The neo-liberal politics and socio-spatial implications of Dutch post-crisis social housing policies

Abstract This review discusses changes in Dutch housing policy that were implemented after the great financial crisis of 2008, notably the 2015 Housing Act and its lead-up legislation, and a landlord levy aimed at taxing housing associations. We argue that these changes should be seen as a decade-long process of institutional re-regulation that may be characterised as neo-liberalisation. In addition to reviewing the neo-liberal politics of these changes, we also provide an assessment of the social and spatial implications of the new legislation. Financial pressures on housing associations and higher income tenants together with the stricter eligibility for new tenants have led to a relative decline and residualisation of the social-rental sector. Also, housing policies are set to increase the spatial concentration of disadvantaged population groups. These socio-spatial implications may further undermine social housing in the future.


Introduction
The Dutch housing context has long been known for its sizeable social housing sector, which is largely managed by around 300 housing associations. These associations are private entities that were instrumental in building-up the social rental sector in the Netherlands. In the 1990s, they were cut loose from the Dutch State and were effectively given free reign over their affairs. They were designated as not-for-profit organisations, receiving no direct subsidies but with access to state-backed loans (Aalbers, Van Loon, & Fernandez, 2017;Ekkers & Helderman, 2010;Van Gent, 2010a). While the sector had continued to build and renew social housing vigorously after this deregulation, reports of mismanagement and criminal behaviour severely undermined reputations, leading to a crisis of moral legitimacy (Blessing, 2015). This crisis came to a head in 2011, when the Dutch State bailed out the Vestia housing association for over 2 billion Euros after it (mis)speculated in derivative markets (see Aalbers et al., 2017). Pressure from politicians, the public and many housing associations themselves led to a re-regulation in social housing (Nieboer & Gruis, 2014). New legislation had already been introduced in the years after the 2008 financial crises (Boelhouwer & Priemus, 2014), but this new direction would be cemented in the 2015 Housing Act.
This review evaluates the recent round of re-regulation of social housing in the Netherlands with a particular focus on social and spatial implications. We show how recent political institutional reforms come in various guises, including disciplinary fiscal measures, the sale and privatisation of social housing, forcing housing associations to take a step back and restricting access to social rent to the lowest income groups. The arguments put forward in this review paper have wider relevance as they are essentially about the process of social housing residualisation, which has also been identified and studied in a range of other countries (e.g., Forrest & Murie, 1983;Hoekstra, 2017;Magnusson & Turner, 2008;Malpass, 2004;Orchard & Arthurson, 2005;Schutjens, Van Kempen, & Van Weesep, 2002).
We argue that institutional changes cannot only be seen as a direct response to problems and risks with autonomous housing associations, but should be viewed as part of a longer project of neo-liberalisation in Dutch government (see Oudenampsen & Mellink, 2019). The demise of public expenditures related to welfare and housing in favour of state interventions to secure (re)commodification have been hallmarks of neo-liberal housing policy (Aalbers, 2016;Jacobs, 2019). From this perspective, the potential harmful socio-spatial effects of new housing policies are not an oversight but inherent to their political design (Madden & Marcuse, 2016). Yet, it is not a uniform or consistent project. Neoliberalisation is a variegated, partial, contradictory and uneven process that may include efforts of re-regulation alongside overarching de-regulation, depending on national and local contexts (Brenner, Peck, & Theodore, 2010;Jacobs, 2019;Van Gent, 2013). Processes of neo-liberalisation are also driven by crises, which tend to trigger reforms that displace or postpone crisis effects in favour of shortterm financial stability. The global financial crisis is a testament to that effect. 'Shock therapy' and new rounds of regulation typically follow crises, leading to a spiralling process of reform (Brenner et al., 2010). As these regulatory reforms are building on previous rounds within specific political configurations, 'institutional geographies of neo-liberalisation have been repeatedly reconstituted and remade' (Brenner et al., 2010, p. 210).
