rise flats and houses, now managed by housing associations. The resident population within social housing has remained largely the same throughout the process, with minimal displacement. In contrast to the previous case, a constructive support network amongst residents facilitated the process, which has successfully increased access to resources and opportunities, in particular to skills and employment (Buck et al., 2002). In fact, this programme was complemented by many employment and training initiatives within the neighbourhood and across the wider borough, mainly coordinated through the Peckham Partnership, which was set up in 1996. For example, ‘Youth Plus’ is a local initiative to encourage youngsters to enter vocational training, which operates alongside a variety of initiatives run by housing associations for their residents. The neighbourhood is also located in close proximity to the new ‘Peckham Pulse’ Healthy Living Centre. Such initiatives seem to be reflected in the employment profile of this neighbourhood, with the dominant employment sector being ‘real estate, renting and business activities’, catering for 20% of the working-age population (see Table 3). As part of this programme, a small element of market housing has also been developed recently, which has both helped to fund the programme and introduce more of a tenure mix in the area (Southwark, 2011). Although the large quantity and variety of social housing in the neighbourhood offers residents more choice, allowing a high degree of internal movement, interviewees also argue that progression out of social-rent housing to intermediate housing is almost impossible, with only a couple of examples in the past 10 years (see footnote 5). Therefore, despite residents’ satisfaction with the design-led renewal and improved living standards, there are extremely limited opportunities for progression across tenures (Watt, 2009). Furthermore, it is debatable whether owner- occupiers in market housing will trigger an even greater rate of socioeconomic opportunities amongst social tenants, owing to the success of employment and training initiatives implemented in the area. Affordability of the neighbourhood’s shops and services has so far been retained; however, an increased influx of middle-class owner- occupiers could lead to gentrification processes, as in Islington, with consequences for neighbourhood affordability. Southwark is a paradigmatic case; because of a lack of tenure mix, training and employment initiatives have been provided, thereby increasing socioeconomic opportunities, as suggested in Part 1. Concentration rather than tenure mix has, in

International Journal of Urban and Regional Research 37.2 © 2012 Urban Research Publications Limited. Mixed-tenure neighbourhoods in London 473 fact, made public intervention possible that focused only on local residents, without attracting new residents, while maintaining neighbourhood affordability. In particular, concentration and ethnicity led to a strong support network (Kesteloot et al., 2006). The reduction in the IMD score can thus be representative of an increase in life chances in cases of limited population and tenure change. Housing progression within the neighbourhood has been reached in terms of availability and quality but not in terms of tenure, as a result of the high cost of market housing in London (Hamnett, 2009). Overall, Southwark demonstrates that mixed tenure is not a prerequisite or a tool for increasing opportunities and local resources.

White leafy mixed neighbourhood in Islington (17) We now move on to the mixed-tenure cases. In contrast to the other neighbourhoods we examined, in this part of Islington the mix of housing tenure is not accompanied by a significant mix of ethnicities (ONS, 2001). This predominantly white neighbourhood has become an increasingly affluent place over the past two decades and is regarded by professionals and middle-class families as a desirable permanent residence (see Table 3). Local-authority-owned and managed blocks of social housing, such as the Chestnuts Estate, are located alongside rows of three-storey houses that have been converted into expensive owner-occupied apartments, interspersed with further social dwellings managed by a housing association, although a large proportion of social housing has been sold to very wealthy professionals (Islington, 2011). The influx of wealthy owner-occupiers accounts for a reduction in deprivation scores, masking the real levels of socioeconomic deprivation social tenants experience, which are higher than in Southwark (9). In addition, the ongoing gentrification of the neighbourhood has triggered changes in the provision of retail and other services (expensive organic shops, wine bars), thus jeopardizing neighbourhood affordability for long-term residents (Cheshire, 2009). However, interviewees suggest that the mixed nature of the neighbourhood has some positive impacts on the tenants of social housing, including an increase in expectations in the standards of their own accommodation. As a result of new owner-occupiers’ demands and lobbying for higher standards of amenities, the exteriors of social properties and public spaces (for example, the Chestnuts Estate) are maintained to a higher standard than in some of the larger social-housing concentrations in nearby areas, so that they do not give an immediate impression of being less affluent. However, as widely argued in literature on gentrification, interviewees also warn about the negative impacts of mixed tenure on residents of social housing. Rich owner-occupiers are attracted by the growing popularity of Islington, which has been heightened by new high-profile developments such as the nearby ‘Arsenal’project (opened in 2009) and close