DISPLACEMENT OR FORCED RELOCATION? a Comparative Analysis, Following the Relocatees After Urban Restructuring in Amsterdam and Gainesville, Florida

Research Master Urban Studies Graduate School of Social Sciences University of Amsterdam Thesis

Amsterdam, June 13th, 2013

Supervisor: Prof. Dr. Sako Musterd, UvA Second Reader: Dr. Fenne Pinkster, UvA

DISPLACEMENT OR FORCED RELOCATION? A comparative analysis, following the relocatees after urban restructuring in Amsterdam and Gainesville, Florida

Koen Tieskens Student#: 0542520 [email protected]

Foreword

This thesis is the result of a research conducted to obtain the degree of Master of Science finishing the research master Urban Studies at the University of Amsterdam. The research was conducted in Amsterdam, NL and in Gainesville, Florida. The fieldwork in Gainesville was conducted during my exchange program at the Florida State University in Tallahassee. I would like to use this opportunity to thank my supervisor from FSU Dr. Andrew Aurand, who assisted me in finding appropriate data sources, which appeared to be more difficult than I expected. Moreover I would like to thank Elizabeth Thompson from the Shimberg Center at University of Florida, who was very kind to provide me with data on relocation of tenants in Gainesville.

The fieldwork in the Netherlands was conducted very smoothly not in the least due to the assistance of the relocation staff from Amsterdam housing association Rochdale. Thanks to their willingness to share their archives with me and their professional insights, I was able to investigate forced relocation in the Netherlands as it was never done before.

Parts of the results of this thesis were also used to write an article that was accepted by Journal of Urban Research and Practice and will most likely be published in the second issue of 20131. This article was written under the supervision and co-authorship with prof. Dr. Sako Musterd. He not only assisted me with the article, but throughout the whole process of this thesis. I would like to thank him as well for the assistance and inspiration. And finally I would like to thank my family, friends and my girlfriend Sarah who supported me in writing this thesis.

June 13, 2013

Koen Tieskens

This thesis was written to obtain the degree of MSc at the University of Amsterdam in co-operation with housing association Rochdale.

1 Parts of this thesis will therefore also be published in this article: Tieskens KF and Musterd S. (2013) Displacement and urban restructuring in Amsterdam, following relocatees after demolition of social housing. Journal of Urban Research and Practice.

Table of Contents 1. Introduction ...... 4 2: Assisted Housing: an historical background ...... 8 2.1: US Housing market ...... 8 2.1.1: Public Housing ...... 9 2.1.2: Urban Decline ...... 10 2.2: Europe ...... 13 2.2.1: Decline ...... 15 2.3: Social mix and neighborhood effect ...... 18 3: Tracking the relocatees ...... 21 3.1: US ...... 21 3.1.1: Gautreaux and MTO ...... 21 3.1.2: HOPE VI ...... 23 3.2: The Netherlands ...... 25 3.3: Conclusion...... 27 4: Methodology ...... 30 4.1: Questions ...... 30 4.2: Macro-Micro-Macro ...... 31 4.3: Conceptual Model ...... 33 4.4: Data ...... 34 4.4.1: Population ...... 35 4.5: Operationalization ...... 36 4.5.1: Relocatee Characteristics ...... 38 4.5.2: Preferences ...... 39 4.5.3: Information and Assistance ...... 39 4.5.4: Institutional Constraints...... 40 4.5.5: Relocation ...... 40 4.5.6: Housing Improvement ...... 41 4.5.7: Neighborhood Quality ...... 41 5. Gainesville ...... 43 5.1 Introduction ...... 43

5.2 General Characterization Gainesville...... 45 5.2.1 Racial segregation ...... 46 5.3: Description of the three projects ...... 48 5.3.1: Glen Springs Manor ...... 48 5.3.2: Seminary Lane ...... 52 5.3.3: Kennedy Homes ...... 55 5.4 Outcomes ...... 58 5.4.1: Section 8 ...... 58 5.4.2: Local Housing Market ...... 60 5.4.3: Relocation ...... 62 5.5: Discussion ...... 67 6. Amsterdam ...... 69 6.1 Context ...... 69 6.1.1 Amsterdam...... 69 6.1.2 Amsterdam Nieuw West ...... 70 6.2 Case description ...... 71 6.2.1 Bakema Sloop Noord ...... 71 6.2.2 Van Tijenbuurt ...... 72 6.2.3 Kolenkitbuurt ...... 73 6.3 Outcomes ...... 77 6.3.1 Preferences ...... 77 6.3.2 Actual Relocation ...... 79 6.3.3 Quality of housing ...... 84 6.3.4 Rent ...... 86 6.4 Summary ...... 86 7. Discussion and Conclusions ...... 87 7.1 Comparison ...... 89 7.2 Conclusion and recommendations ...... 90 References ...... 93

1 Introduction Many cities are facing in Western post-industrialist societies are facing the accumulating problems in inner city neighborhoods. Throughout the world and throughout time, city governments have tried to combat concentrations of poverty in order to prevent the rise of slums and the accumulation of problems. One of the great examples is the attempts of Robert Moses to redevelop large parts of New York City in the decades after WWII. On his account many low- income dwellings were demolished and replaced with high- and middle-income housing. The low-income families –mostly black or Latino- were forced to move to the public housing projects elsewhere or find other alternatives. The urban redevelopment or slum clearance became also known as “negro removal” (Schwartz, 1993). One of the most notable critics of this policy, Jane Jacobs, argued that removing the poor out of a neighborhood was not solving any problem; at best it was relocating the problem (Jacobs, 1993). Strikingly the issue of urban redevelopment, slum clearance and relocating the poor is anno 2013 more than actual, more than 40 years after Jacobs’ seminal work on the great American cities. However the words and rhetoric have slightly changed. ‘Mixed neighborhood’ appears to be the magic word in urban policy, not only in US mega cities but also in Europe (Bridge et al., 2012).

Across Europe, concentrated poverty tends to be treated as an undesirable concept with unwanted segregated neighborhoods as a result. Social and ‘ethnic’ mixing is generally seen as the suitable policy to fight against it (Musterd 2005).These mixed neighborhoods would cater for better opportunities in terms of housing, social mobility opportunities for less affluent residents, and would supposedly create better conditions for the integration of immigrants, many of whom have different ethnic backgrounds. In many western cities, low-income households tend to be over represented in areas characterized by relatively low-quality housing, while frequently these areas also have higher crime rates and experience higher levels of unemployment.

A rather rigorous yet widely used instrument to de-concentrate relative poverty and obtain heterogeneous neighborhoods is encapsulated in the policy of urban restructuring. The extreme forms of such urban restructuring are characterized by objectives such as tearing down poor- quality housing structures, moving tenants to other parts of the city, building new housing of higher quality, with higher rents or for ownership, and attracting more affluent residents in order to create the transcendent mixed neighborhoods. This type of radical intervention is based on two

4 assumptions: social mix – which overlaps with ethnic mix – would enhance the opportunities of the urban disadvantaged, and second, mix of housing would create social and ethnic mix (Musterd and Andersson, 2005).

Social mixing policies are said to stem from worries about lack of ‘integration’, which eventually might even result in the development of ‘parallel societies’ that are considered a threat to national unity and social order (Uitermark, 2003; Münch, 2009; Phillips, 2009). However, even in the US, where concentrations of poverty and race are much more intense than in Europe (Peach, 2009), evidence in support for social mix and the assumed effects is limited and often weak (for an overview, see Goetz and Chapple, 2010). In Europe, the negative effects of distressed neighborhoods on the opportunities of its residents are even more contested (Van Ham and Manley, 2010; Ostendorf et al., 2001; Musterd et al., 2012).

While social mix policies have been applied to an increasing extent, academic debate reflecting on it developed at the same pace. Comparable with Jacobs’ critics of urban redevelopment in Manhattan, Imbroscio (2011) dispraises the ‘mobility paradigm’ for not dealing with the problem at its heart. Instead he calls for a placemaking paradigm which ‘puts the emphasis on securing the necessary supply of affordable housing in urban neighborhoods, in order to mitigate the degree of displacement from the pressures of gentrification’ (Imbroscio, 2011: 13). Lees (2008) pointed out that social mix strategies may also serve completely other objectives. She and others argued that social mixing policies conceal gentrification strategies, which are often actively encouraged by the state, in partnership with private actors, but which, as a result, would displace lower-income households from attractive locations in the urban arena. Other authors also find this displacement effects (Smith, 2002a; Smith, 1996)

They argue that the original low-income population of redeveloped neighborhoods is forced to make place for middle-class ‘gentrifiers’. By being displaced they become the victims of the mixed neighborhood policies of many local governments. Most studies on these types of relocation have had this gentrification perspective with a very negative attitude towards the displacement of the original population (Kleinhans and Kearns, 2013). However recent studies show that displacement after forced relocation in the Netherlands should not be perceived negatively by definition (cf. Posthumus, 2013). Many scholars have tried to find out what happens with relocatees after they are being forced to relocate to make place for those mixed

5 neighborhoods. The very fact that these people have been relocated makes them hard to study since it is often unknown where they have moved to. In the US many studies describe the relocation but often lack good data on the characteristics of relocatees while Dutch studies often have detailed data on relocatees but rely on questionnaires with response rates around 30%.

In order to find out what happens with relocatees after forced relocation I have tried to investigate this process by using highly detailed data on four relocation projects in Amsterdam. The data contain the social demographic characteristics of all relocatees and their exact new location. The results of this Amsterdam study will be compared with a similar investigation of three forced relocation processes in the US, in Gainesville, Florida. However the data on Gainesville are less detailed and complete, but nevertheless can reveal patterns of forced relocation due to urban restructuring. The most important goal of this thesis is to find out whether the relocatees who are actually relocated from their original location could hence be regarded as victims of the mixed neighborhood policies. Therefore I have not only looked at the relocation outcomes but I also incorporated the preferences of relocatees and compare them with the actual outcomes. The questions that I will try to answer in this thesis are:

1. What are the location preferences of households who are involved in large-scale restructuring of their neighborhood? 2. How do these preferences relate to the actual residential relocation due to restructuring? 3. To what extent are relocatees (not) improving their housing situation after forced relocation and how does that differ between relocatee types? 4. How do the outcomes of the relocation from Amsterdam differ from those in Gainesville, FL? These four questions should then help to answer the question posed in the title of this thesis: forced relocation or displacement? Relocatees are forced to move but are they also displaced with all its negative connotation? Essential to the comparison is the different institutional contexts in which the forced relocation takes place. The relocatees often are low-income and therefore are dependent on the supply of subsidized housing and the right they gain due to their forced relocation. Not very surprisingly both these rights and the supply of subsidized housing differ enormously between the US and the Netherlands. Therefore the chapter after this

6 introduction consists of a brief history of subsidized housing in both countries and through what kind of policies low-income families are assisted in their housing finance. Then Chapter 3 discusses previous research on forced relocation while in the next chapter the methodology is explained. The following empirical chapters discuss the outcomes of relocation in respectively Gainesville and Amsterdam. The last chapter contains a discussion on the outcomes including a comparison between the two cases and a final conclusion.

2: Assisted Housing: an historical background In the two decades after the Second World War, both in the US (Schwartz, 2010; Florida and Feldman, 1988) as well as in large parts of Europe (Turkington et al., 2004; Rowlands et al., 2009), large mass housing estates emerged in rather similar fashion. The inner city slum dwelling could not meet the demands anymore of the rising urban underclass. An understanding of the history of public and social housing is essential for discussing contemporary urban problems such as displacement, relocation and neighborhood effects. In this chapter I will discuss how mass housing has emerged both in the US as well as in Europe focusing more on the Netherlands. I will start with elaborating on assisted housing in the US.

2.1: US Housing market After the Great Depression of 1929 the US federal government was increasingly supporting homeownership among its residents. Many institutions were established in order to both provide and secure homebuyers mortgages which have significantly decrease the costs of homeownership. Examples were the Federal Housing Administration (FHA) and later the Home Owners Lending Corporation (HOLC) and Fannie Mae. Although the FHA, HOLC and Fannie Mae provided access to the homebuyers market for the working-class, many regard these same institutions at least complicit to growing racial and spatial inequality (Jackson, 1987; Fishman, 2000). FHA was created mainly to insure mortgages. However they could obviously not insure every single mortgage and therefore they had adopted a set of requirements a mortgage had to meet before it was insured by FHA.

Jackson (1987) discusses how federal policies of the HOLC and FHA have encouraged suburbanization and how their discriminatory practices partly initiated a white flight out of the urban centers while African Americans were trapped inside the inner-city boroughs. One of the most notable requirements for an FHA default insurance was based on neighborhood appraisal by FHA officers. A standardized federal measurement system of neighborhood appraisal was created, that decided where insurances and loans were given. The presence of African American families in a neighborhood almost always let to a negative appraisal, which made the neighborhood unattractive for investments. According to HOLC appraisers “an entire area could lose its investment value if rigid white black segregation was not maintained” (Jackson, 1987: 214). In this fashion the appraisals worked as a self-fulfilling prophecy. Jackson explains how

8 the infamous practice of redlining led to stigmatization of, and structural disinvestment in neighborhoods where African Americans were residing. White middle-class families were given the opportunity with a federally insured loan to buy a home in the suburbs with monthly payments that were sometimes even twice as low as their rents in the inner cities (Jackson, 1987: 204). Not only had the white flight catered for the urban decline, but the negative appraisal of FHA officers made sure no significant investments were made in African American populated inner cities. The American dream of owning your own home in the suburbs was only meant for white families. Even the presence of one African-American family within a neighborhood was enough for an FHA appraiser to deny any mortgage insurance, since one of the most important criteria was the “protection from adverse influences” (Schwartz, 2010: 55). The presence of an African American was regarded as being one of these ‘adverse influences’ (Immergluck, 2004; Jackson, 1987).

2.1.1: Public Housing The New Deal was however not only concerned with homeownership and mortgages. One of the last parts of this plan to get the US out of the Great Depression was the public housing program, now by far the most widely known form of subsidized housing in the US (Schwartz, 2010). Interrupted by the Second World War it was seriously taking its shape in 1949 when Congress passed the Public Housing Act. During the 1950s, 1960s and 1970s over a million public housing units were built. In the larger cities public housing consisted mostly of the well-known high rise modernist towers, but a larger share of the total public housing stock was low-rise town homes, row-type housing or single family dwellings. Despite the mixed characteristics, public housing is easily recognizable due to its sober appearance, Spartan amenities and inferior quality. Even when Soviet architects were visiting a project in Chicago during the 1980s, they were devastated by the inferior quality of the public housing projects (Schwartz, 2010). The lack of quality was mainly due to the lack of funding for maintenance. Public housing was built with federal funds, but had to be maintained with the money that was coming in from rent. Since public housing was populated by the poorest of the poo, the rent was obviously not enough to keep the buildings in proper shape.

Originally public housing was meant for working class families who were due to reasons beyond their fault, unable to find a dwelling in the private market. However since FHA was providing or

9 insuring relatively cheap mortgages, more affluent residents of public housing increasingly shifted to the private market or buyers’ market. Public housing projects had to cope with an increasingly poorer population: the median income of public housing dwellers fell from 57 per cent of the national median income in 1950 to less than 20 per cent of the national median income in 1990.

Important to note is that the decision to build public housing was in hands of local authorities. If a jurisdiction did not want to build any public housing it was free to do so. The decision to build public housing therefore became a matter of tax base and vocational support. Affluent white suburbs did not want any public housing projects in their jurisdictions while African American politicians wanted to keep public housing in their own neighborhood for vocational support. Although the overall population of public housing is fairly mixed with 51 per cent white families and 45 per cent African American, the individual projects became increasingly segregated with most large high-rise public housing projects being more than 90 per cent African American. More than 75 per cent of all minorities in public housing lives in central cities compared to 45 per cent of all rental units nationwide. (Turner et al., 2009; Schwartz, 2010).

2.1.2: Urban Decline The worsening conditions of public housing in the inner cities are illustrative for the general decline which occurred widespread across the US in inner cities. The federal policies promoting owner occupation in the suburbs, the segregating construction of mass public housing combined with the increasing dismantling of heavy industries and large factories led to urban decline in almost every large American city. The most important features of urban decline are loss of population, rising unemployment, and a growing number of people relying on social welfare systems (Friedrichs, 1993). As a result cities had to cope with impoverished inner cities, large ghettos and public housing projects as breading grounds for criminality. Perhaps the best example of urban decline could be found in the well-known failure of public housing project Pruitt Igoe in post-war St. Louis. Pruitt and Igoe was a huge project finished in the early 1950s, consisting of 13 eleven story structures, housing 15.000 people at higher densities than in the original slum.

Soon after the opening the project had started to deteriorate and conditions were getting worse every year. Eventually the conditions got so bad; there was no other option than to demolish the

10 entire projects. This demolition took place in 1974; only 20 years after the construction. Many reasons can be given for the decline. However the most often heard argument for the deterioration of Pruitt Igoe is the mismatch between architecture and population (Yancey, 1972). The project consisted of so many long corridors with no explicit owner or group of people feeling responsible to keep things tidy. Newman (1972) explicitly mentions Pruitt Igoe in his theory on the direct relation between physical environment and human behavior. The lack of so- called defensible places (i.e. demarcated semi-public areas where a small group of people feel responsible to ‘defend’ it) invokes unwanted behavior such as happened in Pruitt Igoe. However others have emphasized more structural causes for he failure, which might be generalizable to most public housing projects. Two main arguments can be identified. First some argue the whole public housing program was deemed to fail due to the impossible financing scheme (Bristol, 1991). The money from rents could have never been enough to maintain the massive housing projects; the projects were already a failure before they were even built. Secondly many mention structural forces such as white sub-urbanization, institutional racism and perverse effects of federal policy as causes for urban decline such as occurred in Pruitt Igoe (Jackson, 1987; Bristol, 1991; Fishman, 2000).

The latter arguments can easily be linked to a wider body of literature discussing the general decline of inner cities. Pruitt Igoe definitely did not stand on its own: it is often used as an explicit example of the failure of the modernist mass public housing, and urban decline within inner American cities. Perhaps the most important writing on this matter was by Wilson (1987). Wilson argues that trough racial discrimination and sub-urbanization, inner cities were left with a young black population. The shift from goods production to the service industries caused polarization on the labor market and led to growing unemployment for the young black population of the inner cities. Together with growing out-migration of working- and middle-class families the concentration of the disadvantaged was increasing every year. A mismatch of jobs in the suburbs and residence in the inner city and a lack of social ties with middle-class or working-class has been disastrous for the black population of the inner cities. The civil rights movement and institutional action against discrimination created opportunities for working-class and middle-class black families to escape the ghettos, but leaving an even poorer population behind (Wilson, 1987). Massey and Denton (1993) agree with Wilson on growing concentrations of poverty in inner cities but emphasize it is mainly racial segregation that is causing these

11 concentrations. Worsening conditions, alarming studies and perhaps the riots of 1992 in LA fed the public concern with the ‘ghetto problem’ (Jargowsky, 1997).

