Introduction

Homelessness is a complex policy issue. From a public administration perspective, it can be typified as both a wicked policy problem (Rittel and Webber 1973) and a persistent social issue (Maeseele et al. 2014). Northern European local governments all face similar policy challenges, such as addressing sleeping rough and promoting the flow of people out of temporary shelters (Benjaminsen et al. 2009). Over the past ten years, homelessness has made its way onto the political agenda of many northern European cities because, in a period of relative prosperity, it is no longer considered an acceptable phenomenon. Since various financial and moral strategies have been combined, new opportunities have arisen to tackle the problem of homelessness. As a result, several northern European metropolitan cities have developed new strategies to deal with homelessness (Anderson and Serpa 2013; Benjaminsen et al. 2009; Hermans 2012).

As little is known about the efficacy of these new local administrative and political approaches to the issue of homelessness, recently research has been conducted to gain more insight. The efficacy of local governance arrangements can be assessed on the basis of aspects such as the quality of the services offered and the related societal effects (Bovard and Löffler 2002; Fawcett and Daugbjerg 2012). The central research question was whether possible differences in the administrative-political approach to homelessness between northern European metropolitan cities lead to differences in the quality of the services offered, as well as in the related societal effects.

Theoretical Framework

Lauriks et al. (2008, 2013) offer a useful starting point on how to measure the quality of services for the homeless. On the basis of an international inventory, they selected a number of performance indicators for the effectiveness of a public mental health care (PMHC) system. As homeless people constitute a subgroup of the target group of such a PMHC system, some of these indicators are also useful for measuring the quality of services in the field of homelessness. This applies to the indicators concerning the degree to which services are of an integrated nature, the degree to which they also include psychiatric services, and the supply of temporary or permanent housing (Lauriks et al. 2008, 2013).Footnote 1

The societal effects of a local governance arrangement can be operationalized as the percentage of homeless people in the catchment area of the PMHC system (Benjaminsen et al. 2009; Lauriks et al. 2008, 2013). Another important indicator is the percentage of homeless people who do not make use of temporary or permanent housing services and sleep rough in the city (Benjaminsen et al. 2009). A final indicator concerns public views on homelessness, such as the response to the presence of homeless people in public places such as train stations and shopping centres, the extent and seriousness of the problem of homelessness, the perceived nuisance of homeless people, and the consequences of the gentrification of neighbourhoods for homelessness (Alexandri 2015; Hermans 2012; Moulaert et al. 2001). Public views on homelessness can both enhance and reduce the stigmatization of homeless people (Hodgetts et al. 2011). All of the relevant indicators are presented on the next page in Table 1.

Table 1 Indicators of the quality of services offered to homeless people and the societal effects of a local governance arrangement on homelessness
Full size table

What elements of a local governance arrangement targeting homelessness are relevant to explain the quality of the services offered and the societal effects? According to the general literature in this area, three elements of a local governance arrangement may be crucial: policy, structure and management (Bovard and Löffler 2009a, 2009b; Frederickson 2005; Head and Alford 2015; Hughes 2012; Peters and Pierre 2004, 2007). These three elements serve as a starting point for the formulation of a number of detailed hypotheses.

The first hypothesis pertains to the policy element of a local governance arrangement, which refers to the attempts that are made to serve one or more public interests. Policy goals and policy instruments are relevant in relation to this element (Benjaminsen et al. 2009; Bressers and Klok 2014; Dunn 2012; Fenger and Klok 2014). With respect to policy goals, a distinction can be made between external goals that refer to circumstances in society and internal goals that target circumstances within the administration itself. Internal goals may, for example, be aimed at improving the functioning of the government’s own organization and/or at aligning separate policy sectors such as health, housing, income and justice (Bressers and Klok 2014; Boesveldt 2015), thus avoiding fragmentation (Interdepartementaal Beleidsonderzoek 2003). Policy instruments are methods used by a government body to achieve the desired results. They should accord with the policy goals of the government body to ensure a sufficient level of goal attainment (Fenger and Klok 2014).

Thus, the first hypothesis states that the setting of internal policy to improve the functioning of the local government’s own organization and to align separate policy sectors with each other, as well as the precise matching of policy instruments to these internal policy goals, will have a positive impact on the quality of services, in terms of better integrated service coverage. Both mental health service coverage and overall service coverage for the homeless will be better.