The Netherlands is a peculiar case within the European context. While many European countries witnessed a strong demise of social housing in the last four decades (Scanlon, Fern andez Arrigoitia, & Whitehead, 2015), the Netherlands, for a long time, managed to retain a relatively high share of social housing that also provides for middle-income groups. The ownership by private entities (housing associations) and the relatively broad access to social rental housing have arguably led to a strong institutional embeddedness of, and social support for, social housing in the Netherlands, and have arguably been obstacles to structural reforms. Yet, a gradual demise of this social housing tradition is under way. The 2015 Housing Act codified a set of new rules, introduced after the crises, with regard to housing association investments and to access to social housing. In short, there is less potential and less autonomy for housing associations to invest and access has been curtailed. The act furthermore coincided with a range of other measures diminishing the role of housing associations.
Despite its socio-historical embeddedness, our case adds to broader political understandings of how housing reforms, and their socio-spatial impacts, contribute to the gradual and partial erosion of a strongly institutionalised social-rental sector. The next section will further discuss neoliberal politics of Dutch housing policy. This will be followed by a review of new regulation and their implications. In our conclusion, we will pose that socio-spatial effects are not only an outcome, but also part of reform politics. The reduction of, access to, and the spatial concentration of, social housing will initiate new political logics that push residualisation further.

Neo-liberalisation in Dutch housing 1989-2008
Like many Western European countries, the Netherlands made significant public investments in housing in the post-war era. While a brief period of liberalisation may be observed in the early 1970s, the commitment to social housing as a collective or 'merit' good increased when a Left-wing Catholic coalition came to power. A 1974 memorandum on rent subsidies expanded the social housing sector considerably through various subsidies and loans. The number of subsidies and loans continued to grow during the economic crisis years of the 1980s, when they were part of an anti-cyclical investment strategy (Ekkers & Helderman, 2010). As a result, the number of social units had almost doubled between 1975 and 1990. Yet, the share remained at 38% of the Dutch housing stock, because the number of owner-occupied housing units had increased at a faster rate (from 37% in 1975 to 44% in 1990). More importantly, this system of subsidies had grown from 0.5% of the national income in 1970 to 2.7% in 1990 (Ekkers & Helderman, 2010).
The 1989 memorandum, Volkshuisvesting in de Jaren Negentig, marked a change of direction away from supporting social housing towards the further expansion of owner-occupied housing (Van Gent, 2010a). The rental sector had to be self-supporting with less reliance on loans and subsidies. The following years saw a re-regulation of housing associations, giving them more autonomy over their assets and investments. Commitments to subsidies and outstanding loans between associations and the state were settled in 1995. The 1992 and 2005 Besluit Beheer Sociale Huursector outlined their social responsibilities but also provided greater flexibility in terms of investment strategy and finance.
It was also during the 1990s that state funds for large-scale renewal projects were made available as part of selective urban policies that target deprived neighbourhoods. Priority in urban policy shifted from improving housing conditions to addressing changing social conditions. The modus operandi of regeneration revolved around tenure restructuring to alter the social mix. Typically, the social housing stock would diminish in favour of owner-occupied housing, to accommodate higher income groups in disadvantaged areas, in an effort to break through a perceived spiral of decline. This strategy typically relied on housing associations, supported by municipalities. They had to demolish, sell or deregulate their social rental stock and often develop new housing for sale or increase rents (Van Gent, 2010b). Given their new entrepreneurial role and the availability of state funds, the 1990s and 2000s saw many regeneration projects predicated on renewal and mixing, particularly in low demand urban areas and peripheral post-war housing estates (Musterd & Ostendorf, 2008).