The federal government applied several programs to counter the negative effect of this ghetto problem by assuming that living in a poor neighborhood decreases one’s opportunities, especially due to the mismatch in supply and demand of labor (Friedrichs et al., 2003). This assumption has led to solutions based on the mobility of residents from poor neighborhoods to neighborhoods where there were better opportunities in terms of labor, education and housing (Briggs, 1997; Imbroscio, 2011). One of the first and most famous attempts to do so was the Gautreaux program in Chicago which was established in 1976 and ran until 1998 (Rosenbaum, 1995). In this program African American households could apply to leave their impoverished neighborhood by obtaining a Section 8 voucher. A Section 8 voucher allows one to rent in the private market while paying only 30% of the income. The difference is paid by the federal department of Housing and Urban Development (HUD). Within the Gautreaux program the households would get assistance to use the voucher within a more affluent ‘white’ neighborhood. Early results of this program did reveal positive effects for its participants in terms of education and employment (Rosenbaum, 1995). Inspired on these successes HUD came with an even larger program called Moving To Opportunity (MTO). MTO worked rather similar as Gautreaux but was set up larger and had some minor differences which will be discussed later. A third and last program that relied heavily in residential mobility in order to counter problems of concentrated poverty was the massive revitalization and Section 8 relocation of distressed public housing sites under the name of HOPE VI (Housing Opportunities for Everyone).

However many scholar have been increasingly critical towards HOPE VI since convincing evidence of improving conditions for those displaced by HOPE VI has yet to be found. The lack of evidence for benefits of residential mobility has been fueling discussion on more place-based solutions, dealing with black urban poor neighborhoods ‘as is’ (Pattillo, 2009; Imbroscio, 2011). Noteworthy is Michael Porter’s ongoing research translating his seminal work on the competitive advantage of nations into the competitive advantage of inner cities (Porter, 1995). As opposed to the mobility programs these scholars are arguing to bring jobs to the inner city instead of moving workers out of it (cf. Imbroscio, 2011) (but see Briggs, 2008: for a reply on Imbroscio's critiques on mobility). The question still remains: does neighborhood matter? Can a poor family gain from

12 living in a mixed neighborhood? And inherently do poor families suffer from living next to other poor families? Are poor people bad for each other? The answers to these questions reach beyond the scope of this research. However the results of this study might provide better insights in the actual outcomes of so-called mobility programs.

2.2: Europe As was the case in the US, urbanization due to the industrial revolution housing the new urban underclass became a major issue for countries in Europe. Rowlands et al. (2009) identified five different reasons why mass housing appeared so widely across Europe in the first decades after the Second World War. (1) The wave of industrialization which swept through Europe during the late nineteenth century and the beginning of the twentieth century, catered for a massive urbanization and rapid population growth of cities. The housing which was provided during that period was provisional and resulted in concentrated poor quality housing for the working class that was often described as slums. Already before the war, plans were being made to improve the quality of the housing stock. The war however had major impacts on housing needs (2). Not only due to the devastation it had created was there a greater need for housing, the unprecedented baby-boom resulted in a huge housing shortage. Most Western European countries had to cope with housing shortages of over 10% (Rowlands et al., 2009). Scarcity on the housing market and the need for rapid economic regeneration after the war resulted in rising housing prices and the disability of markets to cope with the huge shortage. Cheap massive construction could only be realized through state interventions (3). Together these demographic changes, the ruins of the war and the emerging welfare-states in Europe, saw and ideal solution in the emerging modernist architecture (4), based on the functionalists ideals of Congrès Internationaux d’Architecture Moderne (CIAM) formed by the modernist architect Le Corbusier. Technocratic solutions to provide housing for the poor by separating function were the adagio of modern architects. Since this offered a relatively cheap solution to urban problems and a tool to gain more control of the urban population, the modernists’ view on mass housing gained wide support across Europe (5). In rapid pace massive high-rise housing estates appeared in Europe during the 1950s 1960s and early 1970s (Wassenberg et al., 2004).

Despite these seemingly similar conditions for many countries in Europe, post-war housing estates developed in different ways and problems arose in different extents and of different kinds

13 in each country or region. Since this study is a comparison between the US and the Netherlands, I will focus on post-war housing in the Netherlands. However I will also point out to what extent the Dutch context differs from the situation in other countries within Europe.

Formal intervention in housing of the less affluent started in the Netherlands in 1901 with the first Housing Act. After this act social housing in the Netherlands was in the hands of non-profit housing associations that were funded with loans from the national government. Housing associations acted under supervision of municipal governments which in turn were responsible for eventual losses. The social sector had a slow start, although some new housing was built by the associations, the national government was reluctant in providing loans (Van der Schaar, 1983). The social sector actually commenced after the Second World War when the demand for new housing was going through the roof and the welfare state was establishing in the Netherlands. Until the early 1990s association kept building new housing. They reached their plateau, owning 42 per cent of the housing stock in the country during the 1990s (Priemus, 2006). Number of Units in Social Rent

1600000

1400000

1200000

1000000

800000 Number of Units in Social Rent 600000

400000

200000

0 1915 1925 1935 1945 1955 1965 1975 1985 Figure 2.1: figure 1 shows the total number of social housing units throughout the 20th century. After WWII a sharp rise in the number of social housing units is visible. Plans that were made before the war could be realized and the big demand for housing was increased by war devastations and the post-war baby boom

Until then, housing associations were very tightly connected to the government. Since they were providing housing to low-income and middle-income tenants below the cost price, they were dependent on subsidies and other state funding. Neoliberal regimes however were transforming housing policies across Europe. For example, the Thatcher administration in the UK introduced ‘Right-to-Buy’ in 1980, increasing owner-occupation from 50 to 67 per cent in 1993 (Priemus, 1996). But even in strong social democracies such as Sweden, housing subsidies were dramatically decreased during the 1990s (Turner and Whitehead, 2002). The Netherlands had their own special way of liberalization of the housing market. Owner occupation increased to 50 percent, but the social rented sector remained stable at over 40 per cent during the 1990s. However the Dutch government created a unique situation by more or less privatizing the social housing sector. On the first of January 1995, the housing association repaid all their loans to the central government and received all their subsidies they were expecting for the coming 50 years all at once (€16.6 billion) to cut all direct ties with the government. Housing associations still have the task to provide housing at reduced prices for those who cannot afford to rent in the private market. To raise money, housing associations now sell property which is written in their book far below market rate and rent out a share of their stock in the private rental market.

2.2.1: Decline Whereas in the US public housing is intended to house the most deprived part of the population, most European countries provide social housing for a much larger population (Schwartz, 2010; Turner and Whitehead, 2002). A percentage of 42 of the total housing stock, means that social housing is not strictly for the under-class. Middle-class and even a few upper-class households live in social housing in the Netherlands (Van Kempen and Priemus, 2002). Projects described as war-zones and breeding grounds for criminality like Pruitt Igoe have never existed in the Netherlands. As Peach (2009) argues, Chicago-style ghettos simply do not exist in Western Europe. Peach argues that even in the UK the neighborhoods with the highest concentrations of minorities do not resemble the ghettos of the US if proper measurement is used.

Nevertheless segregation and poverty concentration is regarded as a serious issue in Europe. Concentrations of poverty and ethnic segregation appeared frequently in those neighborhoods that were built in the first decades after the Second World War. These neighborhoods, at first built for lower middle classes, were increasingly populated by ethnic minorities and lower

15 classes during the 1980s and 1990s. Problems arose such as unemployment, criminality and school drop-out. However not all post-war estates were coping with the same problems to the same extent (Rowlands et al., 2009). Physical determinism, blaming solely modernistic architecture for the decline also emerged in Europe but just as in the US appears to be short- sighted. Many scholars have come up with different sets of reasons for the decay in these neighborhoods. There are so many reasons for the decay, and simultaneously there are so many different post-war neighborhoods in Europe, so it is difficult to mention one specific reason for decline (for an overview see Van Beckhoven et al., 2009).

In the 1990s and onwards most post-war housing estates are far less desirable than they once were when they were constructed. The population of these estates has changed, the middle class has largely left these areas and mostly ethnic minorities, who had less choice in the housing market have replaced them (Musterd and Van Kempen, 2007). Although compared to American ghettos the levels of segregation are relatively low; concentration of poverty and ethnic segregation is by no means a trifle in the European political arena (Phillips, 2009). Even the European Commission warned for growing concentrations of poverty and ethnic minorities, calling for radical action. In many countries such as Belgium, Germany, England, Sweden and the Netherlands the central government planned for action to counter growing segregation and concentrations of poverty (Musterd, 2003). It appeared as though the panacea for emerging problems was social mixing (Uitermark, 2003). The policies that were invented to mitigate the emerging problems are based on two assumptions: social mix – which overlaps with ethnic mix – would enhance the opportunities of the urban disadvantaged, and second, mix of housing would create social and ethnic mix (Musterd and Andersson, 2005). In most Western countries these social mixing instruments were put into practice in more or less similar ways (Andersen and Van Kempen, 2003). Examples are the Danish Urban Renewal Act, revitalizing post-war functionalist neighborhoods in Denmark, whereas in France urban renewal was carried out under the Politique de la Ville of the late 1990s (Gilbert, 2009). Next to physical interventions, these policies tried to change the social and ethnic compositions of the targeted areas, by dispersing the poor population and through an influx of middle-class. In the Netherlands policymakers tried to accomplish social mixing through the Big Cities Policy (Grotestedenbeleid).

In the second half of the twentieth century and the years thereafter, there have been numerous area-based policies to combat problems related to urban decline and stimulating urban economies (for an overview see Musterd and Ostendorf, 2008). The introduction of the Big Cities Policy (BCP) in 1994 can be regarded as a starting point for social mixing by means of introducing middle-class into distressed neighborhoods (Musterd and Ostendorf, 2008). Councils of the four major cities in the Netherlands (Amsterdam, Utrecht, Rotterdam and The Hague) saw problems such as increasing unemployment, criminality and urban decline growing rapidly, especially within these four big cities. To combat these problems, in 1994 they created a memorandum for the newly formed government calling for policy specific attention for problems in their big cities. Since the major goal of the newly installed government was the creation of jobs they were willing to incorporate the memorandum of the four cities and the BCP was initiated (Priemus et al., 1997). The policy was designed to target the most vulnerable groups in society: chronically unemployed, ethnic minorities, elderly, disabled etc. But more importantly it was also designed to target specific neighborhoods where concentration of these people existed to improve these neighborhoods both physically as well as socially (Van Kempen, 2000). Soon after its introduction other municipalities than the four big cities joined the program since they were facing similar problems. BCP I and its follow-ups II and III were all aimed at combating segregation. On the one hand by providing more mixed housing in disadvantaged neighborhoods so middle-class could move in, but on the other hand by providing opportunities for the original population to make housing career. In such a way successful residents of disadvantaged neighborhoods could remain in their neighborhood and out-migration of the successful would be omitted (Musterd and Ostendorf, 2008).

From the period between 1994 and 2002 there was a special minister allocated for the BCP. Between 2002 and 2007, when the tone of the political debate changed from fostering ‘multi- culturalism’ to forced assimilation and integration, the BCP was under the responsibility of the minister of Justice. Since 2007 it is again under a minister of housing and integration. Despite the large role central government took in implementing the BCP, one of the keywords was decentralization (Musterd and Ostendorf, 2008). The BCPs were implemented with money from central government (and to a small extent even from the EU) allocated to municipalities. These municipalities made a plan, approved by the central government, but together with other stakeholders such as housing associations, but also residents. The operationalized version of the

BCPs are characterized by objectives such as tearing down poor-quality housing structures, moving tenants to other parts of the city, building new housing of higher quality, with higher rents or for ownership, and attracting more affluent residents in order to create the transcendent mixed neighborhoods. This has happened in many cities in the Netherlands on very large scale roughly in the first decade of the 21st century. A good example of this large scale restructuring can be found in Amsterdam Nieuw West (ANW).

ANW is more or less completely built in the first decades after WWII and consisted in 2002 of slightly more than 52.000 housing units (O+S Amsterdam, 2002). Under the name of Project Parkstad the entire post-war area in western Amsterdam was restructured. The plan was to demolish 13.300 units and replace them with 24.300 new units. The percentage of social rent would decrease from 76 to 46 which is closer to the national average (Van Beckhoven, 2007). The ten active housing associations in ANW bundled their assets into two consortia to work together with the central government, the municipality of Amsterdam and the residents on the massive urban restructuring, relocating thousands of people and creating many high-end rental and buyers dwellings.

Strikingly the discussed policies received only very little academic support in comparison with the largeness of the interventions. Musterd and Ostendorf (2008: 88) summarize the history of Dutch urban policies as follows: “[Urban policy had] strong focus on area-based approaches in disadvantaged neighborhoods, aiming to change the housing stock in order to create a social mix. With this ambition the picture of the real situation is out of sight. This is problematic, because the empirical situation differs considerably from the political perceptions and discourses”. A mixed neighborhood does not appear to stimulate social interactions between different social classes. Moreover Fenne Pinkster (2008) showed in a Dutch case study of several mixed income neighborhoods that social mixing can even destroy important social ties while not generating more. Therefore social mixing might even have more negative influence on the targeted area than it has positive effects.

2.3: Social mix and neighborhood effect In both the US as well as in the Netherlands (and in the rest of Western Europe) ethnic or income segregation and concentrations of poverty were regarded as one of the most pressing urban problems. Many policies large scale policies such as HOPE VI, MTO and the Dutch BCP have

18 tried to decrease segregation and concentration of poverty in order to mitigate for accumulating problems such as unemployment. All these policies are based on the assumptions that mix of housing creates social mix, which in turn enhances the opportunities of residents of disadvantaged neighborhoods. The relationship between neighborhood and its residents is however one that is very difficult to investigate or understand. There is some sort of consensus about the question if neighborhood matters; although scholars tend to agree that there is definitely an existing neighborhood effect, actual causal mechanisms proved to be difficult to tease out with empirical studies (Friedrichs et al., 2003; Galster, 2007; Ellen and Turner, 1997; Galster, 2012). Social mixing policies are said to stem from worries about lack of ‘integration’, which eventually might even result in the development of ‘parallel societies’ that are considered a threat to national unity and social order (Uitermark, 2003; Münch, 2009; Phillips, 2009). However, even in the US, where concentrations of poverty and race are much more intense than in Europe (Peach, 2009), evidence in support for social mix and the assumed effects is limited and often weak (for an overview, see Goetz and Chapple, 2010). In Europe, the negative effects of distressed neighborhoods on the opportunities of its residents are even more contested (Van Ham and Manley, 2010; Ostendorf et al., 2001; Musterd et al., 2012).

As I already pointed out, some scholars even found neighborhood effects that accomplish the opposite of what policymakers had in mind applying for example the BCP. Other studies have extensively followed families throughout their housing career and tried to find effects of certain neighborhoods on people’s employment and education. If there is so little evidence for positive neighborhood effects through social mixing, why do these social mixing policies persist so widely? Imbroscio (2011) argues that stubborn advocates of liberal urban policy have always relied on mobility, and if mobility does not work, “the preferred corrective is even more mobility” (Imbroscio, 2011: 3). Lees (2008) pointed out that social mix strategies may also serve completely other objectives. She and others argued that social mixing policies conceal gentrification strategies, which are often actively encouraged by the state, in partnership with private actors, but which, as a result, would displace lower-income households from attractive locations in the urban arena. Other authors also find this displacement effects and regard gentrification as a new global strategy (Smith, 2002a; Smith, 1996). In the Netherlands gentrification is however not used as a strategy by local governments to increase tax-bases or serve the middle-class, but explicitly to counter social problems in the city and actively pursue

19 social mix (Uitermark et al., 2007). Gentrification has increasingly been associated with the negative effect of displacement of working-class from attractive neighborhoods (Atkinson, 2004; Bridge et al., 2012). Therefore if policies of urban restructuring, with the neighborhood effect/social mixing assumptions are evaluated, one has to look further than just the targeted neighborhood. What about those who are displaced? Aside from the restructuring effects on the targeted area, which can still mean an upgrading of a previously disadvantaged neighborhood, urban restructuring in the form of demolition of housing and replacement of more expansive dwellings, the relocation or displacement of the original population can at least have two other effect. On the one hand the relocation of less affluent residents can mean that other neighborhoods receive disproportionately many poor residents as a result from the relocation that concentrations of poverty can grow or appear in other parts of a city. In this line of thought urban restructuring is not dispersing poverty but is merely relocating problematic areas to other neighborhoods. Dutch scholars have found these types of patterns of these types of spillover- effects (Slob et al., 2008; Bolt and Van Kempen, 2010). On the other hand, more along the lines of the state-led gentrification theories, the urban restructuring is not causing the promised neighborhood effects and is merely displacing the original population. In this more structuralist and revanchist view, the middle-class is taking over attractive parts of the city while the working-class is paying high tolls (Newman and Wyly, 2006). Especially in the US and in the Netherlands there is a growing interest in displacement after urban restructuring. Several studies have tried to track down households that were relocated due to urban restructuring to define patterns of displacement and find out to what extent these relocatees are victims of urban restructuring and if new concentrations of poverty have risen. In the next paragraph I will elaborate more on these tracking-studies.

3: Tracking the relocatees Research with a specific focus on displacement of tenants of restructuring neighbourhoods has attracted ample attention, especially from American and Dutch scholars. In this section I will discuss findings of tracking studies of relocatees in regard with poverty de-concentration and/or urban restructuring. I will first start with discussing outcomes of relocation in the US.

3.1: US In the US programs that involved the relocation of households explicitly try to disperse disadvantaged households to lessen poverty concentrations. The best known examples are Gautreaux, MTO and HOPE VI (Goetz and Chapple, 2010). An important difference however is that the first two programs involve voluntary relocation where households could sign up to get involved, while HOPE VI forces people to move due to the demolition of public housing (Goetz, 2002).

3.1.1: Gautreaux and MTO Gautreaux was one of the first voucher programs in the US and started as early as 1976, relocating nearly 7600 mostly African American families out of the ghetto to more white neighborhoods. Gautreaux was initiated after a lawsuit of an African American woman challenging the Chicago Housing Authority for only providing housing for African Americans within the public housing black ghettos. After Gautreaux won the lawsuit, the Gautreaux program initiated trying to offer opportunities to African American households to relocate to white suburbs (Rosenbaum, 1991). Since low-skilled jobs were moving more and more from the inner cities to the suburbs and since suburban schools are generally better than their city counterparts a move to the suburbs could provide these African American households with many more opportunities than they had in the inner cities (Wilson, 1987). A disadvantage was however that these African American families can be regarded as black pioneers in white enclaves in the suburbs. Grocery stores in the suburbs never had to cope with food-stamps, suburban schools never experienced children with less affluent parents and most importantly, African American families did receive harassment from their white neighbors moving into their neighborhood (Rosenbaum, 1991). Since the Gautreaux program was one of the first of its kind, it was an opportunity for researchers to find out if a move from a disadvantaged neighborhood in the inner city to a white neighborhood in the suburbs or somewhere else would cater for better opportunities and eventually to improvement in employment and education of African American

21 participants. The first results came in the early 1990s and were strikingly positive. Adult participants showed significant improvement in earnings and employment while children reported better results in school. Especially those who moved to the suburbs experienced these improvements (Popkin et al., 1993; Rosenbaum, 1995; Rosenbaum, 1991). More recent research, investigating long-term effects of Gautreaux moves show that especially moving to a neighborhood with lower concentrations of African Americans is beneficial for employment (Mendenhall et al., 2006).

