The Stockholm Estates—A Tale of the Importance of Initial Conditions, Macroeconomic Dependencies, Tenure and Immigration

Abstract

In this chapter, we define the concept of housing estates in the Swedish context and provide some information about Stockholm and the historical background to the construction of post-war housing estates. The core research question will then be whether and to what extent initial conditions play a key role for later developments of an estate. Approaching this question, we first provide a statistical overview of developments from 1990 onwards, and then use examples from two estates in Stockholm, one built in the mid-1960s (Bredäng) and one built a few years later (Rinkeby), which now have similar problems of ethnic and socio-economic segregation but have arrived at this situation through very different trajectories. We will analyse these trajectories and identify the key moments leading up to present day convergence in terms of the social challenges facing the estates. Until 1990, the socio-economic situation in the 49 estates we analysed was not very different from the average situation in the Stockholm region. However, the economic crisis of the early 1990s had profound effects and initiated diverging trajectories where some estates continued to do well while others did not. We explain this diverging development with reference to tenure composition, geographical context and building period, all important for also understanding the geography of refugee settlement. This set of explanations is based both on the more structural analysis of all 49 estates and on the more detailed study of our two cases. We end the chapter with a discussion of 40 years of recurrent interventions and of how contemporary challenges are perceived and addressed.

1 Introduction

In Sweden, the label ‘large housing estate’ has primarily been associated with multifamily housing constructed during the so-called Million Programme, when one million dwellings were built in 10 years (1965–1974). During this period, the projects were larger and the industrial efficiency of the construction process further driven than in earlier decades, resulting in metropolitan areas in very large estates with little mixing in terms of housing tenure and dwelling sizes. The resulting estates were also the first, in the modern Swedish context, to be associated with problems of ethnic and socio-economic segregation.

In this chapter, we define the concept of housing estates in the Swedish context and provide a brief historical background. We give an account of the location and demographic and socio-economic characteristics of Stockholm County’s housing estates in 1990, a year when all estates we include in the analyses were completed. We then follow the estates forward to 2014 and document their heterogeneous development but also the emergence of concentrated poverty, declining employment levels and the role they have in residential sorting and segregation processes. These later developments are partly well-researched but there are important questions to be asked regarding the early years of the estates’ development— Why do many of the estates today display similar kinds of problems? Is the answer to be found in issues related to design and architecture, in their location, in their tenure structure or in wider contextual matters such as economic and political developments? The core research question will be whether and to what extent initial conditions play a key role for later developments of an estate. Approaching this question, we use examples from two estates in Stockholm, one built mostly before (Bredäng) and one in the middle of the Million Programme period (Rinkeby), which now both have problems of ethnic and socio-economic segregation but have arrived at this situation through very different trajectories. We will analyse these trajectories and identify the key moments leading up to the present situation, in terms of the social challenges facing many of the estates.

Using longitudinal individual-level data from a combination of census and register data going back to 1960, we will trace the population development of the two cases back to about 5 years after construction, showing who the early residents were and where they came from, and how this has shaped later developments of the estates. Accounts of the post-1990 period are primarily based on data from the Geosweden dataset, comprising longitudinal and geocoded annual microdata of the entire population of Sweden. These data are owned by the Institute for Housing and Urban Research, Uppsala University, but originate from a range of Statistics Sweden’s registers (population, income, education and employment (LISA), real estate and property, dwellings registry).

As in many other countries, large-scale housing estates in Sweden have been much discussed politically since the 1970s and discussions have intensified as social conditions have deteriorated in many of them, triggering housing and urban policy responses and area-based interventions. We allocate one section of the paper to these interventions and reflect upon the current challenges facing Stockholm’s estates. This section is partly based on field visits and interviews in the two estates selected for the long-term analysis.

2 Housing Estates in Sweden—A Background

Housing developments and planning for housing show broader international as well as country-specific characteristics. While architectural and design ideas have travelled easily across national borders, the social and political national realities always put severe constraints on both the when and how questions related to the physical manifestations of such ideas. In Sweden, as in many other countries, the philanthropic reactions to the slum housing created in the wake of industrialisation and urbanisation in the nineteenth century gradually shifted into politically induced change in regulations and planning practices. Step by step, the politically organised working class gained influence over housing developments and city planning. By the mid-1930s the Social Democratic Party (SAP) controlled government and did so for 40 years thereafter. Although some key organisations—like municipal rental housing and cooperative housing companies—were sometimes established already before World War II, it was not until after the war that they became instrumental for modernising housing for the working class. The State developed the financial means and laws regulating planning, building and rent setting, but it was the municipalities—most of them governed by Social Democrats—that were to become the key actors in planning for the new era.

Early on, Stockholm planning was internationally recognised for contributing with new visions regarding post-war urban development. The Stockholm City General Plan of 1945–1952, designed by Sven Markelius, envisioned ‘a high-capacity highway network, designed especially to provide circumferential trips (…), to be supplemented by a brand-new subway system’ (Hall 1988, 308). As pointed out by Hall, the new outer suburban units—Vällingby of 1950–4, Farsta of 1953–61, and Skärholmen of 1961–1968—were developed as ABC units, meaning they were supposed to contain housing, workplaces and service. Vällingby and Farsta were well received but later developments faced a range of criticism.

