Performance agreements to ensure societal legitimacy in the social housing sector; an embedded case study of implementation in the Netherlands

In 2015 the Housing Act was revised in order to further regulate the social housing sector in the Netherlands and thereby improve the steering possibilities for the central government to coordinate housing associations. This included local performance agreements for social housing policy obtaining a legal status. By introducing this policy instrument central government seeks to facilitate and ensure the tri-partite cooperation between municipalities, housing associations and tenants’ organisations in order to release funds by housing associations for social benefit. This should improve the position of municipalities and tenants’ organisations in social housing, and improve legitimate policy making. In this paper the main research question is: How are local performance agreements implemented targeting increased societal legitimacy in local social housing policy making, and what are its strengths and weaknesses in three selected cases in the Netherlands? A case study research design was used involving three local embedded case studies. As a theoretical framework the Contextual Interaction Theory was used. Data collection involved expert interviews and review of policy documents. Results reveal several weaknesses that impede the implementation of performance agreements, including issues in the broader governance regime and context, as well as issues with the inter-organisational structure and stakeholder interaction regarding the tri-partite cooperation between the key actors. This has to do with the precarious role of the tenants’ organisations in the process, and the local housing policy as the basis of local performance agreements. Results also show that implementation of performance agreements is more difficult in cities with dense urban areas.


3 Introduction: ensuring democratic and social legitimacy in the social housing sector
Almost one-third of the total housing sector in the Netherlands can be considered social housing (Ministry of the Interior and Kingdom Relations 2016). In the Netherlands housing associations take care of the provision of affordable housing for low income households since the 19th century. Their role, task and position have thereby always been multiform and changeable. Already since the introduction of the Housing Act in 1901 the role, position and task of housing associations are subject to debate. In the past decades, the position of housing associations has changed from tight subsidised and regulated organisations to financially independent, private non-profit organizations that partially have public sector tasks (Nieboer and Gruis 2016;Van Bortel and Elsinga 2007). After the (financial) privatisation of the social housing associations in 1995, the financial relations between the government and the housing associations were severed (Snuverink 2006). The financial and operational autonomy of social landlords implied that the Dutch government's ability to influence the behaviour of housing associations became limited, compared with the past. Legally, the government still provided official supervision, but the state entrusted the monitoring of performance by social housing associations mainly to the self-regulating capacity of the sector and to voluntary performance agreements between local governments and social housing associations (Elsinga et al. 2021). Housing associations were allowed to act as independent non-profit organisations and obtained a lot of freedom. They provided rental housing for low and middle income groups and they were also allowed to build and manage dwellings by means of a so-called "revolving-fund" model (Blessing 2015). Selling newly built and existing dwellings became a measure to generate income for carrying out their social tasks (Nieboer and Gruis 2016). But there was little insight in the social performance of housing associations (Boelhouwer and Priemus 2014;Veenstra et al. 2017). After the Millennium change severe misconduct manifested in parts of the sector, which manifested in incidents like administrative failures and financial mismanagement. This drew political attention, and in 2013 a Parliamentary Survey was issued. The causes for misconduct and the general performance of the sector were critically examined. The results called for revision of the institutional and regulatory structure of the sector. Its main legal act, the Housing Act was finally revised in 2015.
The result-the so-called "Revised Housing Act"-aimed at further regulating the activities of housing associations by providing stricter rules and boundaries, thereby aiming to strengthen the supervision of the social housing sector (Rijksoverheid 2015). Another important goal of the new Housing Act was to force housing associations to focus on their primary task: providing affordable housing for low income groups. The provision of owner-occupied and rental housing for middle and higher income groups became the responsibility of only private developers and investors. Especially for middle income groups, who are not able to access the social rental sector any more while many cannot afford to buy a house, private investors are supposed to provide so-called middle segment rental housing. This development can be characterised as a move away from an integrated unitary rental market towards a dual or residual rental market (Elsinga et al. 2021). The distinction between unitary and residual rental housing markets was developed by Kemeny (Kemeny 1995), who argues that unitary systems promote direct competition between the commercial and the non-profit rental sector on an integrated rental market. In such systems the non-profit rental sector is not intended solely for low-income groups. Dual rental systems on the other hand, are characterised by a market strategy based on profit, which prevents direct competition between the profit and the non-profit sector. Accordingly, the government separates the non-profit sector, which works on a cost-price basis, from the commercial rental market and uses it as a safety net. This model prompts the government, in an effort to maximise market forces, to separate the non-profit sector from the housing market and subject it to stringent regulation (van der Heijden 2013).