The positive results of the Gautreaux inspired HUD in 1994 to launch its widely discussed experimental program Moving To Opportunity (MTO). MTO was launched in five American metropolitan regions, Chicago, Baltimore, Los Angeles, New York and Boston. In this experimental program targeted at relocating people from the poorest neighborhoods to largely suburban neighborhoods with low poverty rates, three groups were defined. The first and experimental group received Section 8 vouchers which they were supposed to use in non-poverty neighborhoods. The second group received Section 8 vouchers as well but were not restricted to non-poverty neighborhoods in their choice of relocation and the third and last group was a control group who did not receive any voucher (Briggs, 1997; Rosenbaum et al., 2002; Katz et al., 2001). Contrary to Gautreaux no racial element was included in MTO. The only demand for the experimental group was they would relocate to a neighborhood with less than 10 per cent of its population under the poverty line. Although MTO in many ways was similar to Gautreaux, the results for participants were far less significant than for the Gautreaux movers (Kling et al., 2004). Many more researchers found disappointing results for MTO (Briggs and Turner, 2006; Briggs, 1997; Katz et al., 2001).

The differences of the outcomes between the two projects might be explained by the differences in design of the two programs. In a recent published book on outcomes of MTO Briggs et al. (2010) explain these differences by pointing at the fact that Gautreaux relocatees had to move to racially less segregated neighborhoods and they generally moved to neighborhoods that were located further away from their original location. MTO relocatees often remained fairly close to the original site and moved to neighborhoods with less poverty, but often with equal shares of African Americans, meaning they would remain resided in deeply segregated neighborhoods.

Critics of MTO and Gautreaux claim that while the programs have a sense of randomness regarding the selection of applicants, there is certainly a selection bias in these programs (Hedman and van Ham, 2012). Applicants to both these programs had to meet certain demands to be eligible for Section 8. This included income demands, but also criminal records were taken into account. Moreover applying for one of these programs means automatically that a household is willing to improve their situation and escape the disadvantaged ghetto. MTO has omitted this problem by assigning two experimental groups, the ‘normal’ Section 8 group and the MTO group which were obliged to use their voucher in a less than 10 per cent poverty neighborhood. However as Galster (2003: 905) explains “even the treatment group receiving intensive mobility counseling and assistance, although constrained to move initially to a neighborhood with less than 10 per cent poverty rates, nevertheless has the ability to choose neighborhoods varying on their school quality, home ownership rates, racial composition, local institutional resources, etc.”.

3.1.2: HOPE VI One of the most discussed relocation programs is HOPE VI. This program started in the 1990s with the objective to revitalize the most severely distressed public housing projects in the US (Popkin et al., 2002). At the end of the 1980s, public housing in the US was viewed as a complete failure. Although some project were faring relatively well, the most dilapidated projects had to deal with extreme economic and racial segregation, massive unemployment and high crime rates which led to dangerous living conditions, in terms of safety and health (Popkin et al., 2004). To overcome these problems in the deeply segregated public housing projects, in 1989 Congress appointed a Commission on Severely Distressed Public Housing, to identify the most distressed public housing projects and come up with a plan to revitalize these projects. As a result of the recommendations of the Commission, HUD initiated the HOPE VI program (Popkin et al., 2004). HOPE VI ran from 1993 and 2010, assigning approximately $6.3 billion allocated to 262 different projects throughout the US (HUD, 2010). A HOPE VI grant typically was used to demolish a severely distressed public housing project, replace it with a New Urbanist Style mixed housing community. The original tenants were either relocated with Section 8; they moved to other projects or could return to the original site. Since the projects were replaced by mixed types of housing only a portion of it would be public housing. In none of the projects was enough replacement housing to offer all tenants to return (Popkin et al., 2004).

Figure 3.1a HOPE VI before intervention

Figure 3.1b HOPE VI after intervention

Figure 3.1a and b show a HOPE VI development in San Francisco, CA before and after the intervention. Typically a high-rise modernistic structure was replaced with New Urbanist style mixed housing. In general New Urbanist style can be characterized by colorful housing, emphasis on ‘walkability’ and livability with attached housing (Schwartz, 2010). Photograph: SFHA (http://www.sfha.org/HOPE-VI.html)

HOPE VI was initiated not only to revitalize distressed projects in a physical manner, but had the explicit goal of “lessening concentration of the very poor and creating mixed neighbourhoods” (Salama, 1999: 97). Residents living in a public housing project to be addressed by HOPE VI were forced to leave their houses, but got the opportunity to return after revitalization, move to private rental housing by using vouchers or they moved to a different public housing project. Just as with the voluntary relocation programs the relocatees have been tracked in order to find out to what type of neighbourhood they moved. Most findings suggest that relocatees generally move to neighbourhoods with less unemployment and significantly less poverty (Popkin et al., 2002; Buron, 2004; Smith, 2002b; Comey and Popkin, 2004). However, relocatees, mostly African American or Hispanic, turned out to stay resided in deeply segregated areas with frequently over 90 per cent of the population being minorities (Clampet-Lundquist, 2004a). Those who managed to use a voucher that enabled them to rent at reduced cost in the private market reported most frequently an improvement in living condition, and generally moved to less segregated areas, while those who moved to other projects remained in neighborhoods with slightly less poverty than their original situation (Buron, 2004). However the process of relocation is often experienced as stressful, while choice for alternatives is limited. Strikingly there is not much research done in order to find out whether HOPE VI relocatees experienced any employment or educational gains. Clampet-Lundquist (2004b) conducted a qualitative survey to find out whether relocatees gained economically or socially by moving out. Strikingly she found no significant gains for relocatees two years after their move beside the fact that they had significantly less complaints about the quality of their new homes. In general most studies conclude that relocatees were more satisfied with their homes after their move (Comey and Popkin, 2004; Popkin et al., 2004; Popkin et al., 2009; Popkin et al., 2002; Clampet-Lundquist, 2004a; Buron, 2004). However, research also shows that relocated households had more difficulties building social ties in their new homes having difficulties integrating in their new neighborhoods (Clampet- Lundquist, 2010; Manzo et al., 2008).

3.2: The Netherlands In the Netherlands relocation policies to the benefit of social mixing have not been set up so explicitly as MTO or Gautreaux. More similar to HOPE VI the Dutch government under the name of the BCP has relocated tenants out of their homes due to demolition. There were no particular restrictions to where relocatees should move to. Although every city in the Netherlands

25 employ their own system of providing alternative housing to relocatees, most cities work by providing priority on waiting lists of social housing to forced relocatees. In this way those who are forced to move face a large set of alternative housing options ranging from similar dwellings close to their original homes to suburban single family homes. In order to find adequate alternative housing, relocatees receive assistance from their local housing association. In the end the housing association is legally responsible for providing adequate alternative housing to every single relocatee before they can start demolishing. Therefore demolition only takes place when every relocatee has found a new dwelling with enough rooms to house their family to a reasonable rent.

The wave of urban restructuring, initiated by the several BCPs was designed to replace dilapidated social housing with more mixed qualitative high standard housing. As was explained the BCPs main goals were social rather than physical. Yet the means of urban restructuring are physical to accomplish the social goal of safer and more livable communities through social mix. Strikingly apart from the discussion if social mix appears after mix of housing, research shows that through the dispersal of low income tenants from the demolished social housing, ethnic segregation is likely to increase (Bolt and Van Kempen, 2010; Bolt et al., 2009; Slob et al., 2008). Ethnic minority relocatees move to neighborhoods with sometimes even greater concentrations of ethnic minorities and low-incomes, but even more so do white relocatees avoid these neighborhoods (Bolt et al., 2008). Forced relocation after urban restructuring can lead to reinforcement of concentrations of low-income and ethnic minorities in other parts of the city. However they emphasize that forced relocatees only form a small part of the relocation flows into these new concentrations (Slob et al., 2008). Other researchers have investigated forced relocation not only to find macro effects of segregation but also looked for the displacement effects on individual relocatees. Up to now most research relied on sample and questionnaires to gain access to residents who experience relocation. Kleinhans (2003) was one of the first to investigate forced relocation in the Netherlands after urban restructuring. He concluded that most relocatees experienced improvement in their housing conditions after their move. However he adds that there is a potential danger in the system that relocatees often move to similar dwellings that are to be demolished soon after relocation. Therefore he warns for the emergence of urban restructuring ‘nomads’; moving from one demolition project to the other. Doff and Kleinhans (2011) confirmed that Dutch forced relocatees in most cases experienced serious improvement in

26 their housing situation after their move. However they stated that there is a difference in improvement among native Dutch relocatees and those belonging to ethnic minorities. Immigrant relocatees had significantly less chance to experience considerable improvement than their native Dutch neighbors. These findings are backed up by some other researchers, although they find that immigrants not necessarily move to qualitatively worse dwellings than native Dutch relocatees, but for sure they move to worse neighborhoods, in terms of employment average income and ethnic segregation (Bolt et al., 2009; Slob et al., 2008; Van Bergeijk et al., 2008). Moreover, Slob et al. (2008) found that ethnic minorities appear to be less satisfied with their new dwellings after forced relocation than native Dutch. Young relocatees however appear to be more satisfied after their move and are believed to be better capable of adapting to their new neighborhood (Visser et al., 2013)

Meerts et al. (2011) have tried, in a more qualitative approach, to find out to what extent choice was limited by institutional and structural constraints. They conducted 150 in-depth interviews with forced relocatees from Dutch urban restructuring and found that these relocatees can be categorized into two groups. The first group is fairly satisfied with their new dwelling after relocation, despite the fact that they are seriously constrained and not all their preferences are met. A second and much smaller group is less satisfied, and find themselves pressured by time and money constraints so they let go most of their own preferences and ended up in a less satisfying housing situation. However, a recently published study showed that involuntarily relocated households have more options to choose from in their relocation process than is often believed. They should be regarded as ‘active agents’ and are not necessarily victims of urban restructuring (Posthumus, 2013). An interesting conclusion of the same study is that immigrant relocatees and those with children are less satisfied after relocation and often remain in close proximity with the old neighborhood, in neighborhoods with relatively high unemployment rates, and a large portion of minorities (Posthumus, 2013).

3.3: Conclusion In general the whole process of urban restructuring and forced relocation does really resemble the classic examples of gentrification with the difference that it is induced by the government. Therefore the term ‘state-led gentrification’ appears to be appropriate. However gentrification is increasingly framed as a negative process where the emphasis is mostly on displacement.

Although forced relocation inevitably implies displacement, studies show that this type of displacement does not necessarily mean that the lower classes are “paying the highest tolls” as Newman and Wyly (2006) claim in case of classic gentrification. Overall in both the US as well as in the Netherlands relocatees move to better housing when they were forced to move due to urban restructuring or revitalization. However new concentrations may arise and relocatees sometimes have difficulties rebuilding social ties in their new neighborhood.

In a recent special issue of Housing Studies (2013, issue 2) on the subject of forced relocation Kleinhans and Kearns (2013) try to steer the discussion on this topic in a different direction. They find that “[o]ver time, the notion of displacement has increasingly acquired a negative loading in the discourse on residential moves that are triggered by gentriﬁcation”. They continue: “the tendency to frame forced relocation connected to state-led restructuring in a gentriﬁcation discourse tends to ignore or downplay fundamental differences between these phenomena” (Kleinhans and Kearns, 2013: 168). In the same issue of Housing Studies Kearns and Mason (2013) go even further by arguing that terms as ‘displacement’ and ‘forced’ relocation should be employed with more care and preferably be avoided. Preferences of relocatees are too often assumed and not investigated; moreover conditions for ‘forced’ relocations can vary widely in terms of information provided, institutional settings, and mechanisms of support and choice (2013: 197).

The research in this thesis will partly be based on the discussion raised in the special of Housing Studies. I investigated four demolition projects in Amsterdam demolished under the name of urban restructuring. I believe I can add substantially to the discussion outlined above for several reasons. First I was able to collect complete datasets of relocation processes including background information on relocatees and their exact new addresses. In contrast with all earlier studies in the Netherlands I do not rely on questionnaires with 30 per cent response rate but I have access to the outcomes of the entire population of four projects in Amsterdam. Hereby I avoid the sample selection bias other studies had to cope with. Households that do not cooperate in academic surveys might very well show different relocating behavior due to the unknown factor that is causing them not to participate in academic research. In addition the data I have used consist of a survey that is conducted prior to the relocation. Next to income ethnicity and family size this survey included neighborhood preferences of relocatees. These preferences prior

28 to relocation give a good view on the relocatees’ preferences not biased by the relocation process itself. These preferences can therefore be compared with actual outcomes to get a better look at the interaction between preferences and (institutional) constraints. All data are made available by the Amsterdam housing association Rochdale. For three months I have gathered data at the main office of Rochdale. This also entailed the opportunity to have daily discussion and conversations with the relocation staff of this particular housing association. In the methodology chapter I will elaborate more on the quality of the data. Moreover I will compare outcomes from forced relocation in Amsterdam with outcomes from forced relocation in three projects in Gainesville, Florida. The data from Florida are less detailed and will be used to understand the Dutch situation even better by comparing it internationally. In the next chapter on methods I will discuss the comparison in more detail.

4: Methodology In the special issue of Housing Studies on forced relocation Kearns and Mason (2013) question the appropriateness of the term displacement regarding those who were forced to move due to urban restructuring. Moreover Posthumus (2013) recently promoted at the University of Utrecht with a dissertation named “Displacement Myths”. It appears as though there is a shifting paradigm in discussing this particular topic. Whereas previous research framed this type of forced relocation as a merely negative consequence of questionable policies implying social mix, there are a growing number of studies emphasizing the merits of forced relocation and urban restructuring.

4.1: Questions Not inclined to favor any of these two explanations of the process, I want to add empirical evidence to this discussion with this thesis. The evidence in this thesis is different from previous research. To my knowledge the empirical comparison between a Dutch case and an American case has only been conducted once and has room for extension. The data for my Dutch case are very unique, in the sense that they consist of detailed information on a complete relocating population of certain projects, not dependent on response rates through questionnaires. Moreover these data contain information on preferences identified right before the process of relocation started and can therefore be compared to final outcomes. These advantages will make it possible to find out whether displacement, with all its negative connotations, occurs after forced relocation. To examine the concept of displacement I will repeat the four questions from the introduction:

1. What are the location preferences of households who are involved in large-scale restructuring of their neighborhood?

2. How do these preferences relate to the actual residential relocation due to restructuring?

3. To what extent are relocatees (not) improving their housing situation after forced relocation and how does that differ between relocatee types?

4. How do the outcomes of the relocation from Amsterdam differ from those in Gainesville, FL?

Displacement will be examined as a macro phenomenon or outcome, induced by the process of urban restructuring. Moreover this will help me answer the question: can we speak of displacement after forced relocation?

4.2: Macro-Micro-Macro Posthumus concludes in her dissertation relocatees or ‘displacees’ should be regarded as ‘active agents’ (Posthumus, 2013: 184-185). Even under highly constraining circumstances, tenants still have a considerable influence on the result of their relocation. This is in stark contrast with the ‘structuralist approach’ as described by Clapham and Kintrea (1984). The “structuralist approach assumes […] that the individual is dependent upon the local authority and plays a completely passive role in the allocation process [of subsidized housing]” (Clapham and Kintrea, 1984: 264). A structuralist explanatory narrative of displacement, that does not consider individual preferences and explains urban renewal from a gentrification perspective as a certain class- struggle therefore appears to be insufficient to explain forced relocation in the Netherlands. The methodology proposed in this study will follow methodological individualist’s model based on Coleman (1986) ‘macro-micro-macro model’. This model will fit more in the individualist approach, also described by Clapham and Kintrea (1984). By focusing more on the actions of individuals, their preferences and their typical circumstances, the relocatees are regarded as active agents. Coleman’s macro-micro-macro model is shown in figure 4.1, using the concepts of interest of this study.

Coleman distinguishes three different types of relations; type 1 relations are found within the domain of the micro, type 2 relations are structural influences on individual action and type 3 relations are structural outcomes of individual agency (see figure 4.1, for clarity reasons I have added type 4 relations, structural causes for structural outcomes on the macro level). If urban restructuring is framed as gentrification within the main discourse focusing on displacement and classes, the focus is mainly on type 4 relations. The only form of agency within this discourse appears to be the actions taken by the (neoliberal) government, representing the middle-class in order for them to regain possession of the inner cities (Lees, 2008; Smith, 2002a; Smith, 1996; Uitermark et al., 2007).

Macro Level Concentrated Mixed (4) Poverty/Ethnic Neighborhoods Segregation

(2) (3)

Micro Level

Order to (1) New Location Relocate

Figure 4.1: Figure 4.1 shows Coleman’s (1986) model of macro-micro-macro relations edited to fit the situation of forced relocation. It is the yellow lined box of micro relation that will be of most interest in this study

This can be categorized under what Coleman calls “"conspiracy theory of history," in which macro-social outcomes are the intentional result of calculations on the part of some subset of actors, rather than the emergent (and often unintended) consequence of interactions among actors with numerous differing purposes”(Coleman, 1986: 1323). Other scholars have focused more on relations of type 2 and type 1 and type 3. These studies have mainly emphasized that relocatees of different backgrounds react differently on urban restructuring (Meerts et al., 2011; Posthumus, 2013: etc.; Doff and Kleinhans, 2011; Kleinhans, 2003; Kleinhans et al., 2007). In this thesis I will also investigate relations of type 1, 2 and 3. With all the advantages I have mentioned above, I hope to tease out some more structural consequences of urban restructuring by looking at the individual preferences and actions. With urban restructuring the relocation, which is a process that happens on the individual level, of all residents from one project taken all together can be categorized as displacement. Displacement is a process that happens at a larger scale at the macro level. The macro-level process of displacement is investigated by looking at micro level

32 actions of relocation. This means that my (macro) displacement question will be answered by answering (micro) questions 1 to 4.

4.3: Conceptual Model Relocation processes, either forced or voluntary, can be explained using three different approaches (Clapham and Kintrea, 1984). Relocation can be explained by looking at institutions and their effects. This approach focuses on existing policy regarding relocation and the possible effects that housing employees can have on the outcome of a relocation process. Especially in the public or social sector the influence of institutions should not be underestimated since relocatees often lack the means to fulfill all their preferences and are dependent on what the system has to offer. The next approach is a structuralist approach. Relocation patterns are explained by mainly looking at class structures. Very often the emergence of residential segregation is explained by pointing at perpetuating class-structures, and ethnic or socio-economic residential segregation. Gentrification theories on urban restructuring fit this category. In a third and more actor-based individualistic approach the emphasis is placed on individual preferences of tenants that result in the eventual relocation.

Since I want to investigate the relocation processes more on a micro level, especially the first and third approach to relocation will apply to this research. These approaches are both nested in my research design. In figure 4.2 I have elaborated these approaches in a conceptual model. This model is simultaneously an elaboration of the type 1 relation shown in figure 4.1.

The conceptual model in figure 4.2 represents the relocation of an individual relocatee. This model is initiated at the moment a relocatee learns about the demolition plans. He will start looking for alternative housing and eventually find improvement (or not) in this alternative housing after he made his relocation. A move from one apartment to another is a result from a consideration made by each relocatee trying to meet his or her preferences within the given institutional constraints. Then again this consideration is influenced by the information known to the relocatee. Assistance by the authority and characteristics of these relocatees influence how much information is known to them (Clapham and Kintrea, 1984). Eventually the differences that for instance other studies have found between different types of relocatees should be explained by either the amount of knowledge, different preferences or institutional constraints.

