Urban renewal without displacement? Belgium’s ‘housing contract experiment’ and the risks of gentrification

Gentrification has become part and parcel of urban policies throughout the world. Critics have argued against those policies but they have not yet developed concrete and comprehensive alternatives. This paper seeks to remedy this omission by investigating the Belgian ‘housing contract’ experiment (2005–2007). Quite exceptionally, Belgium’s ‘housing contract’ experiment was based on the premise that housing policies should improve the quality of life in deprived urban neighborhoods without displacing the poor. We investigate both the philosophy of the housing contract experiment as well as its effects. On the basis of this evaluation, we sketch the contours of a housing policy that incorporates rent gap theory and counters the negative effects arising from disinvestment and gentrification.

'balanced' or 'vital' neighborhoods, even if this means that poor households are forced to relocate. For instance, in the Netherlands, housing corporations are upgrading, demolishing and selling social housing in order to increase livability and improve a neighborhood's position within the urban system (Uitermark et al. 2007). In the United States, policy programs like HOPE VI and Moving to Opportunity aim to disperse low-income households. Such mixed-income housing policies, James DeFilippis and Jim Fraser argue, ''are largely based on the (hegemonic) mantra that low-income people themselves are the problem, and that a benevolent gentry needs to colonize their home space in order to create the conditions necessary to help the poor 'bootstrap' themselves into a better socioeconomic position'' (DeFilippis and Fraser 2010, p. 136). In the United Kingdom, promoting gentrification has been a key component of policy efforts since Blair's governments sought to tackle urban problems by ushering in an ''urban renaissance'' (Lees 2008;MacLeod and Johnstone 2012). These examples illustrate that gentrification has become part and parcel of urban policy and that already disadvantaged groups may bear the burden (Atkinson 2000;Smith 2006). In this context, it has been argued that the rhetoric of social mixing masks displacement processes and thus establishes gentrification by stealth rather than force (Bridge et al. 2012) The urban policies of the Belgian regional governments of Flanders and Brussels are in line with these general trends (Loopmans et al. 2007;Loopmans 2008;Loopmans et al. 2010;Van Criekingen 2008. However, for a brief period of time (2005)(2006)(2007), Belgium's federal urban renewal policy was an exception to the rule. As we will show, Belgium's federal housing contract experiment aimed to improve the living conditions in deprived neighborhoods without causing displacement of vulnerable groups. Although the housing contract experiment is far from radical in that it pursues modest reforms and is constrained by very limited budgets (69.7 million euro divided among 17 local governments between 2005, Rekenhof 2007, it nevertheless provides us with the opportunity to see how policies not aiming at gentrification work in practice and how they might be improved. The critical analysis of this case enables us to address the problem, observed by Lees (2003), that gentrification researchers have focused more on criticizing extant policies than engaging with policy and thinking through concrete alternatives. The most recent contributions to the debate Loopmans et al. 2010;Bridge et al. 2012) make abundantly clear that the rhetoric of social mix or revitalization can mask attempts to displace poor households. But they also highlight a need to explore how renewal can be designed so that it (also or primarily) benefits weaker groups on the housing market.
In this policy review, we first provide an introduction to Belgian's federal housing contract policy and analyze how it obtained its particular character. Secondly, we show how the policy experiment played out between 2005 (the year of its inception) until 2007 (the experiment ended in 2008). On the basis of interviews with policy makers, 17 contracts between the local and the federal governments, the evaluation sheets of federal officials, and secondary literature on the policy, we investigate the nature of the projects sponsored by federal funds. Thirdly, we show how, in translating federal goals into local policy projects, a process of goal displacement took place: although the federal government expressly tried to prevent federal funds from being used for measures that might cause displacement of lower-income groups, local governments in Flanders nevertheless managed to use these funds to buttress their gentrification policies. Fourthly, we draw from gentrification theory to offer a number of suggestions on how a housing policy for urban renewal can be formulated that avoids gentrification and displacement.