In this present paper however, the focus is not on the broader development of the rental market, but on a specific and important policy instrument within the Revised Housing Act 2015: local performance agreements. This instrument, which already existed before 2015 as a voluntary instrument-and was already used by 64% of the municipalities and 75% of the housing associations in 2014 (Ministry of the Interior and Kingdom Relations 2019)-obtained a legal status in the Revised Housing Act 2015. This is consistent with the development from a highly independent and more or less self-regulated social housing sector towards a more regulated and controlled sector. Conceptually speaking, performance agreements pertain to a (policy) instrument that entails a reward system based on target organisations meeting expected performance. These agreements look at actual or future performance. Based on performance a funding authority signs rewards a public service delivery organisation, often in the form of a financial sum or in case performance is not met, a financial penalty (De Boer et al. 2015). An important element of performance agreements pertains to setting specific performance indicators, which is necessary to assess the 'results' of service delivery by a given public body (Boyne and Law 2005). Performance agreements are used in several economic sectors including local government service delivery, public housing, and higher education.
An important goal of using performance agreements, as key policy instrument within the Revised Housing Act 2015 is to support and incentivize the formulation of local social housing policy. In this process three parties participate: housing associations, municipalities, and tenants' organisations. Within local performance agreements housing associations, municipalities and tenants' organisations are expected to establish how the parties involved contribute to the realisation of the local (social) housing objectives within a given period. Here, an important aim of performance agreements is to increase the link between the investment capacity of housing associations and the local social housing needs to ensure social benefit. Hence, by means of this policy national government seeks to facilitate and ensure the co-creation of (social) housing policy between municipalities, housing associations and tenants' organisations. More specifically, it should empower the role of municipalities and tenants in the social housing sector vis-à-vis housing associations (Ministry of the Interior and Kingdom Relations et al. 2015), leading to a situation in which social legitimacy of policy making increases, municipalities and in particular tenants are more empowered, which leads to an outcome which is eventually accepted better by members of the latter. Legitimacy here pertains to, "a generalized perception or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, beliefs and definitions" (Suchman 1995) (p. 574). More specifically, societal legitimacy concerns the acceptance and compliance to norms among members of a social community (Bokhorst 2014), in the present paper this pertains to tenants. We argue that this is only possible when all three participants of the local policy making process perceive the decision-making process as fair and just (with sufficient involvement of all three parties), including experiencing a high degree of procedural fairness (Baxter 2017;Jenkins et al. 2016;Wüstenhagen et al. 2007).
The Dutch system allows for variation between municipalities in the approach how performance agreements are implemented. For example, in some municipalities there is great emphasis on supporting a co-creation process between the municipality and housing associations, whereas in other municipalities housing associations bid on an offer from the municipality. In other municipalities there are multiple housing associations active, whereas in others there is only one monopolist involved. Moreover, the duration of the agreement varies between 1 and 3 years, and there is great variation in scope. Figure 1 presents the process of establishing the afore mentioned performance agreements in the Dutch social housing sector.
The Revised Housing Act clearly defines the procedure for local cooperation, but there are no requirements for the way in which cooperation is formed. Also, the content of the performance agreements is not prescribed, because the local situation can vary greatly. As a consequence, the performance agreements show a wide variety in terms of character, content and duration, both before and after the Revised Housing Act coming into force (Tiggeloven et al. 2016). For example, there will be more co-creation between the municipality and the housing association, if the housing association has been involved in the drafting of the municipal housing vision. In addition, there are municipalities in which more than one housing association is active and there are housing associations that are active in multiple municipalities within a housing market area. Such differences affect the position of the parties involved in the process. For example, it may lead to the choice of a regional housing vision as the basis for making performance agreements ). There are also differences regarding the period agreed upon. Many parties make new agreements with each every other year, but it is increasingly common for parties to make multi-annual agreements that are updated annually on a few points (Tiggeloven and Klouwen 2019).
A comparison of the performance agreements made before and after the Revised Housing Act coming into force (2015) showed that the proportion of agreements signed by tenants' organisations has risen sharply and that more agreements are being made on the way in which housing associations, municipalities and tenants are co-operating. Moreover, there has been a marked increase in the number of concrete agreements. Municipalities and Fig. 1 Step-by-step process for the establishment of performance agreements (Ministry of the Interior and Kingdom Relations et al. 2015) housing associations seem to have used the revised law to make performance agreements more SMART (i.e., specific, measurable, achievable/attainable, relevant, and time-bound) and therefore more measurable (Tiggeloven et al. 2016).