Neighborhood Relocatee Improvement

Characteristics

Institutional

Constraints Preferences

Relocation

Information

Housing Assistance Improvement

Figure 4.2: the above figure shows the conceptualized type 1 relation shown in figure 4.1. The blue boxes represent the explanatory variables, which result in relocation after they have gone through the institutional constraints of the housing market and rules of relocation. Eventually this would lead to an improvement (Or not) of the situation of relocatees.

4.4: Data To answer my research question whether displacement occurs after urban restructuring, I have conducted a comparative case study, studying four cases in Amsterdam and three cases in Gainesville, FL. Due to availability of data the main focus will be however on Amsterdam. The comparison with Gainesville will therefore function as an extra case to illuminate the importance of the institutional context. As was explained in chapter 2 and 3, forced relocation in the US is carried out by Section 8 vouchers while in the Netherlands most municipalities work with priorities on social housing waiting lists.

The data used in the American case are much less detailed than the data in the Amsterdam case. However I will combine the data on relocation in Gainesville with the publicly available data of the US Census and the American Community Survey. These data contain among others information on the geography of poverty, racial minorities, rental vacancies, income, unemployment and much more. In addition I have interviewed several informants about the

34 process of relocation. Moreover there has been ample qualitative research into forced Section 8 movers. I will include an assessment of the qualitative research into Section 8 movers, to help interpret the relocation outcomes in Gainesville. The dataset that is available is derived from the Shimberg Center database on affordable housing in Florida. With their permission I am working with a database containing the new locations of Section 8 relocatees from three projects in Gainesville, Florida. Moreover this dataset gives a ‘picture’ of the population of these projects with the average income, percentage minorities etc.

The Amsterdam case has a much more analytical disposition. Other studies have shown that classic displacement after urban restructuring is not as self-evident as was sometimes assumed. The data used in Amsterdam are very detailed and includes the preferences of relocatees as well as all their demographic characteristics. All information is gathered from the archives of one housing association in Amsterdam, Rochdale. I will investigate four different projects in Amsterdam where the relocation was carried out by Rochdale. The data used consist of information on the old dwelling, information on the new dwelling, information on the relocation process and a survey of the residents, containing their demographic information, income and neighborhood preferences. This survey was carried out by Rochdale, interviewing every resident in the project. Since I was an employee of Rochdale in 2007 and 2008 myself, I was able to conduct several interviews myself. The advantage of this survey is that without completing the survey, residents could not apply for priority on the social housing market or enjoy assistance in relocation from the housing association. Therefore the survey has a response rate of 95%. The data are more explored in the Operationalization section below.

The comparative case studies are both very quantitative in nature. With quantitative data I will try to examine the concept of displacement. Displacement has been described a lot by many authors, but a thorough quantitative analysis with complete datasets is yet so far not available to my knowledge. Therefore this quantitative effort has the potential to substantially contribute to the academic discussion regarding displacement.

4.4.1: Population The research question is based on displacement after urban restructuring in post-war housing estates. Therefore I have selected four demolished projects due to urban restructuring in post-war neighborhoods that involved forced relocation. Generalizing inferences on these four cases is

35 however problematic. The inferences cannot be applied to every single other urban restructuring project in the Netherlands let alone in Europe. As was explained in chapter 2 and 3, urban restructuring took place mostly in early post-war neighborhoods, built largely between 1950 and 1975. However relocation processes seem to differ between different cities in the Netherlands. Relocatees in Rotterdam move to more segregated neighborhoods than relocatees in for instance Breda or Ede (Posthumus et al., 2013a; Posthumus et al., 2013b); or relocatees from The Hague that move very often outside of the city compared to relocatees from Utrecht and Leiden due to allocation systems (Bolt and Van Kempen, 2010). It is therefore not my intention to generalize outcomes of this study to other cities in the Netherlands. However I do want to make outcomes generalizable to other post-war areas in Amsterdam. Most post-war areas in Amsterdam are either located in Amsterdam Nieuw West or in Amsterdam Zuid-Oost. There is however a difference of night and day between these two areas. Where Zuid-Oost has a huge concentration of Surinamese and African minorities, in Nieuw West most ethnic minorities are of Turkish or Moroccan background. To emphasize the difference in relocation patterns between the two districts I have included a very general study of relocation from Zuid-Oost. The projects in ANW are chosen based on their population and neighborhood being representative for the whole region of ANW. In so doing I might not be able to generalize to the rest of the Netherlands, but the testing of the hypothesis around displacement is still valuable. If large scale displacement is found in Amsterdam, is does put some question marks behind conclusions of other scholars implying that ‘displacement’ is a bit overstated in most literature. However the opposite empowers these conclusions and shows with more accurate data that displacement after urban restructuring is indeed a myth (Posthumus, 2013).

Therefore the population of this research consists of all demolition projects in Amsterdam Nieuw West due to urban restructuring. The sample consists of four cases/projects in Amsterdam Nieuw West and the observations consist of 572 households being relocated.

4.5: Operationalization The main concept I will be looking for in this thesis is the more macro concept of displacement. Displacement has become the most significant topic in gentrification and/or urban restructuring literature (Kleinhans and Kearns, 2013), however a clear definition of the concept often lacks. The increasingly negative connotation of the concept implies that displacement it is a negative

36 consequence for lower class residents of a change in population of a neighborhood. Since displacement has received so much negative connotation, equating it with the only possible result of a forced relocation appears simplistic. It would ignore the possibility that relocatees use the right that comes with forced relocation in order to make significant housing career. A household that moves to a dwelling of significant higher quality within the neighborhood of his preference, for a reasonable price and within a reasonable duration of time should not be considered displaced with all its negative connotations. Kleinhans and Kearns (2013: 168) state that “[t]he common premise is that low-income groups suffer most from displacement, in terms of the quality of their dwelling, increased rents and fewer housing opportunities in general”. However the forced character of moves also entails that people have not asked for it, and should be compensated. If relocatees are not compensated well enough or move to the dwelling that does not match their ‘reasonable’ preferences, relocatees can be regarded as victims of the relocation. If relocatees are considered victims, displacement with its entire negative connotation is present. A definition of displacement should therefore include the results of the relocation process and the process itself. I will adopt the following definition of displacement:

Displacement occurs when after forced relocation (a) certain types(s) of relocatees

- were forced to leave their home AND - did not find improvement of housing quality OR - moved to a neighborhood outside of their preference OR - are cost burdened due to their move, meaning they are paying more than 30% of their income on rent (McClure, 2005) OR - did not have enough housing opportunities to choose from.

If one type of relocatees, based on their characteristics, significantly suffers from at least one of the above mentioned criteria, this type of relocatee will be regarded as displaced. The degree of displacement is dependent on the percentage of that type of relocatee that suffers from at least one of the criteria.

During the proposal phase of this research I planned to operationalize the other concepts used in figure 4.2 in the same way for both the American case as well as for the Dutch case. However the available data differs greatly between both cases that some adjustments have been made to fit

37 each case better. The data for the US case did not give any information on individual characteristics of relocatees nor did it reveal individual information on housing quality of alternative dwellings. The information that is available is based on the characteristics of the new neighborhoods and qualitative information in the process of relocation. Therefore the concepts of relocatee characteristics, preferences and housing improvement will only be used for the Dutch case.

4.5.1: Relocatee Characteristics All data on relocatee characteristics is derived from a survey that was conducted by the housing association that was responsible for the relocation of the tenants of the four Dutch cases. In order for relocatees to receive assistance from the housing association and a priority on social housing waiting lists they had to participate in this survey. This survey includes the annual income, family size and duration of occupancy of original dwelling. Not all residents however participated in the survey since they rejected assistance from the housing association and decided to move outside the system; for instance by moving to owner occupation or private rent. In a few cases the administration of the housing association did not provide adequate information or some information got lost.

Project Total number of households Complete Survey Percentage Bakema Sloop Noord 98 86 88% Van Tijenbuurt 177 167 94% Zuidelijk Veld Fase 1 193 167 87% Zuiderlijk Veld Fase 2 104 88 85% Total 572 508 89%

Table 4.1: table 1 shows the number of observation (relocatees) used in the Dutch case. Almost 90% of all relocatees in the four projects are used in this thesis. Of slightly more than 10% the data got either lost or did not participate in the relocation process

To define different types of relocatees based on their characteristics three variables are used: family size, ethnicity and income. These three variables are considered to be the most influential on relocation (Clapham and Kintrea, 1984). The three variables are used dichotomously to create

38 eight different types of relocatees (23=8). Ethnicity is dichotomized using the categories ethnic minority or native Dutch; incomes are divided by lower or higher than a gross annual income of €28.000 (a rough estimate for the area median income) and family size are categorized as large (two or more children) or small (one ore no children).

4.5.2: Preferences Just like relocatee characteristics this concept will only be dealt with in the Dutch case. Similarly preferences were part of the housing association survey. Relocatees could identify neighborhood preferences and type of dwelling preferences in the housing association’s survey. In order to assist residents in finding alternative housing tenants had to identify a number of neighborhoods to which they desired to relocate and certain types of dwellings. For simplicity’s sake preferences are solely operationalized by neighborhood preferences. Although type of dwelling preferences are interesting, they are difficult to test against actual relocations since type of dwelling of addresses are difficult to trace. The association prepared a list of districts in Amsterdam which could be checked by residents. Since the waiting list priority is effective in surrounding suburbs as well, tenants could check those suburbs as well. The housing association used former ‘stadsdelen’ as districts. Tenants had the opportunity to check an unlimited number of districts. 38 households who did participate in the survey did not check any district and are therefore regarded as indifferent for new neighborhood. Most households identified one, two or three districts to which they desired to relocate. These preferences are the least possible influenced by constraints such as availability of housing within each district since they were identified prior to the process of looking for housing.

4.5.3: Information and Assistance The information that is known to relocatees prior to their relocation is essential to where they will move to. Residents are believed to have ‘mental map’ of the city and apply this mental map when looking for alternative housing (Donaldson and Johnston, 1973). These mental maps are often sector-based, meaning that the knowledge is greatest within one’s own sector of the city (Hoyt, 1943). Moreover the amount of knowledge on the city might differ between different type of relocatees due to levels of education or experience in other parts of the city. However a qualitative assessment through questionnaires on the mental maps of relocatees is not part of this study. Although structural differences might exist between different types of relocatees I will focus on the assistance that is provided by either the housing association or the housing

39 authority. With a qualitative assessment on the relocation process and the role of the HA will reveal how much assistance was provided to relocatees in order for them to find alternative housing. Especially the comparison between the US case and the Dutch case is of interest in this case.

4.5.4: Institutional Constraints Similar to Information and Assistance I will investigate institutional constraints by assessing the relocation process. For both the US case as well as for the Dutch case I will give a detailed overview of what is offered to relocatees and what right they have in order to find alternative housing. In this respect also especially the comparison between the two different institutional contexts is interesting. American relocatees in Gainesville will have different institutional constraints than relocatees from Amsterdam.

4.5.5: Relocation Relocation is the moment when a relocatee signs the rent contract for his new dwelling. In the Dutch case the exact date of this event is known. This means that it is possible to measure how long it took for relocatees to actually find alternative housing. Some relocatees found alternative housing within a couple of months after the survey of the housing association while others sometimes took more than three years to accomplish this. Those who relocate within a year are generally regarded as those who were waiting for the restructuring to begin. Those who took longer are the ones that did not anticipate on the restructuring process2. During the process the stress for relocatees increases with every day the planed demolition comes closer. However this goes for the housing association as well since they are not allowed to start demolition when not every household is relocated. By law the housing association is responsible to offer ‘reasonable’ and ‘adequate’ alternatives to the relocatees. When tenants and the housing authority cannot agree on the terms of adequacy and reason, a judge must decide whether the relocatee has to accept a move or whether a housing association should provide better alternatives. Since both parties generally benefit from a smooth process this happens only in exceptional circumstances, on average once or twice per project.

2 This information is derived from interviews with relocation assistants from the housing association

4.5.6: Housing Improvement Only for the Dutch case some information is available on the characteristics of the new housing. Housing quality is difficult to measure since data on the physical condition of individual dwellings are scarce. I was able to approximate the physical quality of dwellings by looking at two quantitative variables: the size of a dwelling the year it has been built. This information will be compared to the same variable of the old dwelling to find improvement or not. Generally speaking larger dwellings are considered of higher quality than smaller dwellings. Building year however is not as linear. Often older houses are regarded as of lower quality than newer ones; in Amsterdam this assumption is however false. In the period after WWII, many houses of relatively low quality were built speedy in order to comply with the housing shortage. The Amsterdam socially rented housing stock that was built between 1945 and 1975 is believed to be of considerably lower quality than housing of other building years (Rowlands et al., 2009). Since the original site was built in this post-war period, moving to a dwelling of another building year can be interpreted as an improvement of housing quality while staying in housing built in the same period is not. However, only the housing qualities of those who did not leave the public housing sector in Amsterdam were measured. For dwelling size, those who moved to a dwelling that is at least five square meters larger than their previous apartment, will be regarded as having had an improvement in dwelling size.

4.5.7: Neighborhood Quality For the Dutch case I will measure neighborhood quality using two different approaches. On the one hand neighborhood characteristics can indicate whether a neighborhood can be considered a ‘good’ or a ‘not so good’ neighborhood especially regarding the social mix goals. Within the discourse of social mix, good neighborhoods are those with low levels of poverty and no over representation of minorities. For data on neighborhood quality I will use data from Statistics Amsterdam on neighborhood level. A typical neighborhood consists of approximately 8.000 inhabitants. Since the relocation process of the four projects ran from 2006 to 2010 I will use data from the year of 2010 to measure neighborhood statistics of the new location. Using these two variables I will be able to find out to what type of neighborhood different types of relocatees have moved. But next to these statistical measures I will use a more subjective measure comparing the actual location with the preferences given before relocations. This is in accordance with the definition provided for displacement.

Since I do not have any information on neighborhood preferences in the US I will rely on pure statistical data. To discover patterns in the displacement of the Gainesville tenants, I will try to find the driving force behind relocation choices. I will investigate the geography of the Gainesville housing market, and the geographical distribution of different types of residents to find out to what type of neighborhoods relocatees move to. Again this analysis will be very descriptive in nature. For neighborhoods I will use Census Tracts which are comparable in size with the Amsterdam neighborhoods. Since the relocation process took place between 2003 and 2009 I will use data of the US Census 2010.

5. Gainesville

5.1 Introduction During my work at Delta Forte BV, a daughter company of Rochdale responsible for the relocation of tenants of many demolition projects I received many complaints from residents from Bakema Sloop Noord (one of the four projects in the Dutch case) about deficiencies of their to be demolished dwellings. The complaints ranged from broken mailboxes to nuisances about children playing soccer on the streets in front of them. Most claims were considered serious harms to the residents’ ‘woongenot’ which can be translated as living enjoyment. If this was the case, complaints were many times resolved by the staff of Rochdale. These complaints however are in stark contrast with those from resident of for example the Kennedy Homes, one of the three projects under study in Gainesville:

“One of my daughters contracted a disease from being bitten on her lip by a rat. I sat up many a night shootin’ rats with a pellet gun. I used a flashlight. I killed 25 one night. When I’d shine a light on them, they’d freeze. I put them in a box and showed it to the maintenance men, but they didn’t do nothin’ about it.” (Kelley, 2009: 19)

“Kennedy Homes was nicknamed the ‘Hell hole’ and that wasn’t far from the truth. Although we were there only two years, I made so many complaints, but now they tell me they don’t have any copies of any of that. Said they lost all the records.” (Kelley, 2009: 20)

“As far as the roaches problem is concerned, whatever you put down, you had to share with the roaches, they could care less if people were there or not. You could spray and you can’t kill them all, but more would be back the next day. There were big ones, medium sized ones and little ones, and they could fly, too. At night, the walls would be covered with them. I constantly sprayed, cleaned up, sprayed everything, trying to get rid of them every night.” (Kelley, 2009: 21)

“When the stove exploded, I went outside, thinking someone was shooting. My neighbor said ‘it’s your house!’ I went back in and saw a hole in the ceiling. The lines to the gas pipe was all mangled. There was grease all over the walls and ceiling. Dishes were all over the place. It was a huge explosion. I immediately called them, cut the gas off and they said we’ll come back in the morning. I said I still smell gas. They wouldn’t do anything about it, so I cut the gas off myself. I started keeping the window up in the kitchen due to the gas leak and that helped make it better.” (Kelley, 2009: 20)

“There was many, many times when I’d come home after workin’ all day and the apartment would smell from the feces backin’ up. It made me sick to my stomach. I’d clean it up and pour bleach and chemicals in there, but then there was the gas leak, too. When you inhale gas too long, what happens?” (Kelley, 2009: 21)

All above quotes originate from the book “The Kennedy Homes: and American Tragedy” by Pierce Kelley (2009). Kelley was involved in the relocation process of this particular project in Gainesville as an attorney for the relocated residents. The quotes are shown as quotes from residents in this book. Although Kelley admits that his writing is rather one sided, viewing the relocation process from the residents’ viewpoint, they illustrate the huge differences between conditions in public housing in the US and the Netherlands. Next to The Kennedy Homes, which was abandoned December 2003, and demolished in 2008, two other projects that were demolished or completely refurbished were investigated: tenants from 153 units in Glen Spring Manor and another 54 from Seminary Lane were relocated during the months prior to demolition throughout Gainesville. Both projects were closed due to poor living conditions on site (Rolland, 2009). The three project were 100% occupied by subsidized tenants, however the projects were not classic public housing projects. All three were owned privately, rented to residents with Section 8. Therefore the demolition and reburbishing of the projects was not carried out with a HOPE VI grant. Nevertheless tenants were (if eligible) relocated with a new Section 8 vouchher. The only difference with HOPE VI projects is that no New Urbanist style community replaced the old site, which means that the option to return was not available for relocatees from Gainesville. Figure 5.1 shows the location of the three projects in the city of Gainesville.

In this chapter I will start with a brief description of the city of Gainesville and the three projects. This will include a short history, an overview of the residents and some qualitative information generated through interviews with a local housing authority employee, some locals that happened to be in the neighborhood of the projects during my investigation (including a local drunk, a bartender, a preacher from the local church and a project developer) and with Pierce Kelley, the author of the above mentioned book on the Kennedy Homes. After that the relocation process will be discussed, including a brief assesment of the local housing market and some more information on Section 8 vouchers.

Figure 5.1: this map shows the location of the three projects under study.

5.2 General Characterization Gainesville The city of Gainesville is located in central Florida, 70 miles south west of Jacksonville. Most Americans know Gainesville as the hometown to the University of Florida (UF), best known for its football team the Florida Gators, national champions of 1996, 2006 and 2008. Aside from a big stadium and an even larger campus, this means that Gainesville with its slightly less than 130.000 residents can genuinely be called a college-town. UF alone has over 50.000 students, meaning that more than a third of the population is college-student while roughly one quarter of the entire population is aged between 20 and 24. Next to the Florida Gators, Gainesville was the hometown of musician Tom Petty (and the Heartbreakers).