Another relevant aspect of the policy element is the policy model, which concerns the moral and empirical assumptions underlying the policy at stake. Moral assumptions refer to the values that are considered relevant by policymakers and should be realized in their opinion (Dunn 2012; Tirion 2014). In the field of homelessness, these assumptions may, for example, mean that homeless people are not considered to be fundamentally different from other people and should be facilitated to live in the same circumstances as other citizens. Empirical assumptions pertain to the causes, features, magnitude and nature of a policy problem. Realistic and scientifically sound empirical assumptions are a precondition for achieving the desired effects (Dunn 2012; Tirion 2014).

Thus, the second hypothesis states that a local governance arrangement in which the policy element is characterized by normative assumptions emphasizing the equality of homeless people with other citizens and by empirical assumptions presenting a realistic and empirically grounded reflection of social reality will demonstrate a better quality of services for the homeless. There will be a higher level of integrated services and a greater supply of permanent rather than temporary housing.

The structural element of a local governance arrangement concerns, among other issues, the extent to which responsibilities and budgets in a policy sector have been decentralized to the local level (Fleurke and Hulst 2006). In addition, the extent to which these resources have been divided between local government bodies, private companies and non-profit organizations (Jessop 2004; SER 2010; WRR 2000) may influence the effectiveness and efficiency of public policy (Benner et al. 2004; Bouckaert et al. 2010; Olsen 2009).

In a centralized structure, decision-making powers are highly concentrated in a single organization or discipline (or at least in a smaller number). Responsibilities and budgets are not allocated to lower government levels and/or non-government organizations. Concrete tasks may, however, be outsourced to private organizations under clear directions. Complex social issues demand a minimum level of expertise and knowledge necessary for the formulation and implementation of the policy at stake. A centralized form of policymaking and policy implementation may lead to economies of scale (Byrnes and Dollery 2002; Dollery and Fleming 2006). In decentralized structures, policies that require a high degree of specialist knowledge may therefore score negatively with regard to the efficiency of programme spending (Fleurke et al. 1997; Painter and Peters 2010). This argument leads to the third hypothesis, which states that a more centralized structure will offer a better quality of services for the homeless in terms of efficiency.

Another relevant aspect of the structural element is the composition of the policy network (Sandström and Carlsson 2008). Compared to homogeneous networks, heterogeneous policy networks consist of different kinds of organizations or disciplines (Bressers 2008). The involvement of various relevant organizations in the network, such as housing corporations and welfare organizations, also referred to as mainstream agencies rather than specialized shelter services (Pawson et al. 2007), is probably conducive to a better quality of housing services for the homeless.

In addition, a management style of local government characterized by a pluralist vision of the relationship between the state and society offers all relevant organizations the opportunity to influence the quality of housing services for the homeless. All of the relevant organizations are offered relatively equal opportunities to influence the way in which social issues are resolved (Frederickson and Smith 2003; Pierre and Peters 2000). Thus, the fourth hypothesis states that a heterogeneous network in combination with a management style reflecting a pluralist vision of the relationship between state and society will offer a higher quality of services for the homeless – in terms of permanent rather than temporary housing – than a homogeneous network in combination with a corporatist vision.

The previous four hypotheses pertain to the influence of a local governance arrangement on the quality of services for the homeless, while the two final hypotheses developed below relate to the impact of the quality of services on the position of homeless people in society. The fifth hypothesis states that better mental health service coverage and better overall service coverage for homeless people will lead to lower numbers of homeless people and people sleeping rough in the catchment area of the PMHC system and to more positive public views on homelessness. If homeless people intensively use the medical, psychological and financial aid offered to them, then a substantial number will be able to eventually escape homelessness. In that case, the number of homeless people, as well as those sleeping rough, will decrease in the catchment area of the PMHC system. Furthermore, the intensive use of medical, psychological and financial provisions by the homeless will also positively influence public views on the issue of homelessness, due to the improved conditions and behaviour of homeless people.

The final, sixth hypothesis states that better temporary and permanent housing for homeless people will reduce the number of people reporting as homeless, as well as those sleeping rough. The provision of housing services – as a consequence of participation in the policy network by housing providers – implies both the prevention of evictions as well the rapid availability of permanent housing. If local government, also through its network, succeeds in providing a large proportion of homeless people with adequate housing, fewer people will feel the need to report as homeless, sleep rough or reside in institutions.

The theoretical framework, including the six hypotheses, is outlined in Fig. 1.