Although the introduction of performance agreements appears to be promising in ensuring societal legitimacy, and facilitating the co-creation of (social) housing policy, there is limited insight into how negotiations take place, and what its results and impact are. Moreover, little has been documented on the challenges facing the implementation of this type of performance agreements, let alone in the social housing sector. Therefore, the present paper seeks to bridge this gap by exploring and documenting the challenges facing the implementation of local performance agreements. Due to the introduction of this new format of establishing local performance agreements for housing policy new roles, responsibilities and interrelations between municipality, housing associations and tenants' organisations need to be explored. Therefore we seek to answer the following research question: How are local performance agreements implemented targeting increased societal legitimacy in local social housing policy making, and what are its strengths and weaknesses in three selected cases in the Netherlands?
This question will be answered by applying a policy studies' approach. First, performance agreements are introduced as a policy instrument in Sect. 2. In the present study the theoretical framework of the Contextual Interaction Theory is used to explore and analyse the implementation process. In Sect. 3, the rationale of this theory is addressed. Subsequently, in Sect. 4 research design and methodology are presented. Next, in Sect. 5 the results of the case-study evaluations are provided, and strengths and weaknesses are disclosed. Finally, in Sects. 6 and 7 discussion and conclusion are presented, respectively.

Performance agreements as a policy instrument
By means of local performance agreements the Dutch central government seeks to facilitate and ensure cooperative, but non-permissive, tri-partite networks of municipalities, housing associations and tenants' organisations for the co-creation of social policy and to strengthen the supervision of the housing associations. Although performance agreements have been used in the Dutch social housing sector since the privatisation of housing associations in the 1990s, the current ones differ reasonably from the previous ones, both in objectives, actors involved, and in use. Moreover, whereas performance agreements were permissive and voluntary prior the Revised Housing Act (2015) the latter has made them imperative (Ministry of the Interior and Kingdom Relations et al. 2015). Before further heading into the implementation of performance agreements in the Dutch housing sector information on the meaning, the implementation process, and general experiences with performance agreements in the academic literature are addressed.

The value and effects of performance agreements in public policy
Since the 1980s, when many public sector tasks were privatised or decentralised, there has been a lot of attention to use performance measurement in public service delivery. By means of performance agreements public service providers can be held accountable for the policies they implement and the services they deliver. The introduction of performance agreements in the (semi) public sector can be viewed in light of the emergence of New Public Management (NPM) in the 1990s and early 2000s (Boyne and Law 2005). At the time performance agreements were considered a novel tool of governance, in particular for public service delivery (James 2004). First introduced in the United Kingdom it also found its way to continental European countries like The Netherlands.
In the United Kingdom Local Public Service Agreements were introduced as to increase priority setting and information about performance and incentive effects of local government public service delivery. In this reward system national government acted as a funding organisation, and local governments as agents with whom national government would agree on performance, and establish contracts. In case performance was met financial rewards would be funded to local governments. The agreements, however, would typically focus on the outcomes of service delivery, rather than processes, inputs or outputs. Moreover, performance indicators used do not by definition give an accurate reflection of actual outcomes (Boyne and Law 2005). A study by James (James 2004) shows that direct incentive effects of performance agreements were generally weak, and that responsibility for performance was limited, with targets viewed as minimum pledges of performance rather than incentives for improving performance. The credibility of the system also suffered from blame avoidance and blame shifting (Ibid.). In addition, it was felt that organisations involved started focusing heavily on cash incentives and target achievements (Young 2005), valuing quantity over quality. The system led local authorities to use the system predominantly as a means to win extra funding. The scheme also suffered from asymmetry, a simple 'tick the box' mentality to performance achievement, and cultural factors in the civil service that impeded performance (Ibid.).
In the United Kingdom New Public Management was also introduced to the public housing sector with housing associations adopting novel management agendas of performance indicators in their model of service delivery. Experience with the implementation of the agreements reveals a conflict between the social purposes of public funding and the business practices of the housing associations that led to social exclusion, and a parallel between 'what works' and 'what matters' (Sprigings 2002). In the Netherlands performance agreements were introduced in the social housing sector as a new procurement method for maintenance of the housing stock. In this case a (semi-public) housing association would reward a contractor based on performance. Performance agreements were developed to replace traditional tenders by longer term cooperation (Straub 2005). When compared to the way performance agreements are implemented under the Revised Housing Act 2015 the performance agreements targeting tenders with subcontractors are of a different nature, having other goals and focusing on a different actor setting.