Race Number Gainesville US White alone 83,194 66.9% 74.0% Black or African American alone 27,646 22.2% 12.5% American Indian and Alaska Native alone 276 0.2% 0.8% Asian alone 8,273 6.7% 4.7% Native Hawaiian and Other Pacific Islander alone 134 0.1% 0.2% Some other race alone 1,901 1.5% 5.5% Two or more races: 2,847 2.3% 2.4% Two races including Some other race 368 0.3% 0.5% Two races excluding Some other race, and three or more 2,479 2.0% 1.9% Total:races 124,271 100.0% 100.0%

Table 5.1: Racial composition Gainesville, Florida Data Source: US Census 2010

5.2.1 Racial segregation Compared to other Floridian cities such as Jacksonville, Miami, Tampa or Tallahassee, Gainesville has relatively few ethnic or racial minorities. The population of Gainesville is (partly due to the larger share of students) predominantly white with slightly more than 22 per cent black or African American. Other racial or ethnic groups are significantly less represented in Gainesville. The African American and white populations live fairly segregated in Gainesville, The dissimilarity index3 for the black or African American population vs. the white population is 0.41.

Figure 5.2 shows the vast majority of the minorities (i.e. African Americans or Blacks) lives in the Eastern part of town. The two South Eastern tracts consist of even more than 75% minorities. The figures of poverty appear to show strikingly similar patterns with racial concentrations. To exclude a bias of poverty figures generated by students who often have little money to spend but can hardly be considered poor, the poverty shown in figure 5.2b only measures the percentage of

∑ |( ) ( )|

families with children within the census tract that has lived under the poverty line in 2010. These figures are estimates of the American Community Survey, which is based on a sample and has a confidence interval of 90%. Although only families with children are taken into account, there still is a concentration of poverty around UF campus in the South of Gainesville, the area which mostly consists of student dorms and fraternity or sorority houses. Two explanations are possible: either many students have children (which is more often the case in the US than in the Netherlands) or poor families are living among the students in student dorms. All in all both poverty and minorities concentrate in the Eastern section of Gainesville.

Figure 5.2: Distribution of poverty and racial minorities Gainesville, Florida

Data Source: US Census 2010; ACS 2010

5.3: Description of the three projects In this paragraph I will discuss the original location of the three projects in this study, Glen Spring Manor, Seminary Lane and the Kennedy Homes. None of the three were actual public housing projects, they were all privately owned, but built with tax credits or other federal money so in return they would be exploited in the affordable sector. This type of subsidized housing falls under the Section 8 project-based programs.

5.3.1: Glen Springs Manor Glen Springs Manor is the only project of the three that did not get demolished but is being refurbished. The project is located in the north-west of Gainesville. The 136 apartments, built in townhouses of two stories, are accessible via two main entrance ways and otherwise cut off of road network of Gainesville.

Figure 5.3: Satellite image of Glen Springs Manor

Due to increasing criminality and drug abuse HUD issued a grant of around $150.000 in 2004 to surround the complex with an iron fence and keep undesirable guests out. Glen Spring Manor (GSM) was until 2005 owned by a private company, named Bates Realty and was acquired in 2005 by the non-profit MacMillan and Haven Economic Development. It was a Section 8 project-based subsidized project. This means that the owner received direct subsidy from HUD in addition to the tenants rent payments. Moreover the owner acquired tax-credits by exploiting affordable housing complexes. HUD does a yearly appraisal of all subsidized property to see if the owners are maintaining standard quality of housing. In 2002 GSM received 85 out of 100 points, which can be considered as more than sufficient. However, problems were accumulating in the apartment complex and seven years later GSM received only 39 points which is more than 20 point below the minimum of 60. Next to criminal activity, there was mold on the ceiling, inoperable air conditioners and doors that would not lock. Figure 5.4 shows an apartment at GSM just before the closing was announced end of fall 2009.

Figure 5.4: A picture of Glen Springs Manor months before closing

Picture: Tricia Coyne, Gainesville Sun

In November 2009 residents were informed about the planned closing of the apartment complex. HUD had withdrawn its subsidy and used the money instead to assist residents in relocation. By the end of May 2010 all tenants should have been relocated. Residents that had enough financial means, no criminal record and were official residents received a Section 8 voucher to find housing in either another assisted housing property or use the Section 8 voucher in the private

rental market. After the last tenants moved out in June 2010 another developer started refurbishing the apartments in order to rent or sell them in the private market without HUD subsidies.

The population of Seminary Lane can be considered the poorest of the poor. Table 5.3 shows that GSM was populated mainly by minority female headed households earning less than 30% of the area median income. Compared to the other two projects the share of minorities was relatively low. From the 71% minorities more than 90% was African American. From the 121 households that were residing at GSM in 2008, 92 relocated using their Section 8 voucher. The other 29 households were either not eligible for Section 8 or moved by themselves. It is unknown where they have moved to and if they were able to find alternative housing. The relocation of the 92 households will be discussed later.

Figure 5.5: Impression of Glen Springs Manor

The pictures below show what Glen Springs Manor looks like in January 2013. The structures are still there but are being refurbished to be rented out and sold in the private sector.

Pictures by author

5.3.2: Seminary Lane Seminary Lane was a complex of duplexes a few blocks away from the UF campus, consisting of 53 dwellings. The complex was owned by the non-profit organization Gainesville Florida Housing Corporations Inc. but was managed and operated by the Gainesville Housing Authority. Similar to Glen Springs Manor, Seminary Lane was a privately owned project based Section 8 housing complex. In 2008 Seminary Lane was appraised by HUD scoring 62 points, 2 more than the critical quality threshold of 60. However the Housing Corporation decided to close the apartments mainly due to monetary reasons. By the end of May 2009 all residents would have to be relocated using the tenant based Section 8 voucher all 35 households received to relocate. After that all apartments were demolished. Yet so far there has not been built anything replacing the apartments.

Figure 5.6: Satellite image of Seminary Lane

Unlike Glen Springs Manor, Seminary Lane was not closed in any way from other parts of town. The apartments were built close to the town center with wide spaces in between the houses and room for green spaces. The occupants of Seminary Lane were not as poor as those from GSM, with ‘only’ 76% earning less than 30% of the area median income. 96% of the population was African American and the mean income was slightly less than an annual $9.000. In accordance the average rent that was paid by the resident was higher than the rents in GSM. Figure 5.7 shows the current (as of January 2013) condition of the site. The lots are vacated and apparently there are no actual plans to rebuild something on site.

Figure 5.7: These pictures show the current conditions of the original site of Seminary Lane. The houses have been demolished, yet no new construction has been built Picture by author

Income group (annual) Number of Rental Range Number of Number of Total Shortage/ Vacancy Rate Households Occupied Vacant-for- Number of Surplus Rentals rent Units

< $9,999 8439 < $250 1047 88 1135 -7304 7.8% $10,000 to $19,999 5983 $250 to $499 4500 490 4990 -993 9.8% $20,000 to $34,999 6393 $500 to $874 14610 1751 16361 9968 10.7% $35,000 to $49,999 3,930 $875 to $1,249 5472 836 6308 2378 13.3% $50,000 to $74,999 2,973 $1,250 to $1,874 1774 276 2050 -923 13.5% $75,000 + 1680 > $1,875 1110 74 1184 -496 6.3%

Table 5.2: Housing Needs Assessment Gainesville

This table shows different price classes of all rental dwellings in Gainesville and shows the number of renters within each income group. The income groups correspond with the price classes of rental dwellings using 30% of income for rent. This method is derived fromMcClure (2005).

Source: ACS 2010

Units Occupation People Average Mean % Extremely % Low % Female % Minority Per Unit Monthly Rent Income Low Income Income Headed Glen Spring Manor 136 89% 2.8 $135 $5.600 90% 10% 95% 71% Seminary Lane 53 80% 2.4 $235 $8.900 76% 20% 96% 96% Kennedy Homes 172 90% 1.7 $109 $6000 83% 14% 93% 97%

Table 5.3: Picture of Population of Projects

This table shows the general characteristics of the population of the three projects under study.

5.3.3: Kennedy Homes The Kennedy Homes are a somewhat different story than the two other projects. Unlike GSM and Seminary Lane, Kennedy Homes was owned by a for-profit owner AIMCO. As was expressed in the introduction of this chapter, the living conditions at the Kennedy Homes were far from ideal. Tenants were sharing their apartments with roaches and rats, poisonous paint was peeling of the walls as gas pipes were leaking gas in the attics of the apartments. The population of the apartments was almost entirely African American and belonging to the extremely low incomes, earning less than 30% of the area median income. Although AIMCO was a for-profit company, it leased out the apartments in a similar fashion as the other two complexes were leased out: via complex-based Section 8 subsidies from HUD.

The worsening conditions at the complex eventually led to a fire, destroying two entire sections of the complex in October 2003. No one was injured, but the fire caught the attention of HUD that started inspection at the Kennedy Homes. Both the City of Gainesville as well as HUD demanded that serious refurbishment was made by the owner in order to make the project livable. In the end the owner refused, meaning that residents had to move out immediately. By mid-December 2003, all remaining residents were relocated to hotels and motels in the surrounding area so the property could be closed down. The tenants did not have to stay in the hotels, but got a Section 8 voucher to find alternative housing.

As is shown in figure 5.8 the Kennedy Homes are demolished, and yet again there is nothing built in replacement. Although there is not much to see on the original side some picture are included figure 5.9

Figure 5.8: Site of Kennedy Homes

Source: Google Maps 2013

Figure 5.9 Kennedy Homes

These pictures show the current conditions if the Kennedy Homes Pictures by author

5.4 Outcomes The data used to investigate the Gainesville-case consist of the mere census tracts relocatees have moved to. As was described in the above descriptions of the project, in none of the three projects the relocatees had ample time to search for alternative housing. In GSM and Seminary Lane the relocatees had six months to find alternatives while the relocatees in the Kennedy Homes had to look for alternative housing from small hotel or motel rooms. Moreover literature shows that moving with Section 8 vouchers often means a small improvement in housing quality but does not involve a lot of relocatee preferences. In this section I will briefly review the existing literature on the process of housing search with a Section 8 voucher to get a better qualitative understanding of this process.

5.4.1: Section 8 The term Section 8 has been used a lot in this thesis and to understand the relocation process it is vital that all is clear on Section 8. The three projects under study are financed with project-based Section 8. A project based Section 8 project is privately owned. Tenants pay a maximum of 30% of their adjusted income to rent, the difference between this and the approved rent for adequate housing is paid by HUD directly to the landlord (HUD, 2006). HUD is no longer providing new subsidies for project-based Section 8 projects. Therefore when these are demolished or closed, tenants receive a “tenant-based Section 8 voucher”. With this voucher a tenant can receive the same type of subsidy, but not in a dedicated project but in the general private market (Schwartz, 2010). The Section 8 subsidy only pays for dwellings of which the rent is lower than 80% of the standard fair market rent as set by the local housing authority. In Gainesville the fair market rent for a 2 bedroom apartment in 2009 was $751. 80% of that means that the maximum rent of a dwelling under Section 8 voucher program can be up to $600. In Gainesville there are many dwellings available that fit this criterion, as is shown in table 5.2.

Vouchers are in theory an ideal solution for public housing. An employee from the Gainesville Housing Authority stated that “the voucher allows you to live where you want to, so if you have a child that you want to go to a certain school, you can move to that neighborhood”. Since in these three cases HUD will pay for all fees and security deposits, there were no extra costs for the residents moving to another apartment. The relocation process therefore seems an ideal opportunity to leave dilapidated project for a nicer home in a nicer neighborhood with less poverty and crime. The voucher offers the relocatee the opportunity to find a house that meets

58 his or her preferences and demands. In a good overview of the literature concerning forced movers, Joseph and Chaskin (2012) find that the preferences of public housing relocatees are driven by attachment to both place and neighbors. Next to these important values for relocatees safety and schooling are often mentioned by Section 8 relocatees as important to their new location (Clampet-Lundquist, 2004b; Smith, 2002b). However, moving with a voucher does not mean that any preference can be fulfilled. Relocatees face many constraints, ranging from financial boundaries, to unwilling landlords. First of all families living in subsidized housing often rely on public transportation (Brooks et al., 2005). Since public transportation is only served roughly within the city-limits of Gainesville, more suburban areas are often off-limits for the relocatees. Some authors also refer to the influence of housing authority staff. Housing authorities often have incentives to speed up the process. They put pressure on the tenants to relocate rather today than tomorrow. As a consequence of time pressure, relocatees find it more important to accept a home than to find the best option from all opportunities (Comey, 2007; Clampet-Lundquist, 2004b; Smith, 2002b).

Section 8 voucher holders also face the stigma of troublesome residents, decreasing the value of complexes when moving in. Landlords are in general not very willing to accept Section 8 renters. Since it is a voluntary program, Landlords keep the right to reject a Section 8 tenant into his property. Although it is prohibited by the Fair Housing Act to reject tenants on the basis of income, race or gender, Section 8 tenants are often rejected. Landlords say they do not want any federal involvement in their business or are concerned about the continuation of payments and evictions; however some landlords bluntly state they do not want any poor resident in their property (Culbreath and Wilkinson, 1999). Although this being against the law, tenants often do not know their rights and do not want the hassle of filing an official complaint. Meaning that denial of Section 8 vouchers is the order of the day (Popkin, 2000). Secondly, there are many misconceptions about Section 8 among residents. Some believe they only last for a year, others see them as part of a conspiracy to drive all blacks out of the city and again others simply do not understand what a voucher can do for them despite all provided information (Popkin, 2000). The relocation can be considered a limited choice within a very limited amount of time. Often not all preferences can be met with the relocation.

5.4.2: Local Housing Market When providing housing subsidies, HUD uses an Area Median Income (HAMI) to qualify households. Table 5.4 shows the number of households within each HUD income group. The HAMI was defined in 2010 by HUD at $28.653. Category Number

Extremely Low Income Households (< 30% HAMI) 12,705 Very Low Income Households (30-50% HAMI) 7,680 Low Income Households (50-80% HAMI) 7,140 Moderate Income Households (80-100% HAMI) 3,905

Table 5.4 Distribution of income groups under HAMI, $28.653 in Gainesville

Source: ACS 2010

In sum more than 30.000 households in Gainesville fall within the categories of HUD who can apply for housing subsidies. The housing supply is however not entirely appropriate for such households. Table 5.2 shows that most available housing is in the segment of rents higher than $500, while there is a huge shortage of housing in the cheaper segments. Since the categories are based on a 30% maximum of income spendable for rents, this means a large share of the less affluent population in Gainesville can be considered cost burdened, paying more than 30%. Data from FHDC (2013) show more than 5000 households were cost-burdened in 2009. There is only limited affordable housing available. However a Section 8 voucher would allow residents to rent in a higher segment without cost burden. Figure 5.10 shows the location of project-based affordable housing. Not so very surprising these locations are situated in areas in the city with relatively high poverty rates. In addition the map shows the rental vacancy rate of each census tract.

Figure 5.10 Housing supply Gainesville, FL.

The above map shows the subsidized housing projects and the available rental units for rent per census tract. These rental numbers are estimate with 90% confidence interval. Source: Shimberg Center; ACS 2010

In Gainesville, 38 per cent of the total housing stock is owner-occupied, while the remaining share is in rental, with a rental vacancy rate of 11.8%. However the vacancy rate of rental units is the largest among the more expensive dwellings. Table 5.2 shows that the vacancy rate for cheaper units is relatively low. An estimate of the American Community Survey shows (with a 90% confidence interval) that rent vacancy is particularly high in the South West of Gainesville while in the North West typically few dwellings are available for rent. Subsidized housing is available throughout the town except for the North West. Glen Springs Manor appears to be situated in an area with hardly any other subsidized complex, while Seminary Lane and Kennedy Homes were situated in areas with relatively many other complexes in the surrounding neighborhoods.

5.4.3: Relocation Tenants from the Kennedy Homes were forced to relocate in 2003 while the tenants from the other two projects had to find alternative housing in the beginning of 2009. In the end only a part of all original tenants relocated using a HUD provided Section 8 voucher. A large share had already left the complexes before vouchers were provided. Either the conditions were too bad for them, they did not want to go through the hassle of Section 8 due to general distrust in the government or they were not eligible for Section 8 due to too low income or criminal backgrounds. 92 households from GSM, 43 from the Kennedy Homes and 35 from Seminary Lane relocated using a Section 8 voucher. For this research it is unknown what happened to other residents who relocated in earlier stages. For the total of 170 Section 8 movers the new location is made available for this research by the University of Florida’s Shimberg Center for Housing Studies. The new location was provided in the form of the new census tract of all 170 Section 8 movers. Figure 5.11 shows the relocation outcomes of the three different projects in separate maps.

The majority of all three projects have moved to the Eastern part of Gainesville In total 101 relocatees have moved to a census tract that lies within the Eastern area highlighted in blue. 69 other relocatees have found housing in other districts of Gainesville. Striking is that these relocates show a much more dispersed pattern than those who have moved to the East. However there is one part just South West of the city limits that has received relocatees from all three projects, and a slight concentration is visible in this spot on all three maps. This area is highlighted in green

Figure 5.11 Gainesville Relocation

The above map shows the number of relocatees per project per receiving census tract. The red census tracts received the most relocatees.

Source: Shimberg Center

Relocatees from the Kennedy Homes and Seminary Lane have only moved outside of the highlighted areas in rare occasions. 10 from the 78 relocatees from these projects moved to the un-highlighted areas. Strikingly many more relocatees from Glen Springs Manor found housing in this area. This difference is also resembled in the summarized neighborhood characteristics of the receiving census tracts.

Project Racial Minorities Families below Units for N Rent

Old New Old New Glens Springs Manor 19.1% 41.6% 9.5% 30.5% 170 92 Kennedy Homes 77.5% 57.3% 30.7% 33.7% 149 43 Seminary Lane 26.0% 53.0% 58.9% 32.6% 164 35 Total X 47.9% x 31.7% 163 170 Gainesville X 33.1% x 18.6% 115 X

Table 5.5: neighborhood characteristics

This table shows the average characteristics of the receiving (new) neighborhoods per census tract. The numbers for the old neighborhood are averages from the census tract of the particular project

Source: US Census 2010, ACS 2010

The poverty level of the new neighborhoods of GSM relocatees is slightly lower than those of other relocatees, but the most striking difference is that the percentage racial minorities is significantly lower. On average both relocatees from Seminary Lane as well as from Kennedy Homes have moved to majority minority census tracts while relocatees from GSM on average moved to neighborhoods with ‘only’ 41.6% racial minorities. In general relocatees have moved to census tract with relatively high percentages of minorities and high percentages of poverty. These findings are however not very surprising since most of the relocatees themselves, belong to both racial minority and living below the poverty line. Relocatees form GSM had an even lower income than those from the two other projects; nevertheless they ended up in tracts with less poverty. Two possible explanations can be given on the basis of the available information. On the one hand GSM was located in an area with relatively low poverty and racial minorities, residents could have been more inclined to move to a similar neighborhood. However another striking difference with the other two projects is that GSM had 30% white residents. The ones that have moved to ‘white’ census tracts could very well have been the white relocatees from GSM. Since the population of GSM is slightly more diverse than the population from the other two projects, the relocation patterns are harder to find causal explanations for. However the

64 relocation from Seminary Lane and Kennedy Homes show some very interesting patterns when neighborhood characteristics are taken into account

Strikingly the characteristics of the old census tracts do not really resemble the characteristics of the population of the complexes. Glen Springs Manor was located in a white neighborhood with a relatively low poverty rate. Seminary Lane was also located in a white neighborhood, but with a seemingly extreme high poverty rate. As was already said, Seminary Lane was located only blocks away from the UF campus, surrounded by student dorms. The Kennedy Homes is the only project that was located in a more homogenous census tract with many African Americans, many of them living below the poverty line. It is therefore questionable to what extent a census tract can be regarded as a neighborhood. However since there is no real alternative, I will focus on census tracts, but be reluctant in conclusions.