The political roots of Belgium's housing contract experiment
While other West-European countries developed increasingly ambitious urban policies in the 1970s and 1980s, the Belgian government did not have policies to combat urban decline until the mid-1990s (Cheshire and Hay 1989). Two factors account for the underdevelopment of federal urban policies. The first one is the historically powerful position of the Christian-Democrats. The Christian-Democrats dominated the Belgian government and used their power to strengthen their base in rural areas through the promotion of homeownership and affordable transportation (Kesteloot and De Maesschalck 2001;Uitermark 2003;De Decker 2008). In effect, home-ownership has become a major pillar of the Belgian welfare state and constitutes an alternative to social security provision (De Decker and Dewilde 2010). The second factor is the difference between the three Belgian regions. These differences are not only related to diverging interests and beliefs between regional policy makers but also to the make-up of cities. Cities of the nineteenth century Walloon industrial axis face deeply entrenched problems, both in the physical condition of the built environment and the social situation of its residents. Flemish cities as well as Brussels have a much stronger profile. One of the most telling indicators is the population evolution of cities (see Loopmans et al. 2007). Various Flemish cities never experienced a population decline in their recent history, and most Flemish cities, as well as Brussels, saw the population in their inner cities increase after 2000. In Wallonia, most inner cities initiated their population decline earlier and have kept on losing population till today. The largest cities in Wallonia even face population decline in their Functional Urban Region and have long been 'shrinking' (Bontje 2004).
Within this context it is not at all evident that federal policy makers have developed renewal policies. Indeed, the first, still comparatively modest, urban policy initiatives were taken at the regional rather than the federal level. The first attempts to develop urban policies at the federal level were made in 1999, when for the first time in postwar history the Christian-Democrats were kept out of government. The largely 'urban-based' coalition of green, liberal and socialist parties (De Maesschalck and Loopmans 2003;De Maesschalck 2010) set up a new federal urban policy fund. Charles Picqué, a francophone socialist minister, was the first minister of urban policy. Over time, the policy acquired an increasingly 'social' profile. To gain insight into urban deprivation in Belgium, Picqué ordered a group of left-leaning academics to identify 'neighborhoods in difficulties'. The resulting 'Neighborhood Atlas' provided a graphic representation of the relation between space and deprivation, thus focusing both public and policy attention on deprived neighborhoods and their residents . The coordination of the policy was relegated to a subdivision of the Federal Department of Social Integration. Together with the socialist background of its minister, this further contributed to the increasingly 'social character' of the federal urban renewal policy. As a civil servant of the Big Cities division explained: Giving control to a 'social affairs' ministry contributed to the social character of our renewal policy; I've met several of my European colleagues who are in economic or planning departments; they do not necessarily have the same innate social reflex as us. (interview member of Big Cities division, Federal Department of Social Integration, 21-12-2010).
In terms of organization, the federal Big Cities Policy takes its inspiration from the French Politique de la Ville, the European URBAN program and the Flemish urban policy. Local authorities enter into a result-oriented contract with the federal authorities, thereby receiving financial support for specific urban development projects. While the federal policy did not stipulate the exact content of the contracts between the cities and the federal government, the Big Cities division kept an eye on 'the social character' of their initiatives and sometimes used its discretionary power as subsidizing agent to curb what they considered 'anti-social' measures.
This structure was in place when, in 2005, the new minister for Big Cities Policy, Christian Dupont, another francophone socialist, took the initiative to introduce 'housing contracts' as a separate, experimental wing in the policy. The minister argued that housing measures were necessary to reinforce social cohesion and social integration in cities. In particular, rack-renting by slumlords (renting substandard housing to easy-to-exploit groups like undocumented migrants and poor people), underinvestment in housing and gentrification-induced displacement were considered negative effects of contemporary urban development (interview federal coordinator Big Cities Policy, 21-12-2010). The strategy of the housing contract policy is to tackle problems of ''insecurity, impoverishment, speculation and lacking infrastructure in urban neighborhoods by stimulating social mix and social cohesion in deprived areas through housing sector interventions'' (POD Maatschappelijke Integratie 2005, p. 5). Aware of the potential perverse effects of social mixing policies, policy makers stipulated that social mixing should ''not lead to the dislocation of problems of exclusion to other neighborhoods'' (Ibid., p. 15). The federal government in particular wanted to meet the housing needs of the lowest-income groups who have difficulties in securing decent housing on the private market; consequently, households above a (regionally established) maximum income level are not counted among the policy's target groups.