Fig. 1
figure 1

Theoretical relationship between aspects of a local governance arrangement for homelessness and aspects of the arrangement’s efficacy (Source: Boesveldt 2015)

Full size image

Comparative Case Studies

Comparative case studies were conducted in three northern European metropolitan cities to empirically test the six hypotheses. The case studies were undertaken in Copenhagen, Glasgow and Amsterdam. These metropolitan cities are comparable in terms of prosperity and they have demonstrated activity in the setting of policies to address homelessness over the last decade. However, they vary considerably with respect to their governance arrangements in the area of homelessness, differing with regard to governance aspects that are assumed to influence the quality of services and the resulting societal effects (Boesveldt 2015). This is shown in the table below (Table 2).

Table 2 Three western administrative approaches to homelessness (Source: Boesveldt 2015)
Full size table

On the basis of the above trichotomy, Copenhagen, Glasgow and Amsterdam were selected as representatives of the Scandinavian, Anglo-Saxon and Continental traditions respectively.Footnote 2 Each of these cities also offered the right conditions to analyse the relationships between certain aspects of the local governance arrangement and the quality of service delivery and the resulting societal effects, above all because the local authorities were willing to cooperate with the investigation. In each city, the relevant policy documents were studied and confidential semi-structured interviews with ten relevant stakeholders (policymakers, politicians, executive institutions, clients) were carried out. The results of the three case studies offer a well-documented picture of the different aspects of the theoretical framework outlined above, and provide the data necessary to test the six hypotheses. The first step in this process is to examine the quality of service delivery and some of the resulting societal effects.

Variation in the Quality of Service Delivery and Societal Effects

Table 3 presents data on the quality of service delivery and the resulting societal effects of governance arrangements on homelessness in Copenhagen, Glasgow and Amsterdam. Most of the data is of a quantitative nature, although some is qualitative.

Table 3 Quality of service delivery and societal effects in Copenhagen, Glasgow and Amsterdam in 2013 (Source: Boesveldt, 2015)
Full size table

Copenhagen has been able to service half of its homeless population with integrated care but only a small number of those with severe mental health needs. Copenhagen has also been able to temporarily improve the housing situation of half the homeless population; in approximately 20% of cases, permanently. Elements of inefficient spending were observed. Ten to 15 % of homeless people in Copenhagen sleep rough. There is local support (as opposed to national regulations) for a supervised drug consumption room, which is related to changing (more upmarket) demands in certain neighborhoods (gentrification).

Glasgow has been successful in diverting people from specialist to generalist services. Mental health service coverage is probably higher than reported here, since it was not possible to include services for substance abuse. The number of people being permanently housed is greater than the number provided with temporary solutions. In sheer numbers, sleeping rough in Glasgow occurs more frequently than in Amsterdam or Copenhagen. However, as a proportion of the total population of the city this is not the case. The views of the public, according to the respondents, range from indifference to complaints and prejudice.

In Amsterdam, one’s public mental health need determines the extent to which integrated care is being offered to the person concerned. This is reflected in the rather high level of mental health service coverage for homeless people. Integrated services are highly likely to offer an improvement in the housing situation (73%); however, only a small number are offered permanent solutions (17%). Proportionally, sleeping rough occurs just as frequently as in Copenhagen. The public discourse in Amsterdam is centred on issues of safety, illness and nuisance. The addressing of these needs is valued positively by the public.

The data presented in the tables above show significant variation in the quality of service delivery and the societal effects across the three cities. In the next section, it will be determined whether this variation is explained on the basis of differences in local governance arrangements by testing the six hypotheses.

Testing of Hypotheses

All of the six hypotheses were confirmed by the case studies. In relation to each hypothesis, the underlying theoretical reasoning was actually found in all three cases. The first hypothesis stated that the setting of internal policy in order to improve the functioning of the local government’s own organization and to align separate policy sectors with each other, as well as a precise matching of policy instruments to these internal policy goals, would have a positive impact on the quality of services in terms of better integrated service coverage. Both mental health service coverage and overall service coverage for the homeless would be better.

In the Copenhagen case, no objectives were set or attained in close cooperation with any of the adjacent sectors. Administrative respondents therefore found little support from other departments in their attempt to implement their policy goals: ‘We have a lot of arguments with the other departments [trying to convince them] to go in and do what we think they have to do, but they are not obliged to do so’ (Authority Respondent). In its operation, the homelessness sector is rather isolated from other municipal services, for example those services that are responsible for social benefits: ‘They don’t necessarily get state benefits because they are too chaotic for that. Mainly because they fail to meet with social workers or job centres’ (Authority Respondent). This same issue of fragmentation also arises with regard to regional services providing mental health services: ‘Someone should do something, but it isn’t our job. It’s their job for the mentally ill’ (Authority Respondent). None of the policy domains closely related to homelessness policy (psychiatric treatment, income, police or housing) appeared to be sufficiently involved and, accordingly, the quality of output regarding mental health service coverage was particularly low (10%).