proximity to central London (Islington, 2011). In effect, owner-occupiers are buying into a lifestyle and do not necessarily want to live in close proximity to social tenants. For instance, interviewees emphasize that owner-occupiers often seek to purchase an existing social-rent flat on the floor above or below in order to gain an element of exclusivity within two- or three-story town houses that have been subdivided into flats.Affluent professionals are replacing previous populations of Greeks and then Somalians, with social tenants moving further out of London to places such as Enfield and Cheshunt in Hertfordshire. The remaining social tenants are increasingly trapped and restricted, as they cannot afford or gain access to any other accommodation within Islington. Consequently there is a lack of housing choice, and even shared-ownership schemes are taken up almost exclusively by young professionals from outside the neighbourhood, thus limiting progression of housing careers (Islington, 2011). In contrast to the other three case studies, and as already stressed by Butler and Robson (2003) and Burgess et al. (2005), interviewees confirmed that there is a limited mix of pupils at school as a result of the education strategies of middle-class owner- occupiers (for example, by acquiring a home within the catchment area of the best

International Journal of Urban and Regional Research 37.2 © 2012 Urban Research Publications Limited. 474 Sonia Arbaci and Ian Rae schools, or by sending children to private schools). This leads to school segregation, which constrains socioeconomic opportunities for social tenants. Despite this, there are few policy initiatives to address educational and training needs, since the residential presence of the middle classes masks the actual level of deprivation present in the neighbourhood. Islington is a representative case of a mixed neighbourhood that does not provide greater opportunities and resources for social tenants but instead increases inequality, enhances school segregation and constrains housing careers within the neighbourhood, resulting in entrapment or outflow to peripheral outer-London neighbourhoods. Mixing between private and social housing residents appears to be one of distant observation rather than shared investment of social capital. It is a clear case of a reduction in IMD not only being a false indicator of decreasing deprivation, but also being used as a dangerous evidence base for policy, as it masks the need for public funds for infrastructure provision and removes attention from vulnerable population groups. It exemplifies the distorted picture outlined in Part 1. Paradoxically, it also shows how a decreasing IMD could indicate different stages of gentrification as a result of middle-income inward mobility.

Mixed population and tenure in Greenwich (15) This traditionally low-rise working-class neighbourhood shows the greatest diversity in terms of ethnicity (see Table 3), housing typologies and tenures, yet is experiencing increasing levels of deprivation. Residents have a lower skills base than those in all other mixed neighbourhoods, and in the Southwark social-rent neighbourhoods (Buck et al., 2002). Furthermore, figures for manufacturing employment and the Job Seekers Allowance (ONS, 2006) are far higher than for the other mixed-tenure neighbourhoods, with manufacturing employment similar to that in Barking and Dagenham (10). According to interviewees, this diversity is a result of both the nature of the private stock, which favours a transient population, and housing association allocation policy. As a result of the low quality of the private stock, the neighbourhood is often a stepping stone for both owner-occupiers and private renters. This accounts for instance for the area’s marked student population, which leads to a very high housing turnover. Simultaneously, housing associations address social problems by dispersing certain groups of residents away from areas of tension and into close proximity to market housing, which is believed to ‘influence lifestyle aspirations’ of social tenants. Such an approach is reflected in the local authority’s planning policies, which seek to promote a mix of housing tenures and types (Greenwich, 2006).This policy is particularly prevalent in this neighbourhood, because its street layout and its low-rise buildings are thought to foster greater opportunities for integration and interaction between residents of different tenures. Nonetheless, tenants’ mobility within and outside the neighbourhood’s social stock has been limited, as in other mixed-tenure neighbourhoods across Greater London (Watt, 2009). In contrast to the gentrified white mixed Islington neighbourhood, affordability of Greenwich’s shops and services has been retained. Interestingly, the mixed population does not mask neighbourhood-wide deprivation; however, there are no interventions or lobbying, such as those in Southwark and Islington, respectively, as a result of the transient nature of the population. The local ‘Neighbourhood Investment Team’ has carried out some initiatives in the borough, but it is apparent that there is still significant progress to be made in this particular neighbourhood (Watt, 2009). Greenwich is a paradigmatic case of an area in which the presence of a transient population entails a lack of lobbying and social investment in the area. In contrast to predominantly white Islington, this ethnically mixed neighbourhood has retained its relatively affordable cost of living and services, owing to its lack of gentrification. Nonetheless, in both cases the tenure mixing, rather than the high concentration of social housing, is limiting the attraction of public funding and hampering improvements to local services.