Dutch housing and urban policies dating from 1990s into the 2000s reflected the political landscape at the time. During these decades, Third Way social democratic party (PvdA), conservative liberals (VVD) and conservative Christian-democratic party (CDA) shared power in various coalitions, sometimes augmented by liberal democrats (D66) and left-leaning Christian-democrats (ChristenUnie). The continued support for (new) social housing and renewal reflected social democratic ideals related to welfare provision, neighbourhood liveability and economic integration. The new emphasis on ownership, deregulation, privatisation and the cultural integration of minority groups fitted market-liberal and conservative agendas. Thus, Dutch coalition politics resulted in a 'hybrid' status of social housing governance that integrates the spheres of state and market, and infuses collective provisions with market technologies (Blessing, 2012). Such hybridity, and contradictions, in Dutch programmes and strategies is illustrative of how neo-liberalisation of housing policy is rooted in local institutional and political frameworks, and therefore has a geography (see Brenner et al., 2010;Larner, 2003).

Post-crisis housing politics
The late 2000s marked a new direction in Dutch housing policy. Public unease over executives' wages and failing development projects had already resulted in new corporate taxes and state pressure to invest more in area-based urban policies by 2007. These changes were not only the result of a discontent with how housing associations had been conducting their business though. The 2008 financial crises and their political aftermath arguably provided a window of opportunity to push for further neo-liberalisation of the housing system.
The 2008 financial crisis triggered a housing crisis that led to severe financial problems in urban regeneration projects that relied on the sell-off of existing social housing and newly constructed owner-occupied dwellings in lower demand areas. As a consequence, large-scale regeneration projects were heavily curtailed or permanently cancelled (see Boelhouwer, 2017;Savini, 2017). More importantly, the 2010 European debt crisis led to austerity. This year also marked a shift in the Dutch political landscape which saw the conservative liberals becoming the largest party and coalition leaders for the first time since general suffrage (in 1919). Crisis and austerity opened new avenues for state restructuring (Brenner et al., 2010;Ronald & Dol, 2011). Housing allowances were cut and there were proposals to give 75% of housing association tenants the 'right-to-buy' (Boelhouwer & Priemus, 2014). This proposal was impossible as it would have violated the private entities' property rights, but it shows political ambitions. One measure was a landlord levy aimed at housing associations and any other organisation that owned more than 50 rental dwellings. The tax represents half a percent of the tax value of the housing assets per annum. The tax was an austerity measure that served no other purpose than to cover the state budget, with the Dutch state effectively extracting about e1.7 billion a year from the social sector. The levy is still in place at the time of writing and scheduled to incrementally grow for the coming years (see Priemus, 2014).
The room for political manoeuvring offered by the legitimacy failures of associations along with the financial crisis also allowed for the introduction of market-friendly restructuring as had been insisted upon by the European Union. In 2005, the European Commission (EC) had been called on to address the so-called 'problem' of market distortion in Dutch housing development, as housing associations were allowed to develop housing for the private market with various forms of state aid (Priemus & Gruis, 2011). Private actors had no way to access state aid through public tenders as per EU common market rules. In addition, the Dutch Association of Institutional Investors filed a complaint to the European Commission in 2007 over various forms of state aid to housing associations that develop private housing. Their complaint pertained not only to state aid in housing development but also addressed the lack of regulations concerning housing services for non-disadvantaged tenants (Priemus & Gruis, 2011). A 2009 EC ruling was taken as an opportunity to restructure the provision of social housing. After consultation, the Dutch government proposed new housing regulations, which came into effect in 2011, and after some adaptations, would become part of the 2015 Housing Act.