In Glasgow, there was a preventative focus in relation to both the policy goals and policy instruments. Glasgow has been successful in providing homeless people with information/advice and has thereby prevented homelessness. The search for partnerships and cooperation is often mentioned in policy documents. In this way, health services operating regionally have also been involved in the city’s approach. The goals of the city were mirrored in the prevention of evictions by its housing associations and legal partners.

In Amsterdam, as in Copenhagen, a lack of coherence has been observed between policy goals and instruments. The low overall service coverage (37%) is due to the policy goals being targeted to prevent homelessness, but few additional instruments have been introduced to prevent homelessness among those with lighter care needs or other groups, or to deal with homeless people who do not have public mental health needs. The policy instruments remain strongly aimed at people with health needs and/or those posing a public safety risk. In addition, the internal goals are aimed at the mental health and justice groups. Finally, this case does relatively well in terms of the involvement of the adjacent policy areas required to achieve the internal goals, especially the health and justice sectors, and consequently it appeared successful in assessing people with priority needs, in particular, public mental health needs,Footnote 3 also reflected in a high level of mental health service coverage (87%).

The second hypothesis stated that a local governance arrangement in which the policy element is characterized by normative assumptions emphasizing the equality of homeless people with other citizens and by empirical assumptions presenting a realistic and empirically grounded reflection of social reality will offer a better quality of service for the homeless. This would be reflected in a higher level of integrated services and a greater supply of permanent rather than temporary housing.

In Copenhagen, the policy model can be characterized as having a rather restricted (non-realistic) focus on treatment results, which is contrary to the harm-reduction based instruments proposed (Housing First).Footnote 4 Another Copenhagen approach concerns a reluctance to interfere with homeless people’s lives, as they are perceived as strong individuals who lead a good life on the streets. The failure to target more vulnerable groups, which is reflected in a mental health service coverage for homeless people of 10%, can thus now be explained by this reluctance to interfere, as well as the less pragmatic (treatment focused) assumptions of the policy. Contrary to its policy goals, most homeless people, when housed, are given temporary accommodation, expressed by a ratio of 0.48.

In the Glaswegian case, the policy model is characterized by a generalist approach (‘Anyone is only three pay checks away from being homeless’)Footnote 5 as well as the idea that people who could have prevented themselves from becoming homeless are not entitled to homeless services and rapid rehousing. Homeless people are actually rehoused (34%) the most rapidly, and stay in temporary accommodation (11%) for relatively short periods. Overall service coverage (55%) and mental health service coverage (65%) exhibit the most moderate picture, with no notable exceptions.

Under the Amsterdam policy model, reference has been made to a specialist (medicalized) image of homelessness. The Amsterdam discourse has also witnessed an increase in a security domain mentality in relation to social relief (‘The dangerous homeless’). In addition to the evidence presented under the first hypothesis, these models explain the bias in service coverage towards mental health needs (87%). In terms of housing, there is also a bias towards institutionalized temporary and specialized provisions to care for the homeless, with a ratio between permanent and temporary housing of 0.29.

The third hypothesis stated that a more centralized structure would provide a better quality of services in terms of efficiency. This was confirmed by the research findings. In the Scandinavian case, a multi-level structure was found,Footnote 6 which was characterized by a complex financial configuration, at times offering the opportunity and room for initiatives, but also preventing initiatives that did not fit a specific direction or policy goals. There had been discussions between the local homelessness policy unit and the national Ministry of Social Affairs (MSA),Footnote 7 which led to insufficient returns on means intended for the purpose of homelessness; funds that were spent on psychiatric care that could have been spent elsewhere; and there was non-accountable, untargeted policy. For example, in spite of the decentralizing trend, the national MSA continued to generously finance a local social work organization. A particular national financial arrangement allowed this Copenhagen service provider to work independently of the local Copenhagen Homelessness Strategy, disregarding municipal guidelines and targets.