Conclusions This article seeks to contribute empirically to the wider theoretical and political debate that considers diversification of housing tenure as a device for alleviating deprivation. The diversity and complexity of this alleged relationship has been explored through a wide selection of London case studies, varying in context between tenure mix, deprivation, ethnicity, income and location. In fact, individual domains of deprivation show a complex and inconsistent relationship with housing tenure, with four paradigmatic cases displaying diverging patterns of change. They demonstrate that social tenants do experience differing socioeconomic opportunities and access to resources, but that these are not dependent on, or improved by, the level of tenure mix within the neighbourhood. Further research will be required to better explore the relationships between these factors. In particular, Southwark with its black enclave is a paradigmatic example of increasing opportunities in an area of concentrated social housing, facilitated primarily by people-based policies (training and employment) with retention of neighbourhood affordability, the absence of which characterizes the gentrification of mixed-tenure Islington. Here the decreasing IMD, driven by an influx of the white middle classes and their associated school strategy, masks the actual deprivation and entrapment of social tenants. Similarly, the lack of people-based policy in Barking and Dagenham drives entrapment and socioeconomic isolation, which in this case is exacerbated by geographical isolation (outer London), concentration of social housing and white working-class dissatisfaction. Paradoxically, the relationship between gentrification and ethnicity is complicated further by the outward movement of ethnic minorities from gentrified neighbourhoods, such as Islington, to outer-London locations to seek more affordability and housing choice, at the same time creating tension and competition for resources. Surprisingly, Greenwich, the most genuinely mixed neighbourhood in social, ethnic and tenure terms, provides the most complex and unique findings of the neighbourhoods studied. The mix results in a transient population, which in turn is characterized by a lack of lobbying and social investment in the area. Far from being a catalyst for investment in education, training and employment, mixed tenure — both gentrified and transient — can hamper the implementation of such programmes when founded upon area-based policy. This comparative analysis has led to three essential conclusions that can inform policy. First, the IMD as a policy tool and measure of outcomes is misleading, given the inherent contradiction in a people-based indicator being used to calculate place-based measures of deprivation. In fact, the indicator can only measure a neighbourhood’s relative affluence, and not levels or changes of socioeconomic opportunities and access to resources for residents within. A decreasing IMD is not a reflection of socioeconomic improvement or increase in life chances if interpreted in isolation from in- and out-migration, social/urban processes (such as gentrification, professionalization), consolidation of ethnic neighbourhoods and clustering strategies to safeguard middle- class areas. Paradoxically, changes in IMD have usefully identified areas that are experiencing differing stages of gentrification. Secondly, all four cases dispute the alleged benefits of mixing that underpin government policy and empirically contribute to two main threads of theoretical debate. Greenwich and Southwark are paradigmatic in challenging the neighbourhood-effects theory. Dilution of poverty through housing-tenure diversification diverts attention from the causes of inequality that are rooted within the wider welfare system. Instead, concentration of poverty (in social housing) and associated strong support networks can facilitate resources, opportunities and delivery mechanisms. Islington adds to the current debate on the far-reaching impact of gentrification across the city and society, which should warrant a serious policy response, particularly in relation to neighbourhood commodification and school segregation, which is driven by the residualization of public services and middle-class strategy.