The policy defines four priorities that can be targeted by cities. First, the number of high-quality dwellings for lower-income groups should be increased. Second, the policy should help low-or middle-income people with insufficient means to attain a house in the city. Third, measures should be taken to counter vacancy, dilapidation and the exploitation of poor households by slumlords. Fourth, interventions in the housing market should be used as a stepping stone to more generally improve the situation of precarious groups like the homeless, long-term unemployed, former prisoners or battered women (POD Maatschappelijke Integratie 2005, p. 8).

Social displacement as goal displacement
Although the federal government had a strong political motivation to develop a federal policy next to regional and local policies, goals were intentionally set in general terms. The formal policy document has to allow the minister to further specify own priorities along the way and local governments to develop innovative projects targeted to their particular local needs. When we examine the allocation of budgets and projects over the four priorities as defined by the federal minister, we observe striking regional differences (Table 1). In Flanders, almost one-third of the projects and 69 % of the budget are devoted to achieving target 2, i.e. to facilitate housing acquisition for low-to middle-income groups. In Wallonia, this target is selected in only 8 % of the projects, while 72 % of the projects and 95 % of the budget are oriented towards the struggle against slumlords, vacancies and abandonment. In Brussels, finally, the budget is split equally across three different priorities (increasing the stock of qualitative rental housing; combating slumlords, vacancies and abandonment; and reintegrating deprived groups through access to housing).
In part, this divergence reflects actual differences on the ground in each of the regions. Walloon cities invest much more in the struggle against vacancies and slumlords than the other two regions, and cities in Flanders sometimes use the funds to include subsidized housing in otherwise market-driven projects. However, the differences also reflect their discrete political choices and ideologies. Flanders, which is under strong Anglo-Saxon ideological influence, has adopted the neoliberal urban development approach as found in the Netherlands or the United Kingdom. For years, Flemish cities have promoted gentrification under the pretext of social mixing (Loopmans 2008;Loopmans et al. 2010).
In this context, the different emphases with respect to the second priority are especially revealing. This is the only priority for which cities have some discretionary authority to target residents who are not among the most vulnerable in the population. The idea behind this priority was that policy makers could help low-and lower-middle income groups to secure a place in cities. With this in mind, both the Brussels and Walloon region defined maximum income levels to delineate the 'lower-middle class'. The Flemish region, however, did not formulate such a general criterion, and cities were allowed to delimit this target group autonomously. Consequently, Flemish cities exploited the opportunity to expand the target group so as to divert funds away from the most marginalized to the middle class. Moreover, some cities even refused to define maximum income levels so as to allow subsidizing of market-oriented property development. As one federal official noted in an evaluation of a local plan: To our regret, the regulations don't stipulate the conditions concerning the income of the tenant which allows the private owner to let its premises to higher-income groups, which is not the intention. There is a risk that the system is abused and that this subsidy benefits a group which does not need it. (except from an evaluation form on renovation subsidies in private rental housing in a Flemish city).
Belgian officials noted that cities used a range of measures-or 'tricks'-to interpret the rules in relation to subsidized housing in such a way as to facilitate gentrification: Another trick of the cities was not to fix in advance the ratio between rental apartments and apartments for sale. As buying an apartment is not accessible for the lowest-income groups, the overprovision of apartments for sale was a trick to exclude those groups from the project. (interview with federal official, 09/07/2008).
Federal officials intending to halt what they regarded as the abuse of federal means 'for the poor' to the benefit 'of the rich' regularly clashed with local governments over this issue (interview federal official, 11/12/2010). In its evaluation of the federal housing contracts, the Audit Office raised a similar criticism: ''too often, 'social mixing' meant the attraction of higher-income groups in poor neighborhoods, with little or no means to check possible negative outcomes such as displacement of lower-income groups'' (Rekenhof 2007, p. 48).