Another domain in which performance agreements have been introduced is higher education. De Boer et al. (2015) analysed the implementation and effects of performance agreements in fourteen higher education systems across the EU and found that the design and the process around implementation of performance agreement schemes differ strongly between countries and systems. Implementation depends on situational context, but also that that in the countries where the schemes were implemented performance agreements had eventually become accepted. Another study (Jongbloed et al. 2019) analysed the introduction of performance agreements among universities and universities of applied sciences. Implementation among universities was found to increase quality and completion in education; however, universities of applied sciences experienced problems with regard to completion rates. Pros related to performance agreement implementation include putting outcome indicators (like students' success rate) more prominently on the agenda of institutions, and experiencing increased transparency and accountability. Cons included issues like the decline of the universities' autonomy due to the setting of national targets, use of mandatory indicators, and the bureaucracy and administrative costs that go along with the focus on indicators.
Additional studies on the possible effects of performance measurement in the public sector show that such performance measurement systems are also able to increase accountability, credibility and legitimacy besides transparency. However, the use of such systems might also lead to sub-optimal effects like setting less ambitious goals to make sure they are feasible (the so-called tunnel vision; focussing on the objectives agreed on while losing sight on other objectives; emphasis on short term targets at the expense of long term objectives (Bruijn 2006;Johnsen 2005;Smith 1995). Experiences also show that public sectors that have implemented performance agreements systems are inclined to maintain performance agreements, once the approach has been introduced (Commission 2014; De Boer et al. 2015). Table 1 presents an overview of propositions from academic studies that address implementation of performance agreements.

Implications of the Revised Housing act to formulating performance agreements
By using local performance agreements in the social housing sector the Dutch national government seeks to create cooperative, but non-permissive tri-partite collaboration between municipalities, housing associations and tenants in order to release funds from housing associations for social benefit (Koopman and Hopstaken 2017;Relations et al. 2015). By implementing performance agreements it is intended that the position of municipalities and tenants' organisations in social housing are strengthened to ensure democratic and social legitimacy in the social housing sector. Although local performance agreements have legal power following the introduction of the Revised Housing Act in 2015, national Table 1 Propositions based on the experiences with performance agreements (PAs) in the public and semipublic sector Propositions with a positive connotation to PAs PAs make public service delivery more transparent Through PAs public service organisations can be held more accountable PAs lead to an increase of quality in public service delivery PAs increase performance of public service delivery The higher the financial incentive of the PA the higher the public service quality delivered by a public service organization Propositions with a negative connotation to PAs PAs lead to an increase in the bureaucracy and administrative costs of public service delivery The introduction of PAs leads to a conflict of public service organisation focusing between public values on the one hand and business practices on the other, causing adverse effects PAs impact the adoption of key performance indicators on policy agendas Due to the use of (mandatory) indicators the autonomy of certain public bodies decreases Introduction of PAs might lead policy makers to set less ambitious policy goals, focusing on short term goals (and costs) while losing sight of long term benefits Implementation of PAs is not straightforward; it depends on situational context, and varies strongly between countries government has not defined how local performance agreements should look like. This is left to the three main actors at the local level in repetitive deliberative negotiation rounds. National government facilitates and stimulates these processes, resulting in the establishment of performance agreements, for example by setting clear deadlines for the process of decision-making and by providing the possibility to refer disputes to an independent committee (Blok 2015). Performance agreements are intended to attain the following goals. First, they should lead to a change in the relationship between local authorities and housing associations. Performance agreements should empower local authorities to have a stronger position in influencing the local housing development, with the local housing policy (resulting from deliberative negotiations) forming the formal basis to local goal setting and planning of local housing projects. The local housing policy will also form the basis of future local performance agreements in the construction, renovation and maintenance of local housing stock. The Revised Housing Act makes the inclusion of tenants (or their representatives) in the performance agreement negotiation process mandatory (it was awarded a legal status), giving a the tenants a stronger position in these processes. This is done to strengthen the accountability towards the users of social rental housing and to ensure the housing associations' social legitimacy. Another issue pertains to strict deadlines set in the process on which housing associations need to provide its offer to the tenants' organisation and municipality. A deadline is set for the date when the final set of mutual agreements should be established (Ministry of the Interior and Kingdom Relations et al. 2015).