Art Historian Lisbeth Söderqvist wrote her thesis in 1999 about ‘The Record Years’ (1960–1975) with a focus on multifamily large-scale housing in Sweden. She is not the only one to study this period of booming construction activity but her analyses of three particular cases in the Stockholm region convincingly sum up some key aspects we need to understand as background for approaching the Stockholm case. As housing shortage, along with poor conditions in most of the existing housing stock, were the main rationales for stepping up building activity after the war, the actual design of new housing and neighbourhoods was inspired by developments in the 1930s, such as the quest for modernism displayed not least at the influential 1930 Stockholm Exhibition, where a plea for acceptance of functionalism, standardisation and mass production as a cultural change was formulated (Pred 1995). In 1945, only 21% of the country’s dwellings had a bathroom or a shower. This figure rose to 53% in 1960 and 87% in 1975 (Söderqvist 1999, p. 11). Modernistic design ideas along with a strong emphasis on technological advancements and industrial building methods formed the context of developments converging towards building massive amounts of new housing.

A State housing construction investigation was put in place in 1959 (Bostadsbyggnadsutredningen) and the experts estimated the need of new housing to be in the range of 1.5 million units for the 1960–1975 period (SOU 1965, p. 32). As noted by Söderqvist (1999, p. 12), the forthcoming politically formulated project ‘The Million Programme’—meaning the construction of one million new dwellings between 1965 and 1974—was therefore not dramatic, with these plans and forecasts already in place. All 1.5 million dwellings were also realised, truly making the ‘Record Years’ an appropriate label.

For a country with eight million people in 1970, the volume of new housing was indeed remarkable and it took more than new production methods to make it happen. The 1960s saw a range of economic reforms aiming to rationalise the building process. State investment funds for ‘element factories’ were one such new feature. Planning routines were also changed so that big projects (those with more than 1,000 new dwellings) could get a pre-decision from the planning authorities. The concept of ‘total entrepreneur’—i.e. giving one company the responsibility for both project planning and construction—was introduced, further escalating the industrialisation of construction activity. Realising the new modernist housing projects was not possible on already built-up land, and the fact that Stockholm and other cities from the late nineteenth century onwards had bought substantial amounts of land on the outskirts provided another key factor in shaping the necessary conditions for large-scale housing construction. According to Söderqvist (1999, p. 161), municipal planning authorities had to develop general plans but these did not determine the real outcome. Instead, what later has become known as urban governance and public–private partnerships (‘negotiating planning’) characterised planning practices. Private actors had substantial influence while the State, primarily represented by the State Board of Construction and the provincial architect at the County administration, were rather weak and typically critical of high-rise developments (Söderqvist 1999, 164).

With the quantitative goals achieved by the mid-1970s, the debate immediately turned to questioning the more qualitative aspects of the new housing—the high-rise buildings, density, the repetitive character of buildings and layout, poor services and so on. Some blamed architects, others blamed developers and builders and most blamed politicians for failing to put the individual resident in focus. Rådberg (1988, 1997) argues that the modernist visions from the 1920s and 1930s, inspired by Le Corbusier and Walter Gropius but with strong involvement of Swedish modernist architects like Sven Markelius were realised in the 1960s and that architects had a great deal of responsibility for what took place. Markelius and a handful of other Swedish architects were affiliated with CIAM (Les Congreès Internationaux d’Archetecture Moderne), an organisation established in 1928 for discussing and promoting modern architecture (Söderqvist 2008, pp. 31–32). In a U.S.–Swedish comparative study, Popenoe (1977) argues that the construction of the subway (the political decision made in 1941) and the location of subway stations had a profound importance for the scale and density of new housing estates. Concentrating consumers to accommodate the economic interests of service providers also played a key role. It is clear that much of this makes sense but also that most of the criticism came later, after the completion of the estates.

Söderqvist concludes her thesis by stating: ‘My interpretation (…) is (….) that the influence of state authorities on the design of the criticised large-scale housing estates was negligible. It was municipalities, developers, and building companies that, together with the architects, shaped the housing environments’ (Söderqvist 2008, 185). However, sociologists Franzén and Sandstedt (1981) arrive at a somewhat different conclusion in terms of the state’s overall influence. Applying Habermas’ theory on the role of the state and the public sphere, they argue that the emergence of the modern social state meant a transition from a separation of the state and the private sphere to a situation where the state took over many more duties than earlier, and also that different interest groups gained more influence over the state. Their focus is on neighbourhood planning and housing construction but parallel developments took place in other policy fields, such as traffic planning and the control over retail businesses and location. This massive state expansion was made possible by a sustained period of post-war economic growth, but the notion of community planning did in fact not change much from before.

We conclude this background by noting that many circumstances aligned to produce foundations for and realisation of large-scale housing in Sweden. Therefore, explanations are offered at different levels of abstraction. It is clear that after the ‘Record Years’, overcrowding was more or less abolished, most people lived in high-standard modern housing and the debate soon shifted focus to issues related to those who were housed in the new housing estates. It is also with these issues we next engage.