Finally, the analysis shows that mixed-tenure policy cannot be used uniformly across Greater London, let alone the country, and that each neighbourhood should be viewed in its own context, with people-based policy applied regardless of concentration or dispersal of poverty, concentration of social housing, and location in inner or outer London. This offers support for a combination between an area-based and a comprehensive policy approach (decommodification of key services targeting people rather than areas), where the former is effective in engaging people and the latter in widening access to and knowledge of public services. Overall, all four cases show that services and mechanisms that retain neighbourhood affordability are fundamental in fostering socioeconomic opportunity (production) and access to resources (consumption). The degree of decommodification of services, resources and neighbourhood plays a vital role in reducing deprivation. If we are serious about reducing inequalities, or rather tackling the sources of inequality, then solutions cannot be sought in spatial reordering, or in the dilution of poverty through tenure diversification, but in whether and how (local) social welfare can be reconsidered in light of a more equitable society.

Sonia Arbaci ([email protected]), Bartlett School of Planning, University College London, 22 Gordon Street, London WC1H 0QB, UK and Ian Rae ([email protected]), London Borough of Redbridge, Town Hall, 128–142 High Road, Ilford, London IG1 1DD, UK.

References Allen, C. (2008) Housing market renewal and Bolster, A., S. Burgess, R. Johnston, K. social class. Routledge, London. Jones, C. Propper and R. Starker (2007) Amin, A. (2002) Ethnicity and the Neighborhoods, households and income multicultural city. Environment and dynamics: a semi-parametric investigation Planning A 34.6, 959–80. of neighborhood effects. Journal of Arbaci, S. (2007) Ethnic segregation, housing Economic Geography 7.1, 1–38. systems and welfare regimes in Europe. Briggs, X. (2005) The geography of European Journal of Housing Policy 7.4, opportunity. Brookings Institution Press, 401–33. Washington DC. Atkinson, R. and K. Kintrea (2001) Buck, N., I. Gordon, P. Hall, M. Harloe and Disentangling area effects: evidence from M. Kleinman (2002) Working capital: life deprived and non-deprived and labour in contemporary London. neighbourhoods. Urban Studies 38.12, Routledge, London. 2277–98. Burgess, S., D. Wilson and R. Lupton (2005) Bagguley, P. and Y. Hussain (2003) Conflict Parallel lives? Ethnic segregation in and cohesion: construction of ‘community’ schools and neighbourhoods. Urban and the 2001 ‘riots’. Paper Communities Studies 42.7, 1027–56. Conference, 18 September. Butler, T. and C. Hamnett (2011) Location, Barking and Dagenham, London Borough of education: place, choice, constraint in (2010) Local development framework core London. Children’s Geographies 9.1, strategy. [WWW document]. URL http:// 35–48. www.lbbd.gov.uk/8-leisure-envir/planning/ Butler, T. and B. Robson (2003) Plotting the local-dev-framework/plan-ldf-core- middle classes: gentrification and circuits strategy.html (accessed 15 April of education in London. Housing Studies 2011). 18.1, 5–28. Beaumont, J. (2006) London: deprivation, Cantle, T. (2002) Community cohesion: a social isolation and regeneration. In S. report of the independent review team. Musterd, A. Murie and C. Kesteloot (eds.), Home Office, London. Neighbourhoods of poverty: urban social Cheshire, P. (2006) Resurgent cities, urban exclusion and integration in Europe, myths and policy hubris: what we need to Palgrave, Basingstoke. know. Urban Studies 43.8, 1231–46.