Central government has set priorities that should be considered substantive requirements for the housing association's offer on the local housing policy. These priorities ought to give direction to the content of the local housing policy and subsequently the performance agreements. Although the priorities set by the central government do not have a legal base, it is intended to make sure the parties involved in the local network will contribute to these priorities in creating performance agreements (Blok 2015). Priorities for 2016-2019 pertained to: affordability and availability of dwellings for households who are depending on social housing sector (i.e. low-income households); realisation of energy-efficient social housing stock in accordance with targets of National Energy Agreement and Energy Saving Agreement for the Social Housing Sector, accommodation of urgent target groups, and realisation of housing-care facilities for elderly people and other vulnerable communities.
According to the revised Housing Act housing associations should provide the municipalities and tenants' organisations with (financial) information which these parties consider necessary to assess the offer made by the housing association. Amongst others the Housing Act holds that the municipalities will be provided annually with an indication of the investment capacity of the housing association by the minister. Possible disputes that impede the establishment of performance agreements can be submitted to the minister, who then makes a decision regarding the further continuation of the process. However, the primary responsibility for solving a dispute lies with the three main actors locally.

The Contextual Interaction Theory: a conceptual framework for policy implementation
Because the present paper focuses on the implementation of a policy instrument (in this case performance agreements) a theory is used that accommodates understanding of policy implementation. Converging insights from a long tradition in policy implementation theory (Bressers et al. 2000;Bressers and Klok 1988;Bressers and O'Toole Jr 1998;Hill and Hupe 2002;O'Toole 2000) the Contextual Interaction Theory (CIT) is used in the present study. According to Bressers et al. (2000) policy implementation research should focus on the interaction processes between the actors involved in the policy field, instead of focusing on the policy instrument itself, by evaluating policy implementation. The characteristics of the actors have a major influence on policy implementation. In addition, the implementation of policy (instruments) should not be separated from the context in which they are used, since characteristics of the environment also influence the actors involved in these implementation processes. Therefore, the Contextual Interaction Theory focuses on the involved actors in the policy implementation process and the interactions between them (Bressers et al. 2000). The implementation of policy involves three important components: inputs, interaction processes and outputs (see Fig. 2). The first component, inputs, is comprised of the (formal) rules of the game and resources which are considered required for a successful implementation of the policy. However, this input component is also comprised of contextual factors such as structures, positions and processes which already exist in the environment in which the policy or policy instrument is implemented. Thereby it is considered that any policy is never implemented in a blank policy field, but the new policy (instrument) will add an additional element to this field (Hoppe 2009). All these factors influence the actors and the interaction processes between them and therefore also the result of the implementation process. The second component, the interaction process, implies a conversion process as a result of the interaction of various involved actors during the policy implementation. The third component, outputs, indicates the result of the interaction process (Bressers 2007).
The arena indicates the 'place' were these interaction processes and decision making between the involved actors take place. Aside from the actors, this interaction arena (which resembles Ostrom's action arena) embodies the rules of the game, and the issues at stake at a given time and place (de Boer and Bressers 2011;Ostrom 2009). This arena, including its actors, issues and rules of the game, can either be set up and bound explicitly by agreement, or the characteristics of this arena are in a perpetual flux. The characteristics of the involved actors and the interaction processes between those actors have according to Bressers et al. (2000) a major influence on policy implementation processes. In CIT the influence of such factors is exercised via three key actor characteristics: cognition, motivation and resources, as shown in Fig. 3.
The CIT can be used to analyze and evaluate policy implementation processes by determining whether and to what extent the characteristics of the involved actors influence the implementation process and where and why these actors are influenced by the external context (Bressers et al. 2013). This is done by means of an analytical model. In addition, the simplicity of this model offers opportunities to make practical recommendations to improve the implementation processes and the policy instrument itself. The present study follows the development line of CIT. In a sense that it allows to explore various inputs, including contextual factors, actors and their characteristics, in the implementation of local performance agreements for housing policy which might hamper the implementation process and the achievement of the objectives pursued in the light of the Revised Housing Act 2015.

Research design
The present study seeks to explore how performance agreements are implemented as instrument to support societal legitimacy in local social housing policy making. Researching implementation of performance agreements is new; there is no previous research that studied performance agreements in relation to impact on societal legitimate policy making. At the same time the present study is to some extent also of an empirical-evaluative nature, analysing the implementation process of performance agreements and its outcomes and effects. To address this exploratory yet evaluative nature of the present study a case study research design will be used.