3 The Stockholm Estates

We define a housing estate in the Stockholm context as a major concentration of multifamily housing constructed in the 1951–1990 period. We make our selection in 1990, at a time when all larger estates were built (only one was added after 1980). Our study area is Stockholm County and we define a ‘major concentration’ as a statistical area (SAMS) with at least 1,000 dwellings, where 75% or more of the residents are living in multifamily housing built during these four decades. We believe the definition is wide enough to capture the essence of the meaning of a large estate in our context but also narrow enough to avoid including areas that are seldom considered as being of an estate character. Naturally, all definitions of the concept ‘housing estate’ are somewhat arbitrary and the bulk of Swedish housing studies tend to focus instead on housing constructed during the Million Programme period. Although the Million Programme is a shorthand in the Swedish debate for discussing a range of housing related problems, it is less appropriate in this context for two reasons: (1) it is too narrow time-wise considering that housing construction was already industrialised and large-scale towards the end of the 1950s and that a similar type of planning (albeit somewhat adjusted) and housing construction continued throughout the 1970s and into the 1980s; (2) the Million Programme comprised not only multifamily housing but also a large proportion suburban single-family housing, produced fairly large-scale but without the characteristics of housing estates.

Although we use the SAMS neighbourhood division for identifying the estates, the data we provide refers to the multifamily section of each selected neighbourhood, i.e. the estates proper. When applying our definition, we find about 50 estates. We exclude one estate because it had an extended construction period (not resulting in a coherent housing estate). There are also a few estates further out in Stockholm County that we decided not to include. Some that are included from the 1950s are of a transitional character, in the sense that building methods were not as industrialised and design sometimes more varied than were later to follow in the 1960s. Our final selection of 49 estates is displayed in Fig. 16.1, and key data provided in Table 16.1 show the location and basic characteristics of Stockholm County’s housing estates as of 1990. Many of the estates were built over a period of two to three years and in some cases, the construction period crossed a decade shift (such as 1960s/1970s). The period of construction indicated in the map refers to the decade when at least half of the dwellings were completed and occupied. Most of the estates are located 10–20 km from the city centre and along subway lines and main highways. There is a tendency of later built projects to be located somewhat further away from the centre. Administratively, the 49 are dispersed over 10 municipalities (out of 26 in the county), but almost half of them (22) are located within the borders of Stockholm city, including our two cases Bredäng and Rinkeby. The proportion residing in these large housing estates varies across municipalities in 1990, from 36% in Sundbyberg to 11% in Täby and 24% in Stockholm.

Fig. 16.1

Location of large housing estates in Stockholm by period of construction and population size in 1990

Table 16.1 Characteristics of the 49 selected housing estates in 1990

3.1 Key Data on Housing Estates in 1990 with an Overview of Subsequent Changes Until 2014

As we will focus in particular on two of these 49 estates, data is provided separately for these two (Bredäng and Rinkeby). All 49 had about 264,000 multifamily-housed residents in 1990 (circa 16% of the county population and 25% of all living in multifamily housing) with an average of 5,400 people. Our two cases are in the upper range population-wise (7,000 and 13,000, respectively in 1990) but they share many housing and demographic characteristics with other estates from the same time period: most estates are heavily dominated by rental housing, and except for the 1970s estates, buildings are not very high but densities are still above mean for multifamily housing in the county (255 residents per hectare compared with 306 for our selection). Furthermore, three-room dwellings dominate, females are slightly over-represented and the immigrant proportion is clearly above mean values for neighbourhoods in the Stockholm region.

There are also differences between estates, related partly to the period of construction; there are fewer high-rise buildings from the 1950s and housing estates from this early period have fewer immigrants, and related to this, an older population and also smaller households. Overcrowding is very much discussed today but was not on the agenda in early 1990s, and on a group level, housing estates were only slightly more overcrowded than the wider Stockholm region. Estates from the 1960s and 1970s, however, did indeed house many that had to share a room with at least two other persons. These differences also apply for our two cases and Rinkeby’s higher proportion of overcrowded households, younger population (including students), more immigrants and poorer socio-economic position are clearly displayed; the latter two indicators, in particular, stand out when compared to most other housing estates, including Bredäng and other early built estates.

We will extend our study period both backwards and forwards in time, asking: (1) Whether the socio-economic and ethnic differences visible in 1990 were established already when the estates were newly constructed, and (2) Whether developments after 1990 have converged towards Rinkeby’s position and if so, why? The early period will be studied in detail using our two case study areas. The more general development after 1990 is summarised in Table 16.2 and clearly indicates a more problematic situation today than in 1990.

Table 16.2 Characteristics of the 49 housing estates and Stockholm County, 2002 and 2014

It is well known in Sweden that the large housing estates have faced a quarter-century of increasing signs of social exclusion and stigmatisation. Some argue that this is a consequence of increasing racialisation in the wake of the influx of non-European migrants (Molina 1997; Ericsson et al. 2002), while others have tended to primarily stress the geographically uneven effects of declining employment levels, housing policy change, and restructuring of the welfare state (Andersson and Hedman 2016; Andersson and Kährik 2015). All these explanations have relevance and isolating the effect of each component is indeed a very complex endeavour.