Cheshire, P. (2009) Policies for mixed Goodwin, M.J. (2008) Backlash in the ‘hood: communities: faith-based displacement determinants of support for the British activity? International Regional Science National Party (BNP) at the local level. Review 32.3, 343–75. Journal of Contemporary European City (2009) Special issue: Cities for people, Studies 16.3, 347–61. not for proﬁt. City 13.2–3, 173–377. Goux, D. and E. Maurin (2007) Close CLG (2003) Sustainable communities: neighbors matter: neighborhood effects on building for the future. Communities and early performance at school. Economic Local Government, London. Journal 117, 1193–215. CLG (2006) Planning policy statement 3: Greenwich, London Borough of (2006) housing. Communities and Local Unitary development plan. [WWW Government, London. document]. URL http://www.greenwich. CLG (2008) The English indices of gov.uk/Greenwich/Strategies/ deprivation 2007. Communities and Local LandAndBuildingDevelopmentPlan/ Government, London. UnitaryDevelopmentPlan/ReadUPD/ Cruddas, J., P. John, H. Margretts, N. Lowles (accessed 15 April 2011). and D. Rowland (2005) The far right in Grillo, R. (2007) An excess of alterity? London: a challenge to local democracy? Debating difference in a multicultural Joseph Rowntree Charitable Foundation, society. Ethnic and Racial Studies 30.6, Coventry. 979–98. Darcy, M. (2010) De-concentration of Hackworth, J. (2007) The neoliberal city: disadvantage and mixed income housing: a governance, ideology, and development in critical discourse approach. Housing, American urbanism. Cornell University Theory and Society 27.1, 1–22. Press, Ithaca, NY. Davidson, M. and L. Lees (2005) New-build Hamnett, C. (2003) Unequal city: London in ‘gentriﬁcation’ and London’s riverside the global arena. Routledge, New York, renaissance. Environment and Planning A NY. 37.7, 1165–90. Hamnett, C. (2009) Spatially displaced Deas, I., B. Robson, C. Wong and M. demand and the changing geography of Bradford (2003) Measuring neighbourhood house prices in London, 1995–2006. deprivation: a critique of the Index of Housing Studies 24.3, 301–20. Multiple Deprivation. Environment and Hamnett, C. and T. Butler (2010) The Planning C 21.6, 883–903. changing ethnic structure of housing Ellen, I.G. and M.A. Turner (1997) Does tenure in London, 1991–2001. Urban neighbourhood matter? Assessing recent Studies 47.1, 55–74. evidence. Housing Policy Debate 8.4, Housing Studies (1999) Special issue: 833–66. Towards undivided cities: the potential of Fainstein, S. (2005) Cities and diversity. economic revitalisation and housing Should we want it? Can we plan redifferentiation. Housing Studies 14.5, for it? Urban Affairs Review 41.1, 573–716. 3–19. Imbroscio, D. (2008) United and actuated by Finney, N. and L. Simpson (2009) some common impulse of passion: Sleepwalking to segregation? Challenging challenging the dispersal consensus in myths about race and migration. Policy American housing policy research. Journal Press, Bristol. of Urban Affairs 30.2, 111–30. Friedrichs, J. and J. Blasius (2003) Social Islington, London Borough of (2011) norms in distressed neighbourhoods: Islington’s core strategy — adopted testing the Wilson hypothesis. Housing February 2011 [WWW document]. URL Studies 18.6, 807–26. http://www.islington.gov.uk/Environment/ Friedrichs J., G. Galster and S. Musterd Planning/planningpol/local_dev_frame/ (2005) Life in poverty neighbourhoods: pol_corestrat/ (accessed 15 April 2011). European and American perspectives. John, P., H. Margretts, D. Rowland and S. Routledge, New York, NY. Weir (2006) The BNP: the roots of its Galster, G. (2002) Trans-Atlantic perspectives appeal. Human Rights Centre, University on opportunity, deprivation and the of Essex. housing nexus. Housing Studies 17.1, Journal of Education Policy (2008) Special 5–10. issue: Schools and the urban renaissance.