In response to such gentrification efforts, federal officials tried to reason with local policy makers or pressured them to change the plans: In the city of X, there was a project where we subsidized the acquisition of buildings that were in use as brothels in order to turn them into housing. We found out that the buildings were subsequently torn down and that the plot was sold off to a private developer to construct luxury lofts instead. In this case, we firmly opposed. (interview with federal official, 09/07/2008).
While recognizing that such efforts to promote gentrification ran counter to the stated policy objectives and the underlying policy philosophy, federal officials found it difficult to develop a well-grounded argument with local governments about the type of measures most suitable for the cases at hand. Federal policy makers remained dependent upon and locked in by the detailed problem analyses provided by local policy makers.

Rent gap theory and social urban renewal
We observed that the federal housing contract policy was intended to promote 'social urban renewal' but that local governments sometimes used federal funds to promote gentrification. Federal officials had the ambition to implement a policy primarily meant to aid the most vulnerable groups on the housing market but lacked an urban theory that could help them to determine which local measures would be appropriate. Rent gap theory (Smith 1996;Lees et al. 2008, chapter 2) may be of help in remedying this omission as it can inform a logic through which appropriate local interventions can be determined.
The rent gap is the ''disparity between the potential ground rent level and the actual ground rent capitalized under the present land use'' (Smith 1996, p. 67). This disparity is likely to be small immediately after a building is erected. Over time, the building loses its value: its design becomes anachronistic, its fabric dilapidates, its amenities no longer meet the highest standards. As the value of the building goes down, so does the rent and the socio-economic status of the inhabitants. At the end of the cycle, landlords milk the dilapidated properties by renting them out, legally or illegally, to inhabitants lacking access to the more attractive parts of the housing stock. As the building deteriorates, the rent gap grows. The disparity between the potential and actual ground rent is further increased if it is situated in a location that is increasingly central as a result of outward urban expansion. When the difference between the maximum rent and the actual rent is large enough, a reinvestment is made and the plot is upgraded through renovation or new construction for inhabitants able to furnish higher rents.
The concept of the rent gap invites us to take a dynamic and geographical view of housing markets. The cyclical process of investment, disinvestment and reinvestment occurs in different places at different moments, creating a constant ebb and flow of investment capital and inhabitants across space and through time. However, the process of boom and bust is not a given; it only takes place when there are social inequalities and when the housing stock is so much commodified that these inequalities find an expression in urban space. The function of a renewal policy can thus be to attenuate this process: it can ensure that (1) unattractive neighborhoods do not degenerate to the extent that they contain only people left without other options and that (2) attractive neighborhoods do not gentrify to the extent that only higher-income groups can enter or remain there.
From this perspective, the trajectory of a neighborhood is more relevant for determining if and how it should be targeted than its composition. When neighborhoods are experiencing disinvestment, the government can step in and compensate by covering unprofitable investments or supporting provisions suffering from disinvestment. When neighborhoods are receiving investments, government policy would aim at preserving the rent gap. Policies can for instance reduce direct and exclusionary displacement (Marcuse 1986;Davidson and Lees 2005;Slater 2009) through rent regulation, the expansion of social housing, or a combination of both. Finally, there is no need for territorial investment when neighborhoods are stable.
This counter-cyclical investment logic provides a straightforward way to identify where and how investments should be made. Rather than providing extensive lists of the kind of measures that might be taken or the actors that might be involved, it provides a rough but clear guide on how federal funds should be spent. While straightforward, this logic does take into account geographical diversity. In the particular case of Belgium, it suggests focusing efforts to attenuate displacement on the booming Flemish cities and Brussels while attempting to rejuvenate Walloon cities suffering from disinvestment. The countercyclical investment logic moreover facilitates social mixing in the long term. Rather than promoting gentrification in affordable yet attractive areas-the easiest way to promote social mixing in the short term and the surest way to create exclusion in the long termefforts are aimed at preserving diversity in booming areas by blocking a closure of the rent gap. In contrast to current mixing policies in Belgium and elsewhere, these types of measures directly benefit the weakest households on the market and attenuate overall levels of socio-spatial inequality.