Table 16.2 follows the same general format as Table 16.1, but for brevity, we display data only for some key variables. There has been some post-1990 multifamily housing construction in 33 of the 49 estates but these additions explain less than half of the population increase (an additional 59,000 inhabitants in total). For Bredäng and Rinkeby, about one-third of the population increase from 1990 to 2014 comprises people living in post-1990 buildings. While overcrowding is generally not a problem in the Swedish housing market (and it has been further reduced in Stockholm County since 1990), it can be in some housing estates and certainly if an estate has seen in-migration of recently arrived refugee families. Rinkeby illustrates this; by 2014, close to one in four residents lived in dwellings where the available space was less than 0.5 rooms per person (kitchen excluded).

Other housing developments should also be noted. Stockholm has been a showcase for the liberal-conservative political parties’ aims of introducing more market dynamics into the housing market. The primary instrument has been conversion of public rental housing to cooperative housing, and due to a parallel conversion—albeit not politically induced—of private rental housing to cooperative housing, Stockholm County has seen its stock of rental housing substantially reduced. While the overall reduction in Stockholm County is from 45 to 32% it has been very uneven; much more radical in Stockholm inner city while housing estates from the 1960s and 1970s (including Rinkeby) have been less affected (see Table 16.2). These tenure conversions have put extra pressure on the unaffected rental-dominated estates, effects that Andersson and Turner (2014) sum up using the concept of residualisation. Fewer rental units now have to cater to expanding numbers of people without much choice in the competitive housing market.

Besides some of these housing dynamics, three general developments stand out: (1) a radical reduction of the proportion of low-education persons in the region and across the estates; (2) a very fast increase of the immigrant population, following a general trend in the region (although the increasing proportions have been substantial in the housing estates, not least in estates with average proportions in 1990—once again the early and late-built estates such as Bredäng) and (3) the negative employment change affecting many estates from all time periods. Most of the reduction occurred in the early 1990s but recovery has been modest since then (compare 2002 and 2014 in Table 16.2 and see Fig. 16.2).

Fig. 16.2

Employment rates (%) for residents in selected housing estates, 1990–2014. Note This figure is based on annual data for the 49 statistical units (SAMS) that frame the selection of housing estates. Some of the SAMS units thus contain housing other than the housing estate proper. It is clear that most estates follow the same trends (macroeconomic dependency) but also that developments in the estates radically diverged in the early 1990s and that the substantial spread established during the crisis remained for 15 years. Some convergence can be seen lately and preliminary data indicates that this has likely continued through 2017

Large housing estates may to a certain extent be similar in design, construction methods and building material but they exhibit big differences along the socio-economic dimension, for example, when it comes to employment. While some estates, like Rinkeby, now have employment rates around 50%, some, in fact, do better than the county average of 79%. These better performing estates, however, contain entirely or predominantly cooperative tenure, underscoring the key importance of tenure form, and primarily not architectural design, for post-1991 diverging developments.

The overall employment gap between cooperative dwellers and rental dwellers in 2014 was 77 versus 63% for the 49 estates, and housing estates with higher than estate average employment rate (above 66%) had 45% cooperative tenants while those below had just 15%. This explains a great deal of the inter-estate difference and sorting and, as argued, also explains where most immigrants end up. Such tenure-related employment differences did not exist when the estates were new, mainly because cooperative housing was not a market commodity until 1969 (Svensson 1998; Sørvoll 2013) and it took time and increasing housing shortage to establish the currently high entry costs associated with this tenure form.

4 Converging Fates? Bredäng and Rinkeby

The two estates chosen for the case study illustrate how initial conditions at the time of construction can shape the development of the estate over decades but also how developments in society at large, such as the economic recession of the early 1990s and changes in role the municipal housing sector as a whole will put pressure also on the more successful estates. The first case, Bredäng, was built in the early 1960s when local demand for housing was very high and new housing was considered most welcome. When the second case, Rinkeby, was built only about 5 years later, conditions were radically different. Due to extensive construction in the intervening years, the local housing shortage was rapidly transforming into a housing surplus, and media attention had turned towards the more problematic aspects of industrialised housing construction and the resulting estates. Rinkeby has been one of the prime symbols of ‘the segregated estate’ in the public debate ever since it was built (Ericsson et al. 2002) while for Bredäng problems of socio-economic and ethnic segregation began to be noticed only in the mid-1990s.

First, a note on the data used in this section. Data for the period prior to 1990 comes from a combination of registry and census data, where the latter is only available for every fifth year. By using the property codes, we have been able to trace the multifamily properties of these two estates back to about five years after their construction. The choice of variables has mainly been guided by data availability; very few variables are available throughout the entire period from 1970 to 2014.

Bredäng was built in 1962–1967 and can be regarded as Stockholm’s first large-scale housing estate, in the sense that the whole residential area of Bredäng was planned as one architectural unit (Fig. 16.3). Everything, from the street network and green spaces to the placement, form and volume of the buildings was regulated in the physical plan of the area. One governing principle in the physical planning of the estate was to let the landscape of the area guide the placement of streets and houses, and to keep as much as possible of the natural landscape. In order to achieve this, much of the housing had to be concentrated in large high-rise buildings, though there are also a number of low-rise multifamily houses. The result is an estate with winding streets and lots of green space, but where the large-scale concrete high-rise buildings, placed in parallel along the streets, form the dominant visual feature of the estate. In terms of building practices, the construction of Bredäng also marks the beginning of the era of industrialised building. Pre-fabricated building elements like stairs, balconies and some wall elements were delivered from a cement factory, along with cement used for in-place casting (Rittsél 2000).