Journal of Education Policy 23.2, McGhee, D. (2008) The end of 99–198. multiculturalism? Terrorism, integration Journal of Ethnic and Migration Studies and human rights. Open University Press, (2010) Special issue: Linking integration Maidenhead. and residential segregation. Journal of Musterd, S. (2002) Response: mixed housing Ethnic and Migration Studies 36.2, policy: a European (Dutch) perspective. 169–374. Housing Studies 1.1, 139–43. JRF (2005) Mixed tenure twenty years on — Musterd, S. (2005) Social and ethnic nothing out of the ordinary. Joseph segregation in Europe: levels, causes and Rowntree Charitable Foundation (JRF), effects. Journal of Urban Affairs 27.3, Coventry. 331–48. Kesteloot, C. and H. Meert (1999) Informal Musterd, S. and W. Ostendorf (1998) Urban spaces: the geography of informal segregation and the welfare state: economic activities in Brussels. inequality and exclusion in Western cities. International Journal of Urban and Routledge, London. Regional Research 23.2, 232–51. Musterd, S., W. Ostendorf and M. Breebaart Kesteloot, C., A. Murie and S. Musterd (1998) Multi-ethnic metropolis: patterns (2006) European cities: neighbourhood and policies. Kluwer Academic Publishers, matters. In S. Musterd, A. Murie and C. Dordrecht. Kesteloot (eds.), Neighbourhoods of Musterd, S., W. Ostendorf and S. de Vos poverty: urban social exclusion and (2003) Neighbourhood effects and social integration in Europe, Palgrave, mobility: a longitudinal analysis. Housing Basingstoke. Studies 18.6, 877–92. Kling, J., J. Liebman and L. Katz (2007) Musterd, S. and R. Andersson (2005) Experimental analysis of neighbourhood Housing mix, social mix and social effects. Economentrica 7.1, 83–119. opportunities. Urban Affairs Review 40.6, Lees, L. (2008) Gentrification and social 761–90. mixing: towards an inclusive urban Musterd, S., A. Murie and C. Kesteloot renaissance? Urban Studies 45.12, (2006) Neighbourhoods of poverty: urban 2449–70. social exclusion and integration in Europe. Lipman, P. (2008) Mixed-income schools and Palgrave, Basingstoke. housing: advancing the neoliberal urban ODPM (2004) Indices of deprivation 2004. agenda. Journal of Education Policy 23.2, Office of the Deputy Prime Minister, 119–34. London. Lupton, R. (2003) ‘Neighbourhood effects’: ONS (2001) 2001 census data. Office for can we measure them and does it matter? National Statistics, London. CASE Research paper 73, London School ONS (2006) Income support and Job Seekers of Economics, London. Allowance claimant data. Office for Lupton, R. and R. Tunstall (2008) National Statistics, London. Neighbourhood regeneration through Ostendorf, W., S. Musterd and S. de Vos mixed communities: a ‘social justice’ (2001) Social mix and the neighbourhood dilemma? Journal of Education Policy effect: policy ambitions and empirical 23.2, 105–17. evidence. Housing Studies 16.3, 371–80. Marcuse, P., J. Connolly, J. Novy, I. Olivo, C. Popkin, S., D. Levy, and L. Harris (2004) The Potter and J. Steil (2009) Searching for HOPE VI program: what about the the just city: debates in urban theory residents? Housing Policy Debate 15.2, and practice. Routledge, New York, 385–14. NY. Porter, L. and K. Shaw (2009) Whose urban Mayor of London (2008) The London plan: renaissance? An international comparison spatial development strategy for Greater of urban regeneration strategies. London (consolidated with alterations Routledge, London. since 2004). Greater London Authority, Power, A. (2007) City survivors: bringing up London. children in disadvantaged neighbourhoods. Mayor of London (2009) The London Plan: Policy Press, Bristol. spatial development strategy for greater Prime Minister’s Office (2011) PM calls for London (draft replacement). Greater ‘shared national identity’. 5 February 2011 London Authority, London. [WWW document]. URL http://www.