Within the Belgian context, the tools for selecting neighborhoods through this logic were available. In 2007, the federal government commissioned a new Atlas that provided a dynamic analysis of 'neighborhoods in difficulties' ). This Atlas and its planned updates could have become important decision making tools for federal officials to decide about the suitability of proposed projects. However, the Atlas has only been used in the background, for the formulation of general policy goals, and failed to be deployed in the negotiations with local policy makers. Today, the context for developing an anti-cyclical urban policy is less favorable. The housing policy contracts have been subsumed under the general urban policy, decreasing the possibilities to target investments to housing measures for the most deprived groups. This urban policy, finally, is not expected to survive for long: at the moment of writing, Belgium's regions are drifting further apart and urban policies are on the short list to be downscaled to the regions.

Conclusion
Belgium's federal urban renewal policy is exceptional in that it aims to benefit the weakest groups on the housing market and protect them against displacement. However, as we have shown, local governments had the chance to pervert the federal housing policy by using it to promote gentrification. One reason for this has simply to do with power relations-the Flemish cities had enough discretionary authority to infuse their preference for gentrification into the federal policy. But another reason is that federal officials did not develop a policy theory and policy instrument to assess the projects proposed by local policy makers.
The relevance of Belgium's urban renewal policy is clearly limited if we look at its means or future. However, the Belgian experiment remains relevant for thinking through alternatives to neoliberal urban policies. If supra-local governments elsewhere seek to improve the position of the weakest groups in the housing market, they could learn from the implementation problems analyzed here. This is especially important in the context of a recession. Since governments tighten their budgets for social and urban policies, the question of where their expenditure will generate the most benefit becomes more urgent. Alternatives to the neoliberal 'common sense' not only need 'social intentions' as expressed in the positive rhetoric of 'mixing' and 'balance' but also a well thought-out policy framework. The reason is that goal displacement may occur when local governments seek to use federal funds for social programs to promote gentrification. The probability that policies are perverted in this way only increases in times of recession. Desperate to promote at least some development, local governments may be tempted to stimulate growth in areas that already have a strong position. While the gentrification of attractive yet poor neighborhoods may result in social mixing in the short run, the mediumto long-term effect is to accelerate displacement and reinforce segregation. Our argument is that such temptations to develop in the most attractive areas have to be resisted, both institutionally and intellectually. If we believe that residents in gentrifying neighborhoods profit if they can stay, the policy answer should not be to promote gentrification but to let as many people stay as possible by countering displacement. Conversely, in areas that suffer from decline, there is a point in stimulating development. The crucial question is thus when and where to intervene. Our review of the Belgian policy suggests that the designation of neighborhoods is not a residual or marginal part of the policy process. The same sort of measure (promoting home ownership among the lower-middle class, for instance) can have radically different repercussions depending on the context where it is implemented. Rent gap theory might provide a theoretical framework for designating neighborhoods, as it could enable policy makers to identify the dynamism of neighborhoods and assess potential policy measures.
Our suggestions for an anti-cyclical logic underpinning social urban renewal allow critics of gentrification to argue for a different type of social mixing policy. Luring middleand higher-income groups into deprived areas is one way to promote social mixing and an apparently irresistible policy goal. While in the short term the result may indeed be that neighborhoods become diverse, in the somewhat longer term the reduction of the share of low-income groups may lead to homogeneous high-income areas and accentuate socialspatial inequalities. Our proposals aim to achieve social mixing via a different route: by preventing neighborhood decline and protecting lower-income groups against the threats of gentrification-induced displacement. Clearly these sorts of analyses will not sway policy makers determined to cater to the middle classes or to reduce social housing, but they may be helpful to government officials and academics who are looking for concrete and comprehensive alternatives to current neoliberal policies.
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.