Fig. 16.3

Source Arild Vågen, Wikimedia Commons

Bredäng, 2014.

The rented multifamily part of the housing stock, in total 3,900 dwellings built 1963–1965, was mostly owned by municipal housing companies, though private housing companies were (and are) also represented. Multifamily housing was supplemented with about 300 single-family dwellings in terraced housing in 1965–1967, some rented and some owner-occupied. Bredäng centrum, the centre where commercial and public services and the metro station are located, was opened in 1965 (Rittsél 2000). Later additions and changes to the housing stock have until recently been few. Currently, however, several new additions are planned for the coming years. About 500 new dwellings, in both rental and cooperative multifamily housing, are already under way, and another 1,000–1,500 are awaiting decision in the planning process (Stockholms stad: Fokus Skärholmen http://vaxer.stockholm.se/omraden/skarholmen-fokus-skarholmen/). In the physical planning of these additions, the strategy has been to preserve the park-like character of the old, central parts of the estate and instead place new housing on the outskirts of the estate, especially to the east where the additions can help reduce the physical and mental gap between Bredäng and the small-scale single-family housing of neighbouring Mälarhöjden (Bredäng field visit).

In 1970, Bredäng had about 10,500 residents (Table 16.3). The age profile of the residents suggests a dominance of families with children. More than one-third (37%) were children, and adults between ages 20 and 49 made up another 50%. Only two percent of the residents were 65 years or older. About 92% were born in Sweden, which was slightly higher than the corresponding figure for the whole of Stockholm County. Further investigations into place of residence in 1960 of these early residents reveal that an overwhelming majority were ‘locals’. Excluding children born after 1960, about 83% were living in Stockholm city and another four percent in other parts of Stockholm county.

Table 16.3 Socio-economic and demographic composition of early residents of Bredäng (1970) and Rinkeby (1975), in comparison with Stockholm county

Sweden in the 1970s and early 1980s is known as a society characterised by a high degree of economic equality (OECD 2011) and is obvious in our data that shows very small income differences. Even so, available socio-economic indicators suggest that the early residents of Bredäng were rather well off. The average labour income was about six percent higher than the county average, and employment rates were higher, though the latter might be explained by the large proportion of young households, given higher employment rates among younger women than older.

When Rinkeby was built 1968–1971, together with neighbouring estate Tensta, planning ideals had shifted towards more urban forms (Fig. 16.4). That, and criticism in the public debate towards high-rise buildings resulted in lower but more densely placed buildings, interspersed by smaller, artificially landscaped green spaces (Söderström 2003). Originally, 13 different landlords, both public and private, were present on the estate, and 600 of nearly 4,900 dwellings in total were let as student flats.

Fig. 16.4

Source Johan Fredriksson, Wikimedia Commons

Rinkeby, 2014.

Rinkeby and Tensta have received a lot of criticism over the years, for their large-scale and lack of variation in terms of tenure and housing types. At the time of construction, the critics were mainly focusing on the hurried construction process, with residents moving in while neighbouring houses were still being built and shops and other services were still lacking (Lundén 1999). Rinkeby centrum, with commercial and public services, was finished in 1971, but the accompanying metro station did not open until 1975 (Söderström 2003). The dense character of the estate has made further construction difficult, and very few additions have been made to the original housing stock. A small area of terraced housing was added in the early 2000s, but otherwise, all changes are alterations in the original housing stock. Worth mentioning is the conversion in 1993 of all student flats into ordinary private rental housing.

About 5 years after the construction, Rinkeby had 12,700 residents (Table 16.3). The age profile was exceptionally young, with nearly 80% younger than 35, no doubt partly due to the presence of student flats on the estate. The population was arguably even more exceptional in terms of ethnic minority presence, with nearly one out of three residents born abroad. Data on country of birth reveal that the largest immigrant groups were from Finland (12% of the residents), Greece and Turkey (just over 4% each). The socio-economic situation of the residents was rather good, with an employment rate near that of the whole county, and labour income only slightly below the Stockholm average. Thus, our data suggest that a large part of the early residents of Rinkeby were ‘typical immigrants’ of that time, i.e. labour immigrants from Finland and southern Europe. In contrast to Bredäng’s early residents, the share of locals was very low. Only 47% of the 1975 residents were living in Stockholm county in 1965 (excluding children born after 1965).

So far, we have found a number of differences between the early residents of Bredäng in 1970 and those of Rinkeby in 1975. Though the Bredäng ‘founders’ had slightly higher average income, and their age profile was more mixed, the most striking difference had to do with the ethnic profile of the residents. And this aspect of the Rinkeby population was to become even more pronounced. The following 5 years saw a rapid decline in the number of residents with Swedish background and an equally rapid growth of the number of residents with foreign background (Fig. 16.5), resulting in a situation in 1980 where first- and second-generation immigrants constituted 60% of the residents. The transition continued rather rapidly until 1995 when 86% of the residents had foreign background. Since then, the number of residents with Swedish background has decreased at a much slower rate, from about 1,900 in 1995 to under 1,200 in 2014.