In the Glaswegian case, centralized and detailed policy with respect to rules and targets created a clear task at the local level, with the degree of detail leaving little room for local discretion by local civil servants with regard to interpreting policy.Footnote 8 Accordingly, accountable, targeted policy was found and the prioritizing of prevention resulted in additional efficiency. Budgets to provide voluntary services in this case consisted of a combination of local funding and small local and/or national grants. These kinds of public-private budgets reflect a particular stance on public and private responsibilities.

In the Amsterdam case, at the time of this study, the municipal government was expecting that a big trend towards decentralization would result in several inefficient outputs. We saw processes of exclusivity in the budget, resulting in insufficient returns on means intended for the purpose; a lack of prioritizing on prevention; a decrease in expertise; and conflicting silos and networks that all impacted negatively on efficiency. This might be due to the fact that after decentralization, only specialists are able to assess the precise costs of treatment and provisions. Amsterdam’s experts talked about ‘Different prices for the same things’ and ‘No one really knows’. Another example of inefficiency is that according to Amsterdam’s respondents ‘Way too many’ local care networks had been organized in parallel, and based on the same principles. Every specialist discipline has its own team, which creates the image of six separate teams surrounding one individual household in need rather than six helpful professionals providing integrated support.

The fourth hypothesis stated that a heterogeneous network in combination with a management style reflecting a pluralist vision of the relationship between the state and society delivers a higher quality of services – in terms of permanent rather than temporary housing – than a homogeneous network in combination with a corporatist vision. In the Copenhagen case, indeed homogeneous, long-standing relationships between the state and providers of sheltered housing were traced, which for a long time were only managed at the national level. During the interviews it was acknowledged that active networking by large NGOs (shelter institutions, temporary housing) contributed to close relationships. When asked about the inter-ministerial parties involved, it was explained that there was a connection, but it was not formal and cooperation was on an ad-hoc basis. However, when asked about inter-ministerial cooperation, it transpired that some relevant ministries, for example, housing and justice, were not involved at all. This homogeneity and familiarity is reflected in the relatively low ratio of permanent rather than temporary housing (0.48) compared to Glasgow (2.9).

In the Glaswegian case, a more heterogeneous network composition was witnessed that included, in particular, detached relationships with providers of shelters and a clear focus on and more investment in relationships with housing partners. One example of the focus on housing partners who could contribute to policy is the current strategic policy network that emerged after the 1990s. Today, the Glasgow Homelessness Planning and Implementation Group also includes social housing providers. There are four housing associations involved, and it was indicated that the number of housing associations is likely to increase. The involvement of housing providers also appears to be at the right strategic level: ‘We have a joint Board [...] it’s quite high level, you know, taking that responsibility [for] leadership as we’re going forward’ (Social Housing Provider). The homelessness goals within the housing strategy may be just as targeted by the local authority as by its partners. As one housing provider stated poetically: ‘They’re the golden thread going through all our activity’ (Social Housing Provider). The Glaswegian respondents also indicated how the social housing associations had been more successfully involved with the city than the private rental sector. The latter was specifically referred to as having ‘frosty relations’ with the housing department. To address this issue, Glasgow completed a tender in the private sector in an attempt to discharge its duty with respect to long-term housing solutions. Previously, the private sector could only be used for temporary accommodation.

In the Amsterdam case, the network was also found to be more homogeneous than heterogeneous, comprising long-standing relationships and institutionalized coalitions. The network is relatively closed, and new parties rarely become involved. At the time of this study, shelter parties were part of the network but housing parties were not. The exclusivity of the network makes representation of the field difficult at times, as is the dissemination of expert information through the network and to local care networks. Respondents who are part of the network, as well as other stakeholders, referred to the network of social relief as ‘A closed world’, as ‘We know ourselves’ and ‘Very isolated’ ‘Which makes it hard for people to look beyond the framework of social relief’. Amsterdam homeless people with lighter care needs are diverted to the local care networks. Because these local care networks are not connected to the social relief network these people do not gain access to it, which explains why the homelessness of people with lighter care needs is not being prevented. In Amsterdam, the insularity of the network is again seen in the relatively low ratio of 0.29 between the number of people housed permanently and the number housed temporarily compared to Glasgow.

The fifth hypothesis stated that better mental health service coverage and better overall service coverage for homeless people would lead to a lower number of homeless people and people sleeping rough in the catchment area, and to more positive public views on homelessness. The theoretical reasoning underlying this hypothesis was most clearly found in the Glaswegian case, where sleeping rough was relatively low (8% of homeless people). A relatively high service coverage (55%/65%) is likely to account for this to a reasonable extent. According to the local authority, neighbourhood complaints were justified and were taken seriously, and stigma was seen as an issue to be seriously addressed. For the outcome in relation to sleeping rough, it was noted that there was a ‘low tolerance of sleeping rough’ by the public, indicating that this is not considered acceptable.