As was already shown, the relocatees from these two projects move to neighborhoods with high poverty rates and high shares of minorities. Since the Section 8 vouchers are to be used in private rental vacant units, the relocatees also moved to neighborhoods with relatively many vacant rental units. The patterns in figure 5.10 do resemble the patterns of relocation. Tracts with high numbers of vacant rental units attract more relocatees than tracts with relatively low vacant rental units. The pattern that appears, show that relocatees often find alternative housing in census tracts with more than average shares of racial minorities and more than average vacant rental units. Figure 5.12 shows the relocation outcomes of Seminary Lane and Kennedy Homes, compared with the tracts where there are more vacant rental units than the average of Gainesville and where the population consists of more than 33% of minorities, the rate for Gainesville in general. Slightly more than 70% of the relocatees from these two projects move to a tract that fits both these qualifications. A regression analysis reveals that it is mainly the share of racial minorities that is the best predictor of the number of relocatees a census tract received after relocation. Although the N is relatively low (without GSM it is 78), the correlation between percentage minorities and number of relocatees has an adjusted R squared of 0.3 significant at the 95% interval. The correlation between vacant rental units and number of relocatees per census tract was not significant, perhaps due to the low N.

Figure 5.12 Relocation from Seminary Lane and Kennedy Homes and tracts with more than average racial minorities and vacant rental units

Source: Shimberg Center; ACS 2010; US Census 2010

5.5: Discussion This chapter has shown the results of relocation of 170 relocatees from three projects in Gainesville, Florida. The relocatees were able to relocate using a voucher in the private sector, which enabled them to rent in this sector spending only 30% of their income on rents while HUD would come up with the difference. The rental market in Gainesville shows a considerable shortage of rental units dedicated to the population of this study. Together with the short amount of time that was provided to find alternative housing this has severely constrained the relocatees in their options.

With the low supply of cheap rental units and the reluctance of landlords to accept Section 8 vouchers the relocatees have found alternative housing in the limited amount of time. The location of these dwellings was most often in census tracts with relatively many belonging to racial minorities and living under the poverty line. Especially relocatees from the two projects that were almost entirely populated by African Americans moved to high minority tracts. Strikingly the relocatees from Glen Springs Manor, where 30% of the population was white, moved more often to areas with less racial minorities. This hints at a pattern of increasing segregation after forced relocation: whites move from a mixed project to more white neighborhoods while blacks move to neighborhoods that are populated by other blacks. However this pattern cannot be proven by this study, the results hint at such a situation.

The data provided for this study however make an individualist approach for explaining relocation outcomes implausible. Since there is no real information on the individual characteristics of relocatees, nor any information on their preferences, assumptions based on different agency by different relocatees cannot and should not be made. Clapham and Kintrea (1984) provide two other approaches to housing allocation: institutional and structuralist approaches.

Following the institutional approach the emphasis should have been on the available rental units and the operation of Section 8 vouchers, as if all actors involved would have acted with complete information on the rental market, with no special preferences or other constraints than merely institutional ones. Institutional approaches however fail to explain the concentrations of relocatees in particular neighborhoods, especially the concentration of relocatees in the eastside of Gainesville.

Clapham and Kintrea (1984) explain how a structuralist approach on housing allocation often emphasizes how perpetuating class structures shape residential patterns. The relocation patterns found in Gainesville might very well be explained by more structural causes of class structures. The characteristics of the census tract to which relocates moved to, contain more poverty and racial concentrations of minorities than other neighborhoods of Gainesville, even more than the census tracts where the original projects were situated. This might be the consequence of mere individual preferences of the relocatees, but considering more general literature on Section 8 voucher movers, it is more likely that time pressure and few opportunities were important in creating the patterns. Given the fact that the population of all three projects have dispersed through the city, the preference of staying in the same neighborhood could not be met by most relocatees. The data used in this chapter are by far less detailed than the data used in the next chapter about Amsterdam, the individual characteristics of relocatees are lacking and the new location is only known as a census tract. Nevertheless the outcomes provided in this chapter do reveal interesting patterns. Actual information on improvement of housing situation is difficult to distillate from the provided information. However, since the quality of the original housing was in such bad conditions it had to be demolished, most likely residents who moved with a Section 8 voucher have found at least some improvement in terms of housing quality. Only the residents from Kennedy Homes moved to a census tract with less ethnic minorities.

Although the data used in this analysis are limited, the conclusion can very well be made that most relocatees have been displaced. Although most will have encountered housing improvement, the dispersed geographical pattern throughout neighborhoods where relatively many racial minorities and poor live suggests that relocatees moved to neighborhoods where they could find something instead of finding something in the neighborhood of their preference. Moreover the short time period that was provided to find alternative housing will have increased the likelihood that relocatees had not many options for relocation.

6. Amsterdam In this section I will discuss the outcomes of the study of four forced relocation projects in Amsterdam. Since the data provided for this study are extremely detailed and contain information on more than 90% of the relocatees, the outcomes of this section are very unique in its kind. Never before was one able to conduct a research with these highly detailed data. Therefore the outcomes potentially can add a substantial evidence for theories on displacement and forced relocation. Before going into detail on the relocation I will first briefly describe the context of Amsterdam and the borough Amsterdam Nieuw West in which the four projects under study are situated. Then I will discuss the location of the four projects before discussing the outcomes of the relocation analyses.

6.1 Context

6.1.1 Amsterdam Amsterdam is the largest city in the Netherlands; it is also the capital city of the Netherlands. In 2006, the year the first project in this study started, Amsterdam had approximately 750 thousand inhabitants. More than half of the Amsterdam housing stock was social rent, compared to only a third in the Netherlands. Similar as in the other big cities (Rotterdam, Utrecht and The Hague) Amsterdam consisted of relatively many non-Western ethnic minorities. The unemployment was slightly above the national average but under the average of the other big cities.

Total Population Unemployment Non-Western Social Rent Ethnic Minorities

Amsterdam 743027 6.9% 34.5% 51.8% Other Big Cities 1345309 7.8% 29.3% 43.3% Netherlands 16334210 5.5% 10.6% 33.2%

Table 6.1 Picture of Amsterdam compared to other big cities (Rotterdam, Utrecht and The Hague) and the Netherlands as a whole. The data are from 2006, the year the first project started. Source: (O+S Amsterdam, 2006; O+S Amsterdam, 2007) Ethnic minorities from different backgrounds show a strikingly different residential pattern in Amsterdam. As Figure 6.1 shows, Antilleans and Surinamese minorities tend to concentrate in

Amsterdam Zuid Oost while Turks and Moroccans mostly live in the Western part of Amsterdam, particularly in Amsterdam Nieuw West (ANW).

6.1.2 Amsterdam Nieuw West The four projects under study were all part of the large renovation plans of Amsterdam Nieuw West. ANW is the most western part of Amsterdam and consists of the districts Osdorp, Geuzenveld-Slotermeer, Slotervaart and parts of Bos en Lommer (see figure 6.1). This district was more or less built in the 1950s and 1960s. Modernistic virtues such as light air and space have dominated the building process. The pressing housing shortage dictated the architects to build large numbers of dwellings in the shortest period of time. This resulted in the typical post war housing style of post-war housing which can be found in most cities in the Netherlands. The rectangular buildings of four stories and large high-rise apartment complexes are found everywhere in ANW. Since there is virtually no construction from before 1950 in ANW, it looks very distinct from neighboring districts such as de Baarsjes and Oud-West which were built respectively in 1920s/1930s and in the late 19th century (generally speaking).

InFigure the 1990s 6.1: Location increasing of projects poverty, under ethnic study segregation, and the Amsterdam and criminality district were of Amsterdam accumulating Nieuw in West ANW.

ANW became subject of all BCPs that were mentioned in chapter 2. Large-scale revitalization of the post-war housing stock played a big part in these policies and therefore occurred massively in ANW. ANW was as a whole identified by the national government as a so-called ‘krachtwijk’ (power-neighborhood), a list of 40 neighborhoods that would receive national government funding in order to combat accumulating problems such as concentrations of unemployment, criminality and poverty.

6.2 Case description

6.2.1 Bakema Sloop Noord Figure 6.2 shows the location of the two projects in Geuzenveld. Bakema Sloop Noord has been demolished in 2011. The last tenant was relocated in 2008. For the demolition of BSN there was no new construction in replacement. Although there were plans of that at the start of the project, economic downturn made the housing associations involved decide not to build new construction. Therefore relocatees from BSN did not have the opportunity to return to their old location. However the blocks on the other side of the street were being renovated at the same time BSN was demolished; if residents form those blocks decided not to return, dwellings became available and relocatees from BSN had the chance to rent those dwellings before they were advertised. Relocatees could await the renovation of those blocks while living in a temporary home within the neighborhood. The locations of the temporary homes are shown in figure 6.2 as well.

BSN was built in 1956 and named after its architect Jaap Bakema. The buildings were built in the very typical functional modernistic style of four layered dwelling with one main entrance door. The project consisted of three structures at the Jan Bongastraat, Johan Coltermanpad and Sam van Houtenstraat. At the start the three flats were mainly occupied by native Dutch households. Due to the relatively spacious apartments – the apartments varying between 45 and 75 square meters consisted of 3 to 5 rooms- they were mainly popular among larger families. As was the case for the largest part of ANW, the population began to change during the 1980s. Guest workers who did not return to their home country were moving in. In ANW the influx of immigrants consisted of mainly Turks and Moroccans.

Figure 6.2: Satellite image of projects in Geuzenveld. All projects under study in Geuzenveld have been demolished. The eastern part of Van Tijenbuurt has been replaced with new construction while the other sites are still vacant.

Table 6.2 shows the general population characteristics of the project on the days before the reference date, the first of January 2006. In the two years after that date the tenants have relocated. The last household moved out in June 2008.

6.2.2 Van Tijenbuurt Just like BSN the Van Tijenbuurt was built in 1957 and named after its architect, Willem van Tijen in this case. The dwelling types of Van Tijenbuurt are a little more diverse than those of BSN. The dwellings were located on the Dudok de Withof, van Karnebeekstraat, Jasper Warnerhof, Willem Kromhoutstraat, Aalbersestraat, Weissmanstraat and De Savornin Lohmanstraat. Most dwellings were very similar to those from BSM except for the units on the

Willem Kromhoutstraat, Weissmanstraat and van Karnebeekstraat. The 30 dwellings on these streets only had one room and were hence occupied by one or two persons. Besides this difference in population the characteristics of the tenants are very similar to those of the tenants from BSN, with about two thirds of immigrants, from Turkey and Morocco. For an overview on all characteristics see table 6.2.

All relocatees of this project had the opportunity to sign up for returning to the original location in a newly constructed dwelling. However, as is often the case there was not enough new social housing to provide new housing for all tenants. Based on family size income and length of occupancy in the original apartment, the newly constructed dwellings were allocated among the residents. Again they could await the new construction in a temporary home. Coincidentally the construction of the new houses is very nicely shown in Google Streetview. If you put the little orange guy of the web-app on the corner of the Nolenstraat and de Savornin Lohmanstraat and look in south eastern direction, you will see the completely flattened site of where the apartments of the tenants were located (picture taken May 2008). Then, if you ‘walk’ in eastern direction and keep ‘looking’ on the demolished site there is one spot in Google Streetview where you can see the beginning of a new apartment complex (August 2009) and one spot where you can actually see the finished complex, in red stone (October 2010). These apartments were reserved for the relocatees from the Van Tijenbuurt.

6.2.3 Kolenkitbuurt Also the flats in the Kolenkitbuurt are located in ANW and were built during the 1950s. However the Kolenkitbuurt is part of the sub district Bos en Lommer which is only partly within the area of ANW. The two projects in Kolenkitbuurt are actually two phases of one and the same project, Zuidelijk Veld phase 1 and Zuidelijk Veld phase 2. Like the projects in Geuzenveld, phase 1 was owned by housing association Rochdale, phase 2 was owned by another association, Stadgenoot. The relocation in both projects was however led by the relocation staff of Rochdale.

The Kolenkitbuurt received national attention by topping several rankings of for instance neighborhood with worst livability and most problematic neighborhood. One of those lists was the one composed by the national government in order to combat problems in the 40 most deprived neighborhoods in the country. Revitalization is therefore omnipresent in the Kolenkitbuurt. Maybe it is not a coincidence that the main office of housing association

Rochdale is situated in the Kolenkitbuurt (in a highly modern office building bridging the Amsterdam beltway A10). During the 1980s immigrants found the Kolenkitbuurt and moved in en masse. In 2010 roughly 80% of the population of this neighborhood had an immigrant background.

Figure 6.3: Satellite image of projects in Kolenkitbuurt. Zuidelijk Veld Fase 1 has been demolished and even replaced with new construction, which was dedicated to relocatees from other projects. Zuidelijk Veld Fase 2 is at the moment of writing still there. The building bridging the motorway is Rochdale’s main office.

In the two projects more than 80% of the population was belonging to an ethnic minority, mostly Moroccans and Turks. The relocation for phase 1 began in 2006 while phase 2 stared in 2008. The last tenant left phase 1 in 2008 while the last tenant has left phase 2 only in 2011. Again there were a limited number of new dwellings available for tenants which they could return to.

These dwellings are located across the street also in the Kolenkitbuurt. Phase 1 relocatees had to await there move to newly constructed dwellings in temporary homes in the neighborhood, phase 2 relocatees had the opportunity to move right away. The temporary homes were located within the Kolenkitbuurt.

Zuidelijk Veld phase 1 has been demolished, while phase 2 is still awaiting demolition. Due to financial shortages the association did not have the money to start demolishing. The dwellings are being occupied by temporary renters, including many students. Although there are plans to start demolition in 2014, it is very unsure whether these plans will actually take place. Most people working or living in the Kolenkit do not expect demolition in the coming years.

Starting Demolition Average People Average Average % Ethnic Returned Relocated Left Total year year dwelling Per Monthly Annual Minority Regional size Unit Gross Gross Assisted Rent Income Housing

Bakema Sloop Noord 2006 2010 54.05 3 € 421.05 € 23,935.45 65% 11 79 8 98 Van Tijenbuurt 2006 2009 39.73 2 € 232.72 € 27,648.94 67% 61 107 9 177 Zuidelijk Veld Fase 1 2006 2009 53.4 3.1 € 406.58 € 23,837.07 83% 52 110 31 193 Zuidelijk Veld Fase 2 2008 Not yet X 2.6 x € 29,640.07 82% 45 46 13 104

Total X X 48.3 2.7 € 343.86 € 26,102.54 75% 169 342 61 572

Table 6.2: Picture of Population of Projects

This table shows the general characteristics of the population of the four projects under study. Data on the dwellings of Zuidelijk Veld Fase 2 was not available for this research, since the houses were owned by another association than Rochdale.

Data source: Own Survey/Rochdale

6.3 Outcomes From here the outcomes of the analyses of the relocation will be discussed. Each concept discussed in the methodology part will be discussed. For chronology’s sake I will start with discussing the neighborhood preferences, since they were identified before actual relocation.

6.3.1 Preferences In the was shown in table 6.2 that a share of the tenants who had to leave their house in the four projects returned to newly constructed housing that was especially reserved for them. Despite the relatively high rents of these dwellings they were equally popular among different income groups as well as ethnic groups. The combination of a newly constructed dwelling and the location within the original neighborhood appeared very appealing for all different types of relocatees.

Figure 6.3: Neighborhood preferences per district

This figure shows how many times each district was mentioned by the tenants. Relocatees could

mention up to five different districts of Amsterdam in which they preferred to live after relocation.

Data source: Own Survey/Rochdale

The idea that relocatees want to stay in their old neighborhood becomes even clearer in figure 6.3 in which neighborhood preferences are visualized. All tenants had the opportunity to mention districts of Amsterdam to which they wished to move. Clearly, this reveals the common wish among relocatees to remain in the western part of Amsterdam, within or close to the original location. The concentration of neighborhoods mentioned strikingly resembles the wider district of Amsterdam Nieuw West (ANW). Three districts fall completely within the boundaries of ANW (Geuzenveld-Slotermeer, Osdorp and Slotervaart) while one district is partly within ANW (Bos en Lommer). Yet a considerable number of relocatees also mentioned other districts in Amsterdam in which they preferred to live that are not within ANW and often located at larger distance from the original location. There is however no other concentration of preferences than in ANW.

In contrast to the desire to return to newly constructed housing, some striking differences were found looking at the difference in preferences per type of relocatee. Although affluent and less affluent relocatees had similar neighborhood preferences, native Dutch and immigrants had significantly other preferences.

Household Type Preferred Districts in ANW N Small Native Dutch Household 63% 108 Small Immigrant Household 74% 215 Large Native Dutch Household 55% 13 Large Immigrant Household 89% 152 Total 76% 488

Table 6.3 Preferences for Amsterdam New West (ANW) by household type

This table shows the average percentage of districts mentioned by relocatees that is located within the larger district of ANW. Data source: Own Survey/Rochdale

For both large and small households, the desire to remain resided in ANW appears to be more present among immigrants, yet almost all large immigrant families indicated they only wanted to move to a district that was located in ANW. Dutch relocatees not necessarily mentioned fewer

78 neighborhoods in ANW than immigrants, but showed a broader preference by including more districts elsewhere. Opposed to immigrants, small Dutch households prefer to stay in ANW more than larger Dutch households. Due to the limited number of large Dutch households this link appears to be weak. However, there might be a good explanation for this pattern: the small Dutch households have one unique feature: the main occupant of these households is on average ten years older than the main occupant of the other households. Most of the small native Dutch households turn out to belong to the elderly and therefore may have other preferences. Most of these elderly have a long residential history in ANW and were often part of the first wave of residents who settled in ANW when the district was built. They might have more bonding with the neighborhood due to this long history and therefore do not want to move to any other neighborhood in the last phase of their lives (de Bruijn, 2011).

6.3.2 Actual Relocation Preferences are however not limited to a choice for different neighborhoods. Relocatees weigh factors like location, quality and housing costs all in their own way according to their own knowledge. In order to accelerate the relocation process and save costs, the housing association tried to increase this knowledge, for instance by providing bus trips to neighborhoods that were located outside the mental maps of most relocatees, but with plenty of social rented dwellings available, such as the newly built neighborhood IJburg. Nevertheless, the general relocation pattern appears to be pretty much along the lines of the original neighborhood preferences: the vast majority of relocatees stayed in Amsterdam Nieuw West. Many relocatees stayed in or close to the original neighborhood. Figure 6.4 shows the new locations of all relocatees. Just like the preferences however, relocation patterns are different for different types of relocatees. In this part I will try to unravel these different patterns.

The tendency to remain within the same district occurs at even a smaller scale than ANW. Aside from the concentration of relocatees within the boundaries of ANW, within this concentration relocatees from Geuzenveld tend to concentrate in Geuzenveld while similarly relocatees from Bos en Lommer concentrate in Bos en Lommer. To prove that this concentration is not caused by some specific character of ANW and would not occur somewhere else, I have also included the relocation pattern of relocatees from Kleiburg, a demolition project in Zuid-Oost. The relocation pattern of Kleiburg is included in figure 6.5. Relocatees have acted geographically in a similar

79 fashion as did relocatees from ANW, meaning they stayed close to their old neighborhood. Since data on the individual characteristics were not reliable and complete enough this project is excluded in all other calculations.