Fig. 16.5

Ethnic background of Rinkeby residents, 1975–2014

Although Bredäng can be considered ‘immigrant-dense’ now, with first- and second-generation immigrants constituting 72% of the population in 2014 (see Table 16.2), the transition process has been very different from that of Rinkeby. The transition started out much slower, and picked up pace only after 1990 (Fig. 16.6). Between 1990 and 2005 the share of residents with foreign background grew from 29 to 60%. In terms of the development of the number of residents with Swedish background, however, the process resembles that of Rinkeby, but in Bredäng the decline in the native Swedish population was not accompanied by a corresponding growth of the immigrant population during the first decades. Thus, the total number of residents decreased, and only started to grow again after 1990, when the number of immigrants began to grow at a rate that could match the decline in the Swedish population. As of 2014, there are no signs yet of a stabilisation of the number of residents with Swedish background, though the decrease was slower in 2010 to 2014 than in earlier periods.

Fig. 16.6

Ethnic background of residents of Bredäng, 1970–2014

The development over time of the socio-economic indicators (Figs. 16.7, 16.8, 16.9) give a vivid illustration of the general tendency in Swedish society of increasing socio-economic inequality, and the effects of the economic recession of the early 1990s, in particular, are very striking, with rapidly falling employment rates and widening income gaps. Both estates were hit hard, but the negative effects were worse in Rinkeby where conditions had already begun to diverge from county averages (Fig. 16.7). The 1995–2000 period saw a general recovery of employment rates, but in the two estates the recovery was not nearly large enough to match the decline during the recession. From 2000, employment rates have stabilised at around 45% in Rinkeby and between 60 and 65% in Bredäng.

Fig. 16.7

Employment rates of residents aged 20–64 in Bredäng, Rinkeby and Stockholm county, 1970–2014

Fig. 16.8

Average labour income of residents aged 20–64 in Bredäng, Rinkeby and Stockholm County, 1970 to 2014

Fig. 16.9

Average disposable income of residents aged 20–64 in Bredäng, Rinkeby and Stockholm county, 1980–2014

Widening gaps in employment rates are of course also visible in increasing differences in average labour incomes (Fig. 16.8, solid lines), from 1985 and onwards in Rinkeby and about 5 years later in Bredäng. But the income gaps between the residents of the estates and the county averages have also widened among those in employment (Fig. 16.8, dashed lines). Also in this respect, Rinkeby has been hit worse than Bredäng.

A further source of widening income gaps is visible also in the development of average disposable incomes (data available from 1980, see Fig. 16.9), where the county average starts to increase at” policy launched by the liberal-conservative government which meant lower taxes on labour a faster rate after 2005, most likely a result of the so-called “job tax deduction” policy launched by the liberal-conservative government which meant lower taxes on labour income. In addition, the property tax was abolished to be followed by a much lower annual fee.

In summary, both estates are now labelled ‘segregated’ in the public debate, but their trajectories towards the current situation have been very different. Rinkeby had a higher than average minority presence from the beginning and was labelled ‘immigrant-dense’ very early on. It is worth noting, though, that in the 1970s and early 1980s, when income differences were still small and refugee immigrants few, having a high proportion of foreign-born residents was not associated with poverty or deprivation. And as the socio-economic indicators have shown, income levels of the Rinkeby population followed the county averages at least until 1985. It was not until the economic crisis 1992–1993 that the association between immigrant-dense and socially deprived really became established, and this is clearly visible in both estates. As for Bredäng, the early 1990s marks the beginning of both socio-economic decline and ethnic transition. Socio-economic decline has not been as severe here as in Rinkeby, but it is uncertain how the socio-economic situation will be affected if ethnic transition continues and if labour market integration of refugee migrants remains poor.

5 Interventions and Current Challenges

We conclude our analysis of the early years by stating that ethnic segregation preceded socio-economic segregation, indicating that the downward spiral of decline is closely related to an overall integration failure (following a shift in origin countries as the migration type changed from labour market immigration to refugee immigration) which could be related to structural racism in combination with increasing human capital mismatch. The labour market has undergone dramatic shifts from Fordism to post-Fordism, and hence reduced the number of jobs available for people who are less educated and less experienced (affecting many immigrants but certainly also young natives without a higher education). The deteriorating situation was initially strongly triggered by the economic crisis in the early 1990s but recovery seems to have been slow or non-existent in many estates.

This, certainly, has not left political actors passive. On the contrary, the large housing estates have been the focus of a range of interventions starting soon after their completion and followed by others up to this date. Importantly, there have been no demolitions but otherwise, all kinds of measures have been used. Spells of no intervention are indeed few and brief and the format of this chapter makes it impossible to cover all types of interventions for this long time period. The most important ones are already well covered in earlier research, including in comparative European-funded research projects such as UGIS (Urban Governance, Inclusion and Sustainability), Restate (Restructuring Large housing estates in Europe) and Neighbourhood Governance—Capacity for Social Inclusion. (Andersson 1999, 2006; Andersson et al. 2010) provide analyses of the background, aims and outcomes of several of the State-funded initiatives and also references to other research (often published in Swedish) on the efforts to reverse the spiral of decline. A specific report within the Restate project summarises both physically and socially oriented state- and municipally funded efforts launched in the entire post-war era up to about 2004 (Öresjö et al. 2004; Lahti Edmark 2002)