One important outcome of the ruling was the introduction of income eligibility rules for 90% of the housing association stock. Such means-testing and allocation rules had not existed before. Low-income households would be given preference and the majority of the stock would be rented to this group as well as vulnerable target populations. Previously, only about 75% of all social-housing units were allocated to this group (Hoekstra, 2017). This has historical reasons. Since their birth in the early twentieth century until the 1970s, housing associations were meant to provide decent housing to the working population and not necessarily to disadvantaged groups. Yet, in 1989, when ownership housing was held up as the middle class norm, a slow residualisation of social housing began (Van Kempen & Priemus, 2002). The 2011 regulations stated that a gross household income had to be less than e33,000 to be able to enter regulated housing. This meant that about 57% of the Dutch population was excluded from regulated housing (Boelhouwer & Priemus, 2014). The maximum would be indexed every year and stood at e38,035 in 2019. To be sure, while the EC ruled that a maximum income limit was necessary to strictly separate regulated from market rent, the Dutch state determined the exact height and stipulations of this income limit.
In addition to restricting access to social housing for non-disadvantaged groups, the residualisation of social housing was further pushed by yearly additional rent increases for tenants whose household income exceeds the eligibility criterion. The rent increases (up to 5%, plus inflation rate) were meant to displace tenants who were thought to be able to buy a dwelling in a distressed market. The rent increases were presented to housing associations as a way to finance the landlord levy. The decision to effectively tax middle and higher income tenants in times of financial crisis was made possible due to a revanchist discourse on 'skewed living' (scheefwonen). 'Skewed' refers to social tenant incomes being too high for social rent. While the moral hazard of more affluent residents living in regulated rent had been a recurring trope in liberal housing discourse for many years (e.g., Dieleman & Musterd, 1991), mandatory means-testing had solidified the argument that regulated rental housing is not a collective good open to all (see Musterd, 2014).
The opposition of the liberal-led government to regulated rental housing is perhaps best illustrated by their active attempts to push the sale of housing association stock. The Housing Act means that housing associations are discouraged from providing rent-liberalised housing, at least for 90% of their new tenants, yet institutional investors are able to increase rents. Information on this rent gap was helpfully provided to potential buyers by the housing ministry website. Stef Blok (VVD), Housing Minister during the 2012-2017 period, would also visit real estate fairs with the aim of enticing foreign institutional investors to buy up dwellings from financially distressed housing associations (Hendriks, 2018). The narrative of a surplus of social-rental housing is maintained to this day, despite a growing housing shortage, affordability issues and increasing prices. The 2017-2021 VVD party programme explicitly stated that the landlord levy needed to be increased in order to pressure housing association into selling more of their stock (Hendriks, 2018).

Changes in stock and residualisation
New housing regulations instituted after the crisis restricted access to social housing while also putting financial pressure on housing associations and their tenants. The continued demise and residualisation of social housing is evident in the declining share of stock. Whereas housing associations owned about 40% of the housing stock in 1990, this had declined to 30.7% in 2012 and 29.3% in 2018. Between 2007 and 2017, there was a net loss of 23,500 housing association dwellings, while the total housing stock grew by over 600,000 dwellings, as well as households (Figure 1). Housing association dwellings with a monthly rent below the liberalisation threshold 1 even show a 5% decline between 2007 and 2017, with a noticeable drop when the Housing Act came into effect in 2015. Only those dwellings with a rent below this threshold are rent regulated, and therefore social, based on a point scoring system measuring dwelling quality-while dwellings with higher rents are officially rent liberalised. The result of these developments is an increasing gap between the total size of the social-rental stock and the total number of households.
The net decline of social housing is due to the sale of rental dwellings and to low rates of new housing construction. While 25,500 dwellings, in various forms of tenure, were constructed by housing associations in 2009, 2018 saw just 17,300 new dwellings. This latter number falls significantly short of the current ambition of government and housing associations to build 27,000 to 34,000 new dwellings per year (Aedes, 2019). The lower rates are due to rising construction costs, land shortages, municipal regulations and planning rules and new taxes. Moreover, Table 1 indicates that when the levy took effect in 2013, new construction by associations nearly halved. The units that are being built are comparably smaller than before (Ministerie van BZK, 2018), as these can be rented out for low rents.