Figure 6.4 Relocation patterns of the four projects

Each dot represents one relocated household. The households who have returned are also included in this picture, however since they tend to live so close to one another, visually they are represented by only one dot together.

Data Source: Own Survey/Rochdale

Figure 6.5 Relocation patterns of the four projects and relocatees from Kleiburg in Amsterdam Zuid- Oost

Each dot represents one relocated household. The households who have returned are also included in this picture, however since they tend to live so close to one another, visually they are

represented by only one dot together.

6.3.2.1 Differences per relocatee type Again ethnicity and family size seem to be good predictors for location decisions, as table 6.4 shows. Native Dutch households not only more often prefer other neighborhoods, they also show a more diverse actual relocation pattern. Where most immigrant households, especially the large ones, stay in Amsterdam Nieuw West, relatively more Dutch households tend to leave the area. The outcomes of the relocation resemble the preferences. Immigrant families – especially larger ones – stay in ANW while a larger share of the Dutch households moves away, with the exception of small Dutch households with lower incomes, who also stay attached to ANW. However, Dutch small households with higher incomes leave. The last mentioned difference was not present in the preferences.

Relocatee Type Relocated Other Total N to ANW Low Income - Native Dutch - Small Family 68% 32% 100% 82 Low Income - Native Dutch - Large Family 33% 67% 100% 9 Low Income - Immigrant - Small Family 66% 34% 100% 163 Low Income - Immigrant - Large Family 81% 19% 100% 101 High Income - Native Dutch - Small Family 39% 61% 100% 36 High Income - Native Dutch - Large Family 80% 20% 100% 5 High Income - Immigrant - Small Family 66% 34% 100% 67 High Income - Immigrant - Large Family 93% 7% 100% 71 Total 71% 29% 100% 534

Table 6.4 Share of relocatees newly housed in Amsterdam Nieuw West (ANW) after relocation, by relocatee type

Cramer's V= 0.302, p<0.001 Data source: Own Survey/Rochdale

Those who found a new dwelling outside of ANW turned out to be also those who indicated a preference for living outside ANW; only 48 per cent of those who moved outside ANW (also) preferred ANW, while 88 per cent of the households that relocated in ANW (also) expressed an ANW preference. No household left ANW while they indicated they did not want to and no one stayed in ANW while indicating they did not want to.

Not only locational differences between the relocation patterns of different types of relocatees were found, the receiving neighborhoods also differed between the different types, as shown in table 6.5a. The data on neighborhoods in table 6.5 a and b are from Statistics Netherlands, for the year 2010 (CBS, 2010). Since all relocatees came from two neighborhoods in Amsterdam with extraordinary high concentration of relative poverty and more than 60 per cent ethnic minorities, on average they have moved to neighborhoods with lower concentrations of both minorities and households living on social minimal standards. Within the paradigm of mixed neighborhood and poverty de-concentration, on average all groups of relocatees have improved their living situation, supporting earlier findings from the literature (Kleinhans 2003; Doff and Kleinhans

2011). Not surprisingly the geographical differences between relocatee-type are reflected in the corresponding neighborhood characteristics.

Relocatee Type Non-Western Social N Immigrantsa Minimab Low Income - Native Dutch - Small Family 45% 17% 82 Low Income - Native Dutch - Large Family 29% 11% 9 Low Income - Immigrant - Small Family 53% 19% 163 Low Income - Immigrant - Large Family 59% 19% 101 High Income - Native Dutch - Small Family 37% 15% 36 High Income - Native Dutch - Large Family 45% 16% 5 High Income - Immigrant - Small Family 50% 18% 67 High Income - Immigrant - Large Family 63% 20% 71 Total 52% 18% 534

Table 6.5a Ethnic and social characteristics of the ‘new’ neighborhood, by relocatee type

a. R2= 0.135, p<0.001 2 b. R = 0.101, p<0.001 Data source: Own Survey/Rochdale/CBS

Data source: Own Survey/Rochdale Type Non-Western Social Minima Immigrants Geuzenveld 60% 19% Kolenkitbuurt 75% 24% Amsterdam 35% 16%

Table 6.5b Ethnic and social characteristics of the Amsterdam neighborhoods from which households relocated (for 2010)

Data source: CBS, 2010

Again ethnic differences between relocatees are the best predictor for new neighborhood characteristics. Native Dutch relocatees move to neighborhoods with less non-western immigrants and less households living from minimum standards. Only the native Dutch small families with higher incomes show different outcomes than their immigrant counterparts. Large

83 native Dutch families show a striking opposite effect; however, the number of observations is too small to draw meaningful conclusions.

6.3.3 Quality of housing In addition to the analysis of where the displaced end up the physical qualities of the new dwellings themselves were also examined. I tried to answer the same question: to what extent do the qualities of the new housing differ between the different types of relocatees? Housing quality is difficult to measure since data on the physical condition of individual dwellings is scarce. I was able to approximate the physical quality of dwellings by looking at two quantitative variables: the size of a dwelling and the year it has been built. Generally speaking larger dwellings are considered of higher quality than smaller dwellings. Building year however is not as linear. Often older houses are regarded as of lower quality than newer ones; in Amsterdam this assumption is false. In the period after WWII, many houses of relative low quality were built speedy in order to comply with the housing shortage. The Amsterdam socially rented housing stock that was built between 1945 and 1975 is believed to be of considerably lower quality than housing of other building years (Rowlands et al., 2009). Since the original site was built during this post-war period, moving to a dwelling of another building year can be interpreted as an improvement of housing quality while staying in housing built in the same period is not an improvement.

Floor space of the dwellings is generally regarded to be a good indicator of the quality of a dwelling. Most households show an increase in floor space or remain in a similar size dwelling. Only 15 per cent moved to a considerably smaller apartment. Any significant between-group variations in floor space change because of the relocation were lacking. On average tenants relocated to an apartment that was 14 square meters larger than their previous apartment. Even those who returned to the original location in newly built housing did not find significantly different changes in floor space from other relocatees. In general all types of relocatees have experienced an improvement on floor space.

As was mentioned earlier, a large share of the relocatees returned to the original site in new construction housing. Hence it is not surprising to see that on average the relocatees made considerable improvement in terms of building year. Only 35 per cent of the relocatees from which the building year of their new dwelling is known (496 households), moved to a dwelling

84 that was built during the post-war era of reconstruction. Except for a small and barely significant correlation between income level and building year, the different types of relocatees did not show significantly different outcomes in terms of building year of their new dwelling. Since there are no correlations between relocatee characteristics and the dummy variable of returning or relocating, those who returned to the new construction might conceal possible correlation. Table 6.6 shows the housing quality of those that relocated, while returnees have been excluded. Due to their low numbers, native Dutch large families are excluded from table 6.6.

Relocatee Type Post-War Other Total N Period Low Income - Native Dutch - Small Family 42% 58% 100% 57 Low Income - Immigrant - Small Family 43% 57% 100% 118 Low Income - Immigrant - Large Family 67% 33% 100% 51 High Income - Native Dutch - Small Family 29% 71% 100% 21 High Income - Immigrant - Small Family 40% 60% 100% 40 High Income - Immigrant - Large Family 57% 43% 100% 30 Total 47% 53% 100% 317

Table 6.6 Building period of the new dwelling, by relocatee type

Data source: Own Survey/Rochdale

First of all it is striking that even without those who moved to new construction, the majority of the relocatees managed to find a new dwelling that was built in a ‘better’ period. Just as was the case in neighborhood characteristics, the native Dutch small households with relative higher incomes stand out by relocating more often to houses from other building years. Even more striking is that the majority of large immigrant families, especially those with lower incomes, did not manage to find an apartment that was not built during the post-war era. Less than 10 per cent of the Amsterdam housing stock has five or more rooms, and the share is even smaller in the social rented sector. Large households have a difficult time finding adequate alternative housing regardless due to the limited availability. Not only have large families remained in Amsterdam Nieuw West, but a majority of them also ended up in post-war housing, meaning they only found improvement in the size of their dwelling.

6.3.4 Rent In the original situation large and small households, rich or poor, were paying approximately similar rents each month. The difference between average rent of lower income households and higher income households was only €10 per month. But how about rent differences in the newly obtained dwellings? Because housing associations are allowed to harmonize rent prizes when a residential move occurs, changing house may mean a considerable rent increase. Next to possible different preferences among more affluent and not so affluent relocatees, the supply of adequate housing that was built in another period than the post-war era may have influenced these outcomes.

Since all relocatees have moved to a different dwelling, they encountered a different rent. As expected, most relocatees had to cope with significantly higher rents. The highest increase in rent is usually found for those who move to newly constructed dwellings (Kleinhans, 2003: 488). The rents of those who moved to new construction from Van Tijenbuurt and Zuidelijk Veld Phase 1 were compared with the rents they were paying before relocation. Among these tenants on average the rent increased from a monthly €321 to €526 per month. In exchange of this enormous rent increase the relocatees got a newly constructed dwelling with estimated values ranging between €200.000 and €300.000. In order to overcome severe housing costs, those who earned less than roughly €2.500 gross a month were entitled to rent subsidies, which vary between €200 and €350 per month and are dependent on age, family size, income and rent level. When the often used idea that rent levels to be paid should not exceed 30 per cent of gross income is applied (McClure, 2005), no relocatee that moved from Van Tijenbuurt and Zuidelijk Veld Phase 1 to a newly constructed dwelling passed that limit after rent subsidies

6.4 Summary The study of the four Amsterdam cases has showed that most relocatees, especially those with an immigrant background prefer to stay resided in ANW. Actual relocation patterns have confirmed these preferences, meaning that most relocatees were able to fulfill their neighborhood preferences. The difference among different relocatee groups can largely be explained by different preferences. Since nearly all relocatees found improvement in housing conditions and mover to the neighborhood of their preference and had ample time to find alternative housing, they can barely be called displaced from their original place. More elaborate discussion and conclusions will be provided in the next chapter, effectively called discussion and conclusions.

7. Discussion and Conclusions In this section I will discuss the findings and I will try to provide an answer to the four research questions. Since the first three questions only apply to the Amsterdam case I will start with discussing the findings in Amsterdam and answer research question 1, 2 and 3. After that the Amsterdam findings will be compared with outcomes from Gainesville, and thus answering research question 4.

Despite the fact that a forced relocation can have quite a disruptive effect on a family, and can be perceived as unwelcome (Kleinhans, 2003; Allen, 2000; Ekström, 1994), relocation meant an improvement of the housing situation for most relocatees. This supports most findings of earlier studies into forced relocation in the Netherlands (Kleinhans, 2003; Bolt and Van Kempen, 2010; Doff and Kleinhans, 2011; Posthumus et al., 2013a) as well as in the US (Clampet-Lundquist, 2004a; Buron, 2004; Popkin et al., 2009). These findings are however not very surprising since the quality of their original housing was limited, and their neighborhood contained the highest shares of non-Western immigrants, among the lowest income levels and a very large proportion of the housing stock being social rent. The answer to the question whether relocatees are improving their housing situation can nevertheless be answered affirmative. Strikingly relocation patterns appeared to be very different among different types of relocatees. Income, ethnicity and family size all had their own influence on relocation. Large immigrant families tend to stay close to the original site and move to neighborhoods with relatively larger shares of minorities and households dependent on a social minimum income. On the other hand native Dutch relocatees from restructuring neighborhoods move relatively often to places further away from the original site and outside the boundaries of the Amsterdam Nieuw West.

In addition to the complete dataset that was used, another advantage of the data was that it contained information on preferences before the actual process of relocation. I believe that these preferences are less affected by institutional constraints, since they were identified before the relocatees started looking for new housing. The fact that different types of households showed different preferences but correspondingly different relocation patterns suggests that the difference between ethnicities, income groups and different household types, are largely influenced by different preferences. Large families presumably want to remain close to the original location because of amenities such as schools, sport clubs and religious facilities while

87 small households are generally less dependent on these amenities nearby. However, regardless of family size, immigrants wanted to stay in ANW more than native Dutch relocatees. Strikingly income did not affect preferences, but it did have effect on actual outcomes. On the one hand we may assume that those with higher income have a broader range of opportunities, since they can afford to live in more expensive housing. On the other hand, social housing and rent allowances are allocated based on income. Those with higher incomes are excluded from a certain share of the supply.

The actual outcomes are, however, not a perfect representation of mere preferences, since relocatees are limited by several institutional constraints. Vacant socially rented dwellings in Amsterdam are scarce and due to massive restructuring priorities have been subject to inflation; however, within the limits of these constraints, all relocatees were able to find alternative housing and did have at least some choice. The outcomes support earlier findings by Posthumus (2013), that forced relocatees should be treated as active agents.

Tenants relocated to a neighborhood of their preference and to a dwelling of higher quality than their old dwelling. Immigrants and large families end up in neighborhoods with relatively more ethnic minorities and poverty, but apparently also within the neighborhood of their own preference. This finding puts some question marks to research outcomes generated by Bolt et al. (2009), who argue that ethnic minorities would be less able to benefit from the opportunity given to relocate to a ‘better’ neighborhood. Actual behavior of ethnic minority households reflects their preferences. This may not result in a move to physically better neighborhoods, but it does seem to reflect a consolidation of existing social networks and proximity to essential services. Moreover, research shows that living in such neighborhood not necessarily decreases income or educational outcomes (Van Ham and Manley, 2010), while the assumed relation between segregation and integration is highly contested (Peach, 2009; Phillips, 2009). This combined with research that shows that overall relocatees appear to be more satisfied with their living situation after forced relocation (Meerts et al., 2011; Posthumus, 2013; Doff and Kleinhans, 2011; Kleinhans and Van der Laan Bouma-Doff, 2008) sheds a different light on the discussion regarding social mix.

7.1 Comparison As was said earlier the outcomes from Gainesville are not derived from the same type of data as was used in Amsterdam. Caution is therefore required making this comparison. However the outcomes will of Gainesville were in some sense comparable to the outcomes from the Amsterdam case.

The most striking outcome of the relocation in Gainesville was that the largely African American population of the three projects relocated to census tract with significantly more racial minorities than other tracts in Gainesville. The data even suggest –although do not prove- that the share of white relocatees moved to neighborhoods with less racial minorities. At first sight these findings resemble the outcomes from Amsterdam where ethnic minorities more often stayed in ANW while Dutch relocatees tended to leave this area more often. ANW contains considerably more ethnic minorities than most other districts in Amsterdam. However there is one major difference between the two cases.

Gainesville has several different areas which are mainly populated by racial minorities, the so- called majority-minority tracts and those with slightly less than 50% minorities. These districts are located in the east of Gainesville and just south-west of Gainesville’s city-limits. Relocatees from all three projects, regardless of their location within Gainesville moved for a large share to exactly these districts. Disproportionally many relocatees moved to those majority-minority areas, even if their project was not located nearby.

Amsterdam has several districts with disproportionally many ethnic minorities as well. Next to ANW, many ethnic minorities live in the Indische Buurt in the east of Amsterdam and even more ethnic minorities live in the district Amsterdam Zuid-Oost, in the south-east of Amsterdam. ANW and Indische Buurt are largely populated by Turks and Moroccans, while Zuid-Oost is populated by Surinamese, Antilleans and Africans. I showed that relocatees form ANW do not move very often to Zuid-Oost, while the majority of relocatees from Zuid-Oost in contrast remained in more than 70% in this district. Moreover relocatees also ignored neighborhoods such as the Indische Buurt.

Assuming that relocatees from Gainesville had a certain attachment to their old neighborhood and had similarly to the Dutch relocatees a preference to stay within close proximity to their old

89 neighborhood, the Gainesville relocatees were less able to fulfill this preference. In any case they were less able to stay in close proximity to their old location. These outcomes combined with the fact that the American relocatees did not have the two years to find alternative housing but only had a couple of months strengthens the argument that these relocatees were genuinely displaced. All three projects were inadequately maintained, got a negative score at HUD’s survey and the non-governmental institution that was operating the projects had to decide for demolition. Due to either disinvestment, mismanagement or any other reason that was beyond the fault of the residents, the residents became the victim of the eventual demolition since they had to relocate to another dwelling. Although the decision to demolish was also in Amsterdam beyond the influence of the residents, this study does not show the resident can be called genuine victims of the demolition.

The social housing market in Amsterdam is outsourced to semi-private housing associations, but very strict regulations seemed to have catered for a more or less constant diverse supply of social housing for the relocatees to choose from. Moreover it is the responsibility of the housing association to provide actual alternative housing. They can only begin with demolition when every household has found adequate alternative housing. It exactly this responsibility, that is lacking for the operators in Gainesville. In Gainesville the projects were scoring insufficient with HUD’s survey, and when the operator is not able to make the necessary repairs HUD decides to stop its project based Section 8 subsidy. This means the operator can no longer rent out the dwellings without losing too much, so decides to demolish. Subsequently HUD provides the relocatees with Section 8 rental vouchers and the operator can demolish. In this way the operator does not care about eventual successful relocation and has no responsibility for the relocation whatsoever. The relocatees are dependent on the willingness of landlords to accept Section 8 vouchers. The supply of alternatives appears to be less diverse and the time frame was considerably shorter to find alternatives. The institutional context in Gainesville was unable to prevent mass displacement out of the three demolished projects.

7.2 Conclusion and recommendations The growing literature of critique on social mixing policies such as the urban restructuring policy in the Netherlands, is based on the idea that restructuring would in fact be a form of state-led gentrification with no positive effects for the less affluent residents of mixed neighborhoods

(Bridge et al., 2012). Lees (2008) argues that ‘[t]he movement of middle-income groups into low income areas creates overwhelmingly negative effects, the most significant of which is the displacement of low-income groups’ (2008: 2457). Although the arguments against the widespread social mixing paradigm seem convincing, this study shows – at least for Amsterdam – that ‘displacement’ after forced relocation might be perceived less negatively by residents than is often assumed.

This study’s strong point is the unique data that were available. By including nearly all relocatees, I was able to avoid sample selection bias, which other studies had to cope with. A limit however was that only demolition projects in Amsterdam were included, conducted by one housing association. In the Netherlands, rules for forced relocation differ per municipality but the bottom line is that any household forced to relocate should be offered ‘reasonable’ alternative housing. This is enforced by national law. I assume that the outcomes are generalizable to other forced relocation experiences in Amsterdam, and even the Netherlands but more research is required to support this assumption. Significant differences in both preferences and relocation among different groups of relocatees were found.

Key factor to prevent displacement due to urban restructuring appeared to be an institutional context which can provide adequate alternatives. Irreplaceable in this context are that responsibility of providing alternatives lies with the operator of the demolition and a constant diverse supply of housing alternatives must be available. The Gainesville case shows that these two conditions are not a matter of course and should therefore be well guarded if displacement is to be prevented. However the social housing stock in both Amsterdam as well as in the rest of the Netherlands is declining. Housing associations are selling units to remain profitable and are increasingly entering the private rental market. Moreover the housing association system is on the verge of major reforms since an increasing number of associations appear unable to fulfill its task of providing housing at reduced costs.