A series of programmes were initiated, especially in the 1990s, and were to be followed by new initiatives in the new millennium. As pointed out, there have been no demolitions and most of the State and municipal interventions have been people-based rather than placed-based (physical) in character. Typically, these interventions focus on educational efforts (both in schools and for adults), improved matching of unemployed and labour demand, investments in health-related institutions, democratic participation and dialogue between residents and local authorities and supporting cultural life. Although often criticised for being time-limited project interventions, some have worked well. It has nevertheless proven difficult to break the underlying population dynamics, i.e. that more successful households move out to be replaced by newly arrived households (Andersson and Bråmå 2004). Rinkeby has been part of most interventions, more recently Järvalyftet, launched as a planning process in five housing estates in northwestern Stockholm city in 2007 and actively involving the city´s public housing companies. The aims were formulated in terms of ‘good housing in a more varied city environment’, ‘safety’, ‘improved education and language training’ and ‘more jobs and entrepreneurial activity’. This still-ongoing effort has a strong physical component, including major renovations of dwellings and maintenance of public space, but physical interventions sometimes include employment-related activities. The renovations, or ‘renovictions’ (renovation + eviction) as they are sometimes labelled, have in Stockholm and other Swedish cities been harshly criticised for forcing people to move and for striving for gentrification of neighbourhoods. It is argued that what is generally believed to be necessary renovations and energy efficiency improvements of the 50–60-year old estates, are in fact ‘overdone’ in order to increase rents beyond what is affordable for many households (Baeten et al. 2016).

6 Looking Ahead: What Is on the Agenda?

The ‘renoviction’ debate illustrates a growing problem in a context of abolished housing subsidies and a much less active State housing policy, but if a single word could condense and summarise today’s dominating public discourse regarding the poor and immigrant-dense estates, it is safety. Safety has always been a relevant aspect and priority for those living in neighbourhoods experiencing unemployment and relative poverty, but it is now on top of the political agenda and it is strongly related to debates on immigration, refugees, radical Islam, gang violence and terrorism. The police have their own list of the worst and most dangerous places in Sweden, covering 61 estates across the country and 24 in Stockholm County, and many of our selected 49 housing estates are on this list. The government operates its own list of Urban 15, i.e. 15 housing estates in need of special attention and interventions. Rinkeby is on both lists while Bredäng is only on the police list.

Most social scientists do of course realise that the objective conditions created by neo-liberal policies and widening socio-economic gaps in a more segregated city provide fertile ground for social unrest. Particular events such as widespread car burnings and clashes between young men and the police attract national and international media attention and contribute to putting safety issues on top of the agenda. Many local residents do of course share the demands for more police presence, increased surveillance and other safety measures. However, available research on crime and safety finds crime rates dropping over time (2008–2014) and an overwhelming majority of residents find their neighbourhood to be safe (Stockholms Stad 2014). The overall proportion of people in Stockholm city reporting feeling unsafe in their own neighbourhood dropped from 10 to 8% in 2008–2014, but with variations across neighbourhoods (proportions are more than twice as high in socially disadvantaged estates such as Rinkeby and Bredäng). Complaints about damaged property, littering and arson are more common in areas dominated by housing estates compared to elsewhere. One in five residents in Stockholm city housing estates reports having felt anxious about violence over the past year (for comparison, see Kemppainen 2017, analysing disorder and insecurity in Finnish suburban housing estates). Clearly, even if such rates have not increased over time they are not satisfying, and Stockholm city along with the police have intensified their crime prevention activities. In Rinkeby and adjacent estates, the work is led by a security strategist who works in close cooperation with physical planners in order to make public space safer. The introduction of CCTV in May 2017 is another feature intended to make it easier to combat crime and gather evidence when a crime is committed (Rinkeby field visit).

It would, however, take a much more comprehensive social and housing reform agenda to combat the spiral of decline and reverse social exclusion tendencies. Such an agenda will likely not be politically agreed upon, but most share the view that employment is key and that improving school education is decisive for long-term improvements. The current left-green-feminist political majority in Stockholm city launched a new effort in 2015—The Commission on Socially Sustainable Development (Kommissionen för ett socialt hållbart Stockholm)—where a more holistic approach is taken towards current challenges, focusing not only on large housing estates but on planning and local policies more broadly. Priorities are listed under four headings: ‘democracy and safety’, ‘work and income’, ‘housing and living environment’ and ‘childhood and education’. The commission works in close cooperation with the research community, but whether this new broader effort will result in changes on the ground remains to be seen.

It is clear that the severe housing shortage in Stockholm in some ways are seen by planners and politicians as an opportunity to also address problems related to the ‘enclave’ character of many estates. By means of new housing construction Rinkeby will have 600–1,000 new housing units with mixed tenure added to its housing stock. These will be located, for instance, on former parking sites in the estate but also on land freed up due to channelling passing traffic on the highway E18 underground and placing new housing above the tunnel. Other development plans aim at expanding housing on green land immediately south of the estate. Rinkeby will also be the location of a new large local police office employing approximately 300 people, serving all housing estates in northwestern Stockholm.