At the same time, tenants have become more marginalised. While 12% of tenants belonged to the lowest income quintile group in 1990, this share increased to 37% in 2009 and 45% in 2015 (SCP, 2017, p. 311). Social housing tenants were also confronted with higher rents. Between 2009 and 2015, the share of tenants that had insufficient income to pay rent doubled from 9 to 18%, while remaining stable around 3% among owner occupiers (PBL, 2016). During this period, the share of income spent on housing showed a sharp increase among social tenants, from 32.7 to 37.3%. Yet, between 2015 and 2018, this dropped slightly, from 37.3 to 36.3% (CBS, 2019). The most recent drop in housing costs may be explained by improving economic prospects (e.g., less unemployment) in the wake of the post-crisis economic boom, but also by a better allocation of social housing. With the introduction of the 2015 Housing Act, it is formally delineated that the most affordable social-housing units should go to the lowest income households. In 2013, 31% of the new tenants in regulated dwellings had 'expensive rents'. 2 This number had dropped to 5% in 2016 (Ministerie van BZK, 2018), indicating that expensive social rents for lowincome households has become less of an issue. New maximum income criteria for regulated rental housing have effectively barred middle-income groups from entering this tenure. Simultaneously, ongoing labour-market flexibilisation, the rapid increase in house prices since 2013-especially in Amsterdam and Utrecht regionsand the introduction of stricter mortgage lending criteria mean that middle-income groups increasingly struggle to buy a dwelling. A growing group of middle-income households finds itself between the two tenures, partly as a result of housing re-regulation (see Aalbers, Hochstenbach, Bosma, & Fernandez, 2019;Hochstenbach & Ronald, 2018).
The worsening position of middle-income households has become a focal point in Dutch political debates on housing. The dominant solution offered by political leaders and developers is so-called middle-income housing, with monthly rents between e720 (the liberalisation threshold in 2019) and e1000. It is questionable, however, to what extent this is actually a solution: around 30% of all middle-income households are unable to afford a rent above e720 per month (Van Middelkoop & Schilder, 2017), and rents per square metre in this middle segment are, typically, comparatively high. Furthermore, while private developers struggle to develop middle-income housing in high demand (urban) locations, the 2015 Housing Act frustrates housing associations in providing this type of housing.

Segregation
New allocation rules, means-testing and income-dependent rent increases have created conditions for the concentration of disadvantage. For Amsterdam, such an increase of disadvantage has already been observed in concentrations of housing association dwellings, leading to increasing segregation (Musterd & Van Gent, 2016). This tendency towards concentrations of disadvantage in housing association dwellings will be further strengthened by relaxing rent control for private rental. These are not just concentrations of poverty, though housing associations serve an increasingly important task in housing other population groups featuring a range of social problems, such as residents who are being reintegrated from social relief and shelters. The Dutch rental system regulates rent levels for housing owned by both housing associations and private companies based on a point system. Points, and therefore rent, are based on amenities, quality and dwelling size. After a threshold of a number of points, rents may be raised to any level; they are 'liberalised'. This point system means that rent on privately owned dwellings is also regulated. However, this system has been slowly undermined by adding additional points based on the dwelling's tax value in 2015 (Ministerie van BZK, 2015). This means that in attractive neighbourhoods and in high demand regions where tax values are higher, private landlords are able to 'liberalise' their rental dwellings after sitting tenants leave, leading to rent increases in high demand areas.
In Amsterdam, where pressures on the housing market are particularly high, the share of relatively affordable regulated rental dwellings in private ownership decreased from 18% of the total housing stock in 2007 to 13% in 2017, while the share of more expensive rent-liberalised private-rental dwellings increased from 5% to 11% during this same time period (Hochstenbach & Ronald, 2018). When looking at new allocations, the vast majority of private-rental dwellings are rented out in the relatively expensive liberalised segment.