number10.gov.uk/news/latest-news/ in Hoogvliet, Rotterdam. Environment and 2011/02/pm-calls-for-shared-national- Planning A 39.1, 125–41. identity-60300 (accessed 15 April 2011). Urban Studies (2008) Special issue: Raco, M. (2008) Key worker housing, welfare Gentrification and public policy. Urban reform and the new spatial policy in Studies 45.12, 2379–625. England. Regional Studies 42.5, Van Beckhoven, E. and R. van Kempen 737–51. (2003) Social effects and urban Robson, B., M. Parkinson, M. Boddy and D. restructuring: a case study in Amsterdam Maclennan (2000) The state of English and Utrecht. Housing Studies 18.6, cities. Department of the Environment, 853–75. Transport and the Regions, London. Van Gent, W.P.C., S. Musterd and W. Simon, P. (2002) When de-segregation Ostendorf (2009) Disentangling produces stigmatisation: ethnic minorities neighbourhood problems: area-based and urban policies in France. In M. interventions in Western European cities. Martiniello and B. Piquard (eds.), Urban Research and Practice 2.1, Diversity in the city, University of Deusto, 53–67. Bilbao. Van Kempen, R., K. Dekker, S. Hall and I. Southwark, London Borough of (2011) LDF Tosics (2005) Restructuring large housing Core Strategy — adopted April 2011 estates in Europe. Policy Press, Bristol. [WWW document]. URL http:// Watt, P. (2009) Housing stock transfers, www.southwark.gov.uk/downloads/ regeneration and state-led gentrification in download/2648/documents_for_core_ London. Urban Policy and Research 27.3, strategy_adoption ac (accessed 15 April 229–42. 2011). Wilson, W. (1987) The truly disadvantaged: Uitermark, J., J. Duyvendak and R. Kleinhans the inner city, the underclass and public (2007) Gentrification as a governmental policy. University of Chicago Press, strategy: social control and social cohesion Chicago.

Résumé Des deux côtés de l’Atlantique, des politiques sociales et de diversification des régimes d’occupation des logements sont régulièrement mises en æuvre pour lutter contre la pauvreté urbaine et les inégalités sociales (bien que sans preuves solides), certaines mesures entraînant souvent gentrification, polarisation sociale ou dispersion vers les périphéries. Cet article étudie si une diversification des régimes d’occupation permet de freiner la pauvreté en multipliant les opportunités socio-économiques (production) et accès aux ressources (consommation); elle contribue aussi, sur le plan empirique, à des débats généraux sur les incidences des inégalités, de la ségrégation et du voisinage. Des analyses quantitatives et qualitatives longitudinales évaluent dans quelle mesure les opportunités et accès aux ressources, pour les locataires de logements sociaux du Grand Londres, sont supérieurs dans les quartiers où les régimes sont diversifiés par rapport aux quartiers à forte concentration de logements sociaux. Les niveaux de dénuement selon l’Indice de pauvreté multiple (IMD) sont comparés et décomposés. Les corrélations entre régime d’occupation, ethnicité, revenu, lieu, accessibilité financière du quartier et pauvreté sont examinées. D’après les résultats, les habitants de logements sociaux ont effectivement des opportunités et accès aux ressources différents, ceux-ci n’étant ni déterminés ni améliorés par le degré de diversité des régimes d’occupation dans le quartier. En revanche, des programmes propres à un territoire ou à une population, ainsi qu’un accès démarchandisé aux services sociaux (éducation, formation et emploi), sont essentiels. En outre, l’IMD comme argument de base pour formuler des politiques est contesté, et les effets de voisinage sont remis en cause.