As mentioned earlier, new construction aiming at breaking down existing barriers to adjacent neighbourhoods is also planned near Bredäng. Breaking barriers by densification is a general ambition of politicians and of Stockholm city’s planning office but such densification will likely also be contested because it reduces green space and constrains existing leisure activities. Balancing the need for more housing with rearranging accessibility patterns and conserving green space is indeed a challenge in a rapidly expanding urban region like Stockholm.

It is natural to expect municipalities and property owners to try to address what is believed to be problematic housing and neighbourhood conditions. But if anything, our analyses reveal the key importance of broader economic and political developments for what has taken place over the past 25 years. Absent a clearly formulated policy on refugee reception—for instance, regional and local placement policies—migrants have tended to cluster into a limited number of rental-dominated large housing estates. The economic crisis of the early 1990s saw some of these estates fall to the bottom in the rankings of many key neighbourhood indicators, employment being the most important. A better steering of newly arrived migrants’ first housing and a more active education and labour market integration policy are measures more likely than area-based restructuring policies to be fruitful for combatting tendencies towards social exclusion in relation to the housing estates. Some might argue that as the cooperative segment of the estates is doing fine or at least better, why not convert remaining rental housing into ownership form? That might work for individual estates but the resulting gentrification will produce displacements and relocation of poor households elsewhere in the region or even beyond. We judge that the current level of rental housing, in particular, affordable rental housing, is already too low and too geographically concentrated for allowing enough people to move in and out of the Stockholm region and for sustaining a dynamic labour market.

7 Conclusion

Using earlier research, registry data and descriptive methods this chapter has provided an overview of the planning background and subsequent development of large housing estates in the Stockholm region. Despite criticism towards their standardised and dense layout, traffic separation, poor service level and other aspects, most estates did well socioeconomically until 1990. Our two cases, Bredäng and Rinkeby, had very different starts where Bredäng, as one of the early estates (finished by the mid-1960s) was populated with relocated, well established local households while Rinkeby (finished by the early 1970s) to a much higher extent was the destination for younger migrants from outside Stockholm and also from outside Sweden. We judge that these early conditions had a large impact on subsequent developments for more than 20 years.

Later on, Bredäng and many other formally successful estates seem to have undergone selective (in-)migration, where out-moving native-born Swedes, and employed people with higher incomes, have been replaced by successively larger proportions of new immigrants with weak attachment to the labour market. The dynamics of selective migration cannot be understood if not put in the context of the economic crisis occurring in the early 1990s. More than 12% of all jobs in the region disappeared in a few years’ time, and newcomers in the labour market such as immigrants and young people—i.e. those that were concentrated in housing estates like Rinkeby—were among the worst hit. In addition, changes in the housing market have resulted in a reduction of affordable rental housing in many parts of the Stockholm region. With fewer rental units available, those remaining have to absorb even more of the forthcoming expansion of immigrated (refugee) households, reinforcing selective in-migration. Also at this stage, rental-dominated housing estates previously less affected by immigration and expansion of unemployed people, like Bredäng, become destinations, both for immigrants seeking relocation from the most immigrant-dense estates in the region and for those arriving directly from abroad.

In explaining which housing estates that followed this trajectory (converging towards Rinkeby’s position) and which remained relatively unaffected, we can single out three important factors. First and foremost is tenure. Estates overwhelmingly consisting of rental housing have the potential to relatively quickly change their household composition, while estates with a larger proportion market-priced housing have lower turnover and exclude poorer households entirely from moving in. This does not mean that cooperative housing neighbourhoods are immune to selective migration, only that such dynamics will take longer time and that there are counter-balancing factors, especially when the housing market is tight and prices for entering are high. Second, housing estates located closer to existing clusters of poverty run a higher risk of being destinations for horizontal mobility from one area to another. This has to do with the very nature of residential mobility as very distance dependent. The bulk of movers tend to move over short distances because shorter moves have less transaction costs emotionally and socially. The third factor is building period, where the main difference seems not to be between Million Programme estates and non-Million Programme estates, but rather between early and late estates, where the dividing line seems to be around 1968. Our two cases illustrate a more general tendency that estates built later had less favourable initial conditions, i.e. lower local demand for housing and emerging stigmatisation of some estates.

Stockholm’s housing estates face many challenges and the list of interventions by property owners, municipalities and the State is long and covers more than 40 years. Some interventions have targeted physical aspects, others socio-economic and wider cultural developments. To an increasing extent, these are now framed within a safety discourse. However, irrespective of the reason for political and management reactions and the objective needs to intervene and refurbish, we believe too much focus is placed on area-based reasoning and much too little on the wider social and political issues determining the fate also of particular estates.

Field visits, hosts

In Rinkeby, May 29, 2017:

Peter Lundevall, city planner at the planning office, Stockholm City.

Per Granhällen, manager Security-Crime prevention, The Rinkeby-Kista district, Stockholm City.

Ewa Jungstedt-Pilestål, community planner at the Rinkeby-Kista district, Stockholm City.

In Bredäng, May 30, 2017:

Eveliina Hafvenstein Säteri, city planner at the planning office, Stockholm City.

Flor Luna, community planner, Skärholmen city district, Stockholm City.

Love Örsan, community planner, Skärholmen city district, Stockholm City.