This thesis has generated several starting points for future research. Although one of the stronger points of this study was its database, certain aspects could very well improve it so conclusions can be drawn with more robustness. The preferences measured in this research merely consisted of a number of neighborhoods identified by the relocatee in which they desired to live. The reasons behind these preferences are still unknown. Although the staff of the associations has

91 some clues on these reasons and questionnaires with relocatees in previous research has these clues as well, a complete dataset with reasons for preferring certain districts could very well improve the understanding of the found relocating patterns. In addition future research should focus on the supply of alternative housing for relocatees in order to find trends in this supply which may have influenced relocation patterns. Since at least for the Amsterdam region the advertisement of all social housing is operated by only one organization, ‘Woningnet’, it should be possible to map this supply throughout time. However to my knowledge there has not been any study that has done such an analysis.

References

Allen T. (2000) Housing Renewal—Doesn’t it Make You Sick? Housing Studies 15: 443-461.

Andersen HT and Van Kempen R. (2003) New trends in urban policies in Europe: evidence from the Netherlands and Denmark. Cities 20: 77-86.

Atkinson R. (2004) The evidence on the impact of gentrification: new lessons for the urban renaissance? International Journal of Housing Policy 4: 107-131.

Bolt G and Van Kempen R. (2010) Dispersal patterns of households who are forced to move: desegregation by demolition: a case study of Dutch cities. Housing Studies 25: 159-180.

Bolt G, van Kempen R and van Ham M. (2008) Minority Ethnic Groups in the Dutch Housing Market: Spatial Segregation, Relocation Dynamics and Housing Policy. Urban Studies 45: 1359-1384.

Bolt G, Van Kempen R and Van Weesep J. (2009) After urban restructuring: relocations and segregation in Dutch cities. Tijdschrift voor economische en sociale geografie 100: 502- 518.

Bridge G, Butler T and Lees L. (2012) Mixed Communities: Gentrification by Stealth?, Bristol: The Policy Press.

Briggs XdS. (1997) Moving up versus moving out: Neighborhood effects in housing mobility programs. Housing policy debate 8: 195-234.

Briggs XdS. (2008) MAXIMUM FEASIBLE MISDIRECTION: A REPLY TO IMBROSCIO. Journal of Urban Affairs 30: 131-137.

Briggs XdS, Popkin Susan J and Goering John M. (2010) Moving to Opportunity: The Story of an American Experiment to Fight Ghetto Poverty, New York: Oxford University Press.

Briggs XdS and Turner MA. (2006) Assisted housing mobility and the success of low-income minority families: Lessons for policy, practice, and future research. Nw. JL & Soc. Pol'y 1: 25.

Bristol KG. (1991) The Pruitt-Igoe Myth. Journal of Architectural Education: 163-171.

Brooks F, Zugazaga C, Wolk J, et al. (2005) Resident Perceptions of Housing, Neighborhood, and Economic Conditions After Relocation From Public Housing Undergoing HOPE VI Redevelopment. Research on Social Work Practice 15: 481-490.

Buron L. (2004) An improved living environment? Neighborhood outcomes for HOPE VI relocatees. A Roof Over Their Heads: Changes and Challenges for Public Housing Residents. Waschington, DC: Urban Institute.

CBS. (2010) Wijk- en buurtkaart 2010. Den Haag/Heerlen: Centraal Bureau voor de Statistiek.

Clampet-Lundquist S. (2004a) HOPE VI relocation: Moving to new neighborhoods and building new ties. Housing policy debate 15: 415-447.

Clampet-Lundquist S. (2004b) Moving Over or Moving Up? Short-Term Gains and Losses for Relocated HOPE VI Families. Cityscape 7: 57-80.

Clampet-Lundquist S. (2010) “Everyone Had Your Back”: Social Ties, Perceived Safety, and Public Housing Relocation. City & Community 9: 87-108.

Clapham D and Kintrea K. (1984) Allocation systems and housing choice. Urban Studies 21: 261-269.

Coleman JS. (1986) Social Theory, Social Research, and a Theory of Action. American Journal of Sociology 91: 1309-1335.

Comey J. (2007) HOPE VI'd and On the Move. Washington, DC: Urban Institute.

Comey J and Popkin SJ. (2004) An Improved Living Environment?: Housing Quality Outcomes for HOPE VI Relocatees, Washington, DC: Urban Institute.

Culbreath A and Wilkinson JE. (1999) No Section 8 Policies: Combating Landlords' Resistance to Renting to Section 8 Recipients. Clearinghouse Rev. 33: 607.

De Bruijn A. (2011) Ook oude bomen moeten soms worden verplant Een onderzoek naar gedwongen verhuizingen onder ouderen als gevolg van herstructurering. Faculty of Geosciences. Utrecht: Universiteit Utrecht.

Doff W and Kleinhans R. (2011) Residential Outcomes of Forced Relocation Lifting a Corner of the Veil on Neighbourhood Selection. Urban Studies 48: 661-680.

Donaldson B and Johnston RJ. (1973) Intraurban Sectoral Mental Maps: Further Evidence from an Extended Methodology. Geographical Analysis 5: 45-54.

Ekström M. (1994) Elderly people's experiences of housing renewal and forced relocation: Social theories and contextual analysis in explanations of emotional experiences. Housing Studies 9: 369-391.

Ellen IG and Turner MA. (1997) Does neighborhood matter? Assessing recent evidence. Housing policy debate 8: 833-866.

Fishman R. (2000) The American metropolis at century's end: Past and future influences. Housing policy debate 11: 199-213.

Florida RL and Feldman MMA. (1988) Housing in US Fordism*. International Journal of Urban and Regional Research 12: 187-210.

Friedrichs J. (1993) A Theory of Urban Decline: Economy, Demography and Political Elites. Urban Studies 30: 907-917.

Friedrichs J, Galster G and Musterd S. (2003) Neighbourhood effects on social opportunities: the European and American research and policy context. Housing Studies 18: 797-806.

Galster G. (2003) Investigating behavioural impacts of poor neighbourhoods: towards new data and analytic strategies. Housing Studies 18: 893-914.

Galster G. (2007) Neighbourhood Social Mix as a Goal of Housing Policy: A Theoretical Analysis. International Journal of Housing Policy 7: 19-43.

Galster G. (2012) The Mechanism(s) of Neighbourhood Effects: Theory, Evidence, and Policy Implications. In: van Ham M, Manley D, Bailey N, et al. (eds) Neighbourhood Effects Research: New Perspectives. Springer Netherlands, 23-56.

Gilbert P. (2009) Social stakes of urban renewal: recent French housing policy. Building Research & Information 37: 638-648.

Goetz EG. (2002) Forced Relocation vs. Voluntary Mobility: The Effects of Dispersal Programmes on Households. Housing Studies 17: 107-123.

Goetz EG and Chapple K. (2010) You gotta move: advancing the debate on the record of dispersal. Housing policy debate 20: 209-236.

Hedman L and van Ham M. (2012) Understanding Neighbourhood Effects: Selection Bias and Residential Mobility. In: van Ham M, Manley D, Bailey N, et al. (eds) Neighbourhood Effects Research: New Perspectives. Springer Netherlands, 79-99.

Hoyt H. (1943) The structure of American cities in the post-war era. American Journal of Sociology: 475-481.

HUD. (2006) Programs of HUD, Major Mortgage, Grantss, Assistance and Regulatory Programs. Department of Housing and Urban Development.

HUD. (2010) HOPE VI Revitalization Grants. Washington, DC: Department of Housing and Urban Development.

Imbroscio D. (2011) Beyond mobility: The limits of liberal urban policy. Journal of Urban Affairs 34: 1-20.

Immergluck D. (2004) Credit to the community: Community reinvestment and fair lending policy in the United States, New York: ME Sharpe Inc.

Jackson KT. (1987) Crabgrass Frontier: The Suburbanization of the United States, New York: Oxford University Press, USA.

Jacobs J. (1993) The Death and Life of Great American Cities, New York: The Modern Library.

Jargowsky PA. (1997) Poverty and place: Ghettos, barrios, and the American city, New York: Russell Sage Foundation Publications.

Joseph ML and Chaskin RJ. (2012) Mixed-income developments and low rates of return: insights from relocated public housing residents in Chicago. Housing policy debate 22: 377-405.

Katz LF, Kling JR and Liebman JB. (2001) Moving to opportunity in Boston: Early results of a randomized mobility experiment. The Quarterly Journal of Economics 116: 607-654.

Kearns A and Mason P. (2013) Defining and Measuring Displacement: Is Relocation from Restructured Neighbourhoods Always Unwelcome and Disruptive? Housing Studies 28: 177-204.

Kelley P. (2009) The Kennedy Homes, an Amreican Tragedy, New York: iUniverse.

Kleinhans R. (2003) Displaced but still moving upwards in the housing career? Implications of forced residential relocation in the Netherlands. Housing Studies 18: 473-499.

Kleinhans R and Kearns A. (2013) Neighbourhood Restructuring and Residential Relocation: Towards a Balanced Perspective on Relocation Processes and Outcomes. Housing Studies 28: 163-176.

Kleinhans R, Priemus H and Engbersen G. (2007) Understanding social capital in recently restructured urban neighbourhoods: two case studies in Rotterdam. Urban Studies 44: 1069-1091.

Kleinhans R and Van der Laan Bouma-Doff W. (2008) On priority and progress: forced residential relocation and housing chances in Haaglanden, the Netherlands. Housing Studies 23: 565-587.

Kling J, Liebman J, Katz L, et al. (2004) Moving to opportunity and tranquility: Neighborhood effects on adult economic self-sufficiency and health from a randomized housing voucher experiment. Faculty Research Working Papers. Harvard University.

Lees L. (2008) Gentrification and social mixing: towards an inclusive urban renaissance? Urban Studies 45: 2449-2470.

Manzo LC, Kleit RG and Couch D. (2008) “Moving Three Times Is Like Having Your House on Fire Once”: The Experience of Place and Impending Displacement among Public Housing Residents. Urban Studies 45: 1855-1878.

Massey DS and Denton NA. (1993) American apartheid: Segregation and the making of the underclass: Harvard University Press.

McClure K. (2005) Rent burden in the housing choice voucher program. Cityscape: 5-20.

Meerts A, Kleinhans R and Posthumus H. (2011) Choice within limits. How forced relocation affects residents’ dwelling searches and trade-offs in dwelling choice. Enhr Conference 2011. Toulouse.

Mendenhall R, DeLuca S and Duncan G. (2006) Neighborhood resources, racial segregation, and economic mobility: Results from the Gautreaux program. Social Science Research 35: 892-923.

Münch S. (2009) “It’s all in the mix”: constructing ethnic segregation as a social problem in Germany. Journal of Housing and the Built Environment 24: 441-455.

Musterd S. (2003) Segregation and integration: A contested relationship. Journal of Ethnic and Migration Studies 29: 623-641.

Musterd S and Andersson R. (2005) Housing mix, social mix, and social opportunities. Urban Affairs Review 40: 761-790.

Musterd S, De Vos S, Das M, et al. (2012) Neighbourhood composition and economic prospects: a longitudinal study in the Netherlands. Tijdschrift voor economische en sociale geografie 103: 85-100.

Musterd S and Ostendorf W. (2008) Integrated urban renewal in The Netherlands: a critical appraisal. Urban Research & Practice 1: 78-92.

Musterd S and Van Kempen R. (2007) Trapped or on the springboard? Housing careers in large housing estates in European cities. Journal of Urban Affairs 29: 311-329.

Newman K and Wyly EK. (2006) The Right to Stay Put, Revisited: Gentrification and Resistance to Displacement in New York City. Urban Studies 43: 23-57.

Newman O. (1972) Defensible space, New York: Macmillan.

O+S Amsterdam. (2002) Stadsdelen in Cijfers 2002. Amsterdam: Dienst Onderzoek en Statistiek.

O+S Amsterdam. (2006) Amsterdam in Cijfers 2006. Amsterdam: Dienst Onderzoek en Statistiek.

O+S Amsterdam. (2007) Amsterdam in Cijfers 2007. Amsterdam: Dienst Onderzoek en Statistiek.

Ostendorf W, Musterd S and De Vos S. (2001) Social Mix and the Neighbourhood Effect. Policy Ambitions and Empirical Evidence. Housing Studies 16: 371-380.

Pattillo M. (2009) Investing in poor black neighborhoods ‘as is’. In: Turner MA, Popkin SJ and Rawlings L (eds) Public housing and the legacy of segregation. Washington DC.: The Urban Institute Press, 31-46.

Peach C. (2009) Slippery Segregation: Discovering or Manufacturing Ghettos? Journal of Ethnic and Migration Studies 35: 1381-1395.

Phillips D. (2009) Minority Ethnic Segregation, Integration and Citizenship: A European Perspective. Journal of Ethnic and Migration Studies 36: 209-225.

Pinkster FM. (2008) De sociale betekenis van de buurt: Een onderzoek naar de relatie tussen het wonen in een arme wjik en sociale mobilieteit, Amsterdam: Amsterdam University Press.

Popkin SJ. (2000) Searching for Rental Housing with Section 8 in Chicago Region.

Popkin SJ, Katz B, Cunningham MK, et al. (2004) A decade of HOPE VI: Research findings and policy challenges. Washington, DC: Urban Institute.

Popkin SJ, Levy DK and Buron L. (2009) Has HOPE VI transformed residents' lives? New evidence from the HOPE VI panel study. Housing Studies 24: 477-502.

Popkin SJ, Levy DK, Harris LE, et al. (2002) HOPE VI Panel study: Baseline report: Final report. Washongton, DC: Urban Institute.

Popkin SJ, Rosenbaum JE and Meaden PM. (1993) Labor Market Experiences of Low-Income Black Women in Middle-Class Suburbs: Evidence from a Survey of Gautreaux Program Participants. Journal of Policy Analysis and Management 12: 556-573.

Porter ME. (1995) The competitive advantage of the inner city. The city reader: 274-286.

Posthumus H. (2013) Displacement myths: The real and presumed effects of forced relocations resulting from urban renewal. The Hague: Universiteit Utrecht.

Posthumus H, Bolt G and Kempen R. (2013a) Urban Restructuring, Displaced Households and Neighbourhood Change: Results from Three Dutch Cities. In: van Ham M, Manley D, Bailey N, et al. (eds) Understanding Neighbourhood Dynamics. Springer Netherlands, 87-109.

Posthumus H, Bolt G and van Kempen R. (2013b) Why do Displaced Residents Move to Socioeconomically Disadvantaged Neighbourhoods? Housing Studies: 1-22.

Priemus H. (1996) Recent Changes in the Social Rented Sector in The Netherlands. Urban Studies 33: 1891-1908.

Priemus H. (2006) Regeneration of Dutch post-war urban districts: the role of housing associations. Journal of Housing and the Built Environment 21: 365-375.

Priemus H, Boelhouwer P and Kruythoff H. (1997) Dutch Urban Policy: a promising perspective for the big cities. International Journal of Urban and Regional Research 21: 677-690.

Rolland M. (2009) HUD: Glen Spring Manor closing due to poor living conditions. The Gainesville Sun. Gainesville.

Rosenbaum JE. (1991) Black pioneers—do their moves to the suburbs increase economic opportunity for mothers and children? Housing policy debate 2: 1179-1213.

Rosenbaum JE. (1995) Changing the geography of opportunity by expanding residential choice: Lessons from the Gautreaux program. Housing policy debate 6: 231-269.

Rosenbaum JE, Reynolds L and Deluca S. (2002) How Do Places Matter? The Geography of Opportunity, Self-efficacy and a Look Inside the Black Box of Residential Mobility. Housing Studies 17: 71-82.

Rowlands R, Musterd S and Van Kempen R. (2009) Mass housing in Europe: multiple faces of development, change and response: Pallgrave Macmillan.

Salama JJ. (1999) The redevelopment of distressed public housing: early results from HOPE VI projects in Atlanta, Chicago, and San Antonio. Housing policy debate 10: 95-142.

Schwartz AF. (2010) Housing policy in the United States, New York: Routledge.

Schwartz J. (1993) The New York approach: Robert Moses, urban liberals, and redevelopment of the inner city: Ohio State University Press.

Slob A, Bolt GS, van Kempen R, et al. (2008) Na de sloop: waterbedeffecten van gebiedsgericht stedelijk beleid: Nicis Institute.

Smith N. (1996) The new urban frontier: gentrification and the revanchist city Routledge.

Smith N. (2002a) New Globalism, New Urbanism: Gentrification as Global Urban Strategy. Antipode 34: 427-450.

Smith RE. (2002b) Housing choice for HOPE VI relocatees: Final report. Washington, DC: Urban Institute.

Tieskens KF and Musterd S. (2013) Displacement and urban restructuring in Amsterdam, following relocatees after demolition of social housing. Journal of Urban Research and Practice.

Turkington R, van Kempen R and Wassenberg F. (2004) High-rise housing in Europe: Current trends and future prospects. In: Studies ORIfHaM (ed) Housing and Urban Policy Studies. Delft: DUP Science.

Turner B and Whitehead CME. (2002) Reducing Housing Subsidy: Swedish Housing Policy in an International Context. Urban Studies 39: 201-217.

Turner MA, Popkin SJ and Rawlings L. (2009) Public housing and the legacy of segregation: The Urban Insitute.

Uitermark J. (2003) 'Social Mixing' and the Management of Disadvantaged Neighbourhoods: The Dutch Policy of Urban Restructuring Revisited. Urban Studies 40: 531-549.

Uitermark J, Duyvendak JW and Kleinhans R. (2007) Gentrification as a governmental strategy: social control and social cohesion in Hoogvliet, Rotterdam. Environment and Planning A 39: 125-141.

Van Beckhoven E. (2007) Herstructurering succesvol? . Factoren die bijdragen aan het succesvol verlopen van herstructureringsprojecten. Utrecht: Stade Advies BV.

Van Beckhoven E, Bolt G and Van Kempen R. (2009) Theories of Neighbourhood Change and Decline: Their Significance for Post-WWII Large Housing Estates in European Cities. In: Rowlands R, Musterd S and Van Kempen R (eds) Mass Housing in Europe: Multiple Faces of Development Change and Response. London: Palgeave Macmillan, 20 - 50.

Van Bergeijk E, Kokx A, Bolt G, et al. (2008) Helpt Herstructurering? Effecten van stedelijke herstructuring op wijken en bewoners, Delft: Eburon.

Van der Schaar J. (1983) Woningcorporaties en de non-profit verhuur. Working Papers. Delft: DUP.

Van Ham M and Manley D. (2010) The effect of neighbourhood housing tenure mix on labour market outcomes: a longitudinal investigation of neighbourhood effects. Journal of Economic Geography 10: 257-282.

Van Kempen R. (2000) Big cities policy in the Netherlands. Tijdschrift voor economische en sociale geografie 91: 197-203.

Van Kempen R and Priemus H. (2002) Revolution in social housing in the Netherlands: possible effects of new housing policies. Urban Studies 39: 237-253.

Visser K, Bolt G and van Kempen R. (2013) Urban Restructuring and Forced Relocations: Housing Opportunities for Youth? A Case Study in Utrecht, the Netherlands. Housing Studies 28: 294-316.

Wassenberg F, Turkington R and van Kempen R. (2004) High-rise housing estates in Europe. In: Turkington R, Van Kempen R and Wassenberg F (eds) High-Rise Housing in Europe. Delft: DUP Science, 1 - 14.

Wilson WJ. (1987) The truly disadvantaged: The inner city, the underclass, and public policy, Chicago: University of Chicago.

Yancey WL. (1972) Architecture, interaction, and social control: The case of a large-scale housing project. Environment and Behavior 3: 3.

100