In the Copenhagen case, the number of homeless people sleeping rough (15%) might partly be explained by the low mental health service coverage (10%), indicating that mentally ill people on the streets are not assisted by municipal professionals, while regional professionals perceive their tasks in relation to these people differently. These professionals made specific reference to this issue of the most vulnerable sleeping rough. However, the views of the general public in this case merely concern the issue of gentrification. As a result of inhabitants being confronted with drug use in their streets, support for a drug consumption room was expressed by the public.

In the Amsterdam case, the actual homeless rate (0.19%) and street homelessness (12%) were the lowest of the three cases. Also, in this case, a positive change in the needs of the homeless people who come forward is visible, as high care needs are being served (87% mental health service coverage). Communication, the scope of measurement, research and the knowledge base of policy in this case is limited mostly to numbers concerning the reduction of nuisance, with much being invested in the public visibility of reduced nuisance. Public views also reflect this image of the homeless as the neediest but also the most dangerous individuals in society.

The final, sixth hypothesis stated that better temporary and permanent housing for homeless people reduces the number of people reporting as homeless, as well as those sleeping rough, and also to more positive public views on homelessness. In Copenhagen, the relatively low figures in terms of temporary housing (17%) and permanent housing (35%) accord with its poor outcomes, in terms of increases both in the numbers of homeless people sleeping rough and the number of people reporting to be homeless.

In Amsterdam, a bias was witnessed towards institutionalized temporary specialized accommodation (56%), a policy based on public concerns about street homelessness and nuisance and a belief that the public required protection. This stemmed from previous, well-known problems, which began to disappear from the streets around 2006.

In Glasgow, positive housing outcomes appear not to be limited to positive percentages in relation to the rehousing of homeless people (11%/34%), and also show a 1 % drop in applications for assistance concerning homelessness, specifically from people who had previously had a social housing tenancy (Danny Philips Associates 2009). The low tolerance for rough sleeping in this case is also indicative of public support for this housing policy. The three case studies thus also confirm the final hypothesis.

Conclusions and Discussion

This research addressed the question of whether possible differences in the administrative-political approach to homelessness in a number of northern European metropolitan cities had any impact upon the quality of the services offered as well as other related societal effects. The general conclusion is that public governance does matter. What a local authority does to address homelessness in conjunction with the activities of other parties has an impact. All hypotheses were confirmed.

It was found, for example, that the setting of internal policy goals, as one element of a local governance arrangement, is conducive to improved levels of integrated service delivery. Internal policy goals help prevent the multi-level fragmentation of responsibilities and budgets; consequently, the level of integrated service delivery is higher. Another example of the relevance of local governance arrangements to the quality of services and related societal effects concerns the fact that decentralizing trends appear to have a negative impact on the realization of higher levels of integration, such as more integrated and customized services. Further evidence was found that the network structure influences the efficacy of the overarching governance arrangement in terms of the improved housing situation of homeless people. A more heterogeneous constellation of the network, which means a more mixed composition of specialists and generalists in the network, appeared to lead to better outcomes in terms of the housing situation of homeless people.

The research findings justify a plea for an integrated, instrumental approach to governance arrangements for the homeless. During the course of this research, the significance of disenchantment with current solutions became apparent, and that it is important to address the issue on the basis of more solid empirical evidence and in a more rational way. The initial research interest focused on the values underlying the governance of homelessness. However, the study of these values actually revealed a picture in which addicts were being blamed for their drug use and untreated psychiatric patients were blamed for not having paid their rent. These moralizing opinions create obstacles to an instrumental approach. It was for this reason in particular that contractarian concepts related to governance were used, such as efficacy and efficiency, in looking at solutions. It was considered that this approach would be most capable of rationalizing the social issues at stake and in so doing better serve the interests of the homeless.

Finally, interventions to improve the functioning of the social relief sector should focus to a considerable degree on the structural elements and on the governance levels at which the adjacent responsibilities for homelessness are organized, such as the responsibility for mental health policy, youth policy and learning disability policy. The main focus of homelessness policy and sheltered institutions should then be to identify relevant trends and transfer the homeless – or ‘push them back’ – to the adjacent areas which are primarily responsible. The social relief sector should function as a trampoline not only as a safety net.