Urban policies
The deregulation of privately owned rental stock means that lower income households in high demand areas have fewer options, leading to higher demand for social rental housing. To be clear, concentrations of low-income households are not necessarily a problem. Generally, when these concentrations remain relatively small and mixed, there may be no additional negative effects for most residents (see Miltenburg, 2017;Van Ham, Hedman, Manley, Coulter, & € Osth, 2014). Yet, regardless of whether such effects exist, the visibility of disadvantage will most likely lead to policy responses. For instance, in the US, the government has proceeded to intervene and to restructure the housing market in the few public housing projects they have (Goetz, 2003).
Whether Dutch national policy will turn to renewal and social mixing again remains to be seen. The 2010 governing coalition of conservativeliberals and Christian-democrats was kept in power by the radical-right PVV. At their behest, urban policies that targeted deprived urban areas were cancelled altogether, effectively ending two decades of urban policies that relied on direct interventions. The only national urban policy in effect is the 2006 'Act on Extraordinary Measures for Urban Problems', which relies more on the regulatory exclusion of low-income groups than on state investments (see Uitermark, Hochstenbach, & Van Gent, 2017). Initially, there had been a reduction in the landlord levy for investment in renovations and renewal. This has been abolished in 2018 and replaced by a reduction for investment in environmental sustainability and energy saving (e.g., insulation and more efficient heating).
Local governments may still engage regeneration initiatives though. As a way to remedy the former lack of oversight, the Housing Act prescribes housing associations, municipalities and tenant organisations to agree on a housing agenda. This prescription offers municipalities a new way to direct new construction by housing associations and coordinate regional housing questions while also having the tenants involved. Housing associations, however, report that the coordination with municipalities is frustrating new construction (Aedes, 2019). More importantly, while public involvement has increased, the EC-ruling has meant that every renewal programme that relies on the development of private housing has to be subjected to a 'market test'. The effect of the market test is currently unclear, as is the degree of freedom municipalities have to outline their housing programme. The market test may not hinder renewal here, as private developers are usually not keen on investing in deprived areas. Indeed, the necessity of tenure restructuring as a strategy of regeneration is up for debate. Yet, the market test may impede any integrated regional social housing programming that also spans non-deprived areas.

Conclusion
This article reviewed recent Dutch housing policies that sought to restructure the social-rental sector from a collective provision to a last-resort tenure. While the residualisation of social housing has been the topic of research before, our review offers two insights with wider relevance. First, we show how the transformation of a housing system is a long-term, inconsistent and partial process. We particularly highlight how crises can play a crucial role in creating new windows of opportunity for politics to push through and accelerate reform, but also how these reforms may predicate new crises. Second, the reforms have specific social and spatial ramifications, which are in many ways deleterious for low-income populations. We argue these implications are not merely outcomes, but also give birth to new policy waves that further cut back on social housing, creating a selfreinforcing feedback cycle. The remainder of this conclusion discusses these two key points in more detail.
Before 2008, the Dutch housing system already underwent rounds of liberalisation by making housing associations more market-oriented, promulgating home ownership through social mixing and deeper privatisation, deregulating and supporting mortgage markets, and reducing rent subsidies. The new round of housing reforms was a response to the broader budget crisis after 2008, as well as the fall in housing values and the threat of mortgage defaults (see Boelhouwer, 2017). In addition, financial mismanagement and misconduct of some housing associations showed that the lightly supervised entrepreneurship in the social sector represented a systemic risk, as the Dutch State is liable in case of defaults. These currents were navigated by liberal-led governments and housing ministers who were ideologically predisposed to drastically re-regulate rental housing. It should be mentioned that the housing associations themselves, particularly the ones that fully embraced entrepreneurialism after the reforms of the 1990s, had also undermined themselves, at least in the public eye. Their wealth, as shown in over-ambitious projects and in increasing director salaries, was indignantly portrayed in the media as their conduct did not reflect the bearings of a public institution (e.g., Gualth erie van Weezel, 2014). From this perspective, the seeds of the legitimacy crisis were also sown in the previous round of comprehensive housing reforms.
The new regulations are an important step in the decades-long conversion of the regulated rental sector, once the cornerstone of the Dutch housing system, from a collective provision to a temporary service for lower income households and marginalised groups (see Fitzpatrick & Stephens, 2008;Jacobs, 2019). We discussed this transformation of social housing within the framework of neo-liberalisation as a context-dependent-and therefore often contradictory-process of institutional reform that thrives on crises (see Brenner et al. 2010;Larner 2003). The Dutch case illustrates how such transformations may be inconsistent and non-linear. Coalition politics and institutional legacies have meant that social housing was never going to be abolished overnight. Instead, we have seen a series of adaptations, re-regulations and conversions that introduced market elements with the deregulation of housing associations. After 2008, when the previous reforms led to crises of legitimacy and financial risk, a new direction became evident: housing associations were re-regulated and there was political room to restrict access to social housing. The Dutch case highlights how crisis plays a role in institutional housing reform; it is commonly understood that crises may open up spaces for new politics (Jacobs & Manzi, 2013), but regulatory reforms are often temporary fixes that themselves lead to new rounds of crises and reform. Any policy analysis that looks at long-term processes should evaluate the role of crisis beyond the need for remedy-as neo-liberal politics suggest-and to view it, rather, as an integral part of the reform process.
Our review also sought to gauge the social and spatial implications of recent interventions. New taxes, the landlord levy and the 2015 Housing Act and its lead-up have led to demise in new construction and a relative decline of the stock. Particularly, high demand housing regions will see starker dualism between regulated and market rent. This has various implications. First, recent political developments have contributed to a further residualisation of the social-rental sector. The share of social-rental units has reduced and the tenure is increasingly housing the poorest and most vulnerable groups in society. Second, affordability remains an issue with many low-income tenants paying a relatively large share of their income on rents and utilities, although the Housing Act has taken steps to slightly improve rental affordability. Third, provisions in the Housing Act are set to increase spatial segregation, and particularly the spatial concentration of disadvantaged population groups, all the more with market forces currently exacerbating spatial inequalities (Hochstenbach & Arundel, 2019).
In that respect, more than the taxes and levies, the new means-testing rules, rent requirements and rent increases will have a lasting institutional effect: if almost one third of the housing stock is inaccessible to a larger share of the population, that share will be less inclined to support it politically and will call for further reductions in social rent. Such conversion and eligibility strategies have been part and parcel of welfare state reform for decades (Hacker, Pierson, & Thelen, 2015;Pierson, 1996). Also, the spatial concentration of social housing in low demand neighbourhoods and regions will likely result in new calls for regeneration and renewal. Past experiences indicate that this will be accompanied with a programme of tenure mixing which will put further pressure on social housing stock. It remains an open question whether these social and spatial 'effects' are unintentional or whether they are part of the neo-liberal politics of housing, as they will likely precipitate new rounds of re-regulation. Regardless, the socio-spatial effects of residualisation may produce a self-reinforcing tendency for institutional reform. While this tendency has been suggested as an outcome before (e.g., Forrest & Murie, 1983;Jacobs & Manzi 2013;Orchard & Arthurson, 2005), housing policy analyses typically pay little attention to how such a 'social-institutional feedback loop' may be an essential part of reform politics.
The social housing sector has been restructured but, we should also note, it has not been abolished. Housing associations retain a strong institutional presence in Dutch housing policy and it may see new conversions. The issue of housing affordability for middle-income households in high demand regions may also lead to new regulations that would bring the associations back to their original function: housing workers for whom the market does not adequately provide. Such a reversal, of course, would require social and political movements to rally around questions of housing (Hacker et al., 2015;cf. Chen, 2011;Romanos, 2014). Furthermore, with any re-regulation to help middle-income households, the question remains whether this will only come to benefit these households, or whether low-income groups will benefit as well.