Case study research allows the exploration and understanding of complex phenomena within the boundaries of a specific environment, situation or organisation (Yin 2003). This includes research into comprehensive phenomena like studying the implementation process of local performance agreements in the social housing sector. To address this exploratory yet evaluative research nature of qualitative research methods were used to collect and analyse informative, rich data. This involved in-depth expert interviews with actors involved in the establishment of local performance agreements for (social) housing policy. This research approach captures experiences, meaning, perceptions and understanding of the challenges associated with the implementation of local performance agreements for (social) housing policy in the light of the Revised Housing Act 2015. This made it possible to analyse and reflect on the current state of affairs of performance agreements implementation processes, and the outcomes and effects that result from them (also when compared against the original goals they were designed for to achieve).
In the present study societal legitimacy is operationalized in terms of how the process and outcome of local social housing policy was perceived by the three key stakeholders (i.e. housing association, municipality, but in particular tenants) as 'just' in terms of openness and procedural fairness (Baxter 2017).

Case selection
To avoid getting a biased view it was decided not to use a single case study research design but to alternatively use an embedded case-study design that included multiple local case studies, with the case study an aggregated level pertaining to the Netherlands, and the embedded case studies pertaining to three municipalities where performance agreements are implemented. For selecting the three local case studies a deviating case selection criterion was used (Gerring 2007). This included variation in terms of city size, because it was expected that city size positively corresponds to urban complexity, which is expected to affect local social housing policy-making processes in a negative way, although larger municipalities generally are more experienced co-developing local housing policy and establishing of performance agreements (Severijn 2013). In addition, the criterion to which extent municipalities are experienced using performance agreements was used. This was important because at the time the research was conducted performance agreements had only been implemented in a limited number of municipalities with the aim for policy makers to learn from experiences. As such, the local case studies can arguably be perceived as municipalities and housing associations relatively early adopting performance agreements as a new instrument.
Using these criteria eventually led to selecting three cases: the municipalities of Rotterdam, Bodegraven-Reeuwijk and Zoetermeer. All three are located in the province of South-Holland. By selecting three cases varying in size, it is assumed that a proper picture of the implementation process of performance agreements will be provided. The first local case study, the City of Rotterdam concerns a highly urbanized municipality with over 640,000 inhabitants, and the largest city in the province, having experience with performance agreements. The second case, the municipality of Bodegraven-Reeuwijk, containing of the villages of Bodegraven and Reeuwijk, is a small municipality with a very low degree of urban complexity, that is home to 34,000 inhabitants. Since 2011, when the municipality was established after merger of the two previously mentioned smaller municipalities, it has not been involved in developing local housing policy. The third case is the municipality of Zoetermeer. Zoetermeer is home to 125,000 inhabitants, which makes it the third largest population centre in the Province of South Holland after Rotterdam and The Hague. In terms of urban complexity and experience with performance agreements Zoetermeer is in between Rotterdam and Bodegraven-Reeuwijk.

Data collection
Data collection took place between April and June 2018. The interview approach was the same for all interviews. Interviews were conducted with experts in the field of the Revised Housing Act and performance agreements in the social housing sector. This was done to provide a broad picture of the context and the implementation of local performance agreement in light of the revised Housing Act. In addition, interviews were conducted with actors involved in the decision-making processes regarding the establishment of local performance agreements in the three case studies. From the municipal perspective three experts were interviewed, from the perspective of the housing associations six interviews were conducted, and from the tenants' perspective five interviews were conducted. One additional interview was conducted with a law firm to reflect on the legal aspects of performance agreements. Each interview lasted approximately one hour.
Interviews had an open character and were semi-structured. Themes discussed in this interview pertained to key concept deriving from CIT (e.g. on interactions between actors in decision-making, the wider governance structure), but also to issues that address the essence of performance agreements, like the legal basis, differences with previous formats of establishing performance agreements, the changing relationship between municipalities and housing associations, local housing policy as basis for performance agreements, the role of tenants' organisation in the process, and the negotiation process for the establishment of performance agreements itself. Other themes included compliance with performance agreements, support policy offered by central government to facilitate and stimulate negotiation processes, central government priorities for housing policy, time limits for the establishment of performance agreements, and financial information of the housing association. Finally, there was also attention to the dispute settlement body, and influence other aspects of the revised Housing Act have on negotiation processes for the establishment local performance agreements.

Data treatment and management
Right after the interviews data collected were processed in detail by means of the recorded transcriptions and writings. All information gathered via the respondents was thereby grouped by the themes mentioned above . After the data were processed, interviews were grouped by case and by type of respondent when for example multiple housing associations have been interviewed for one specific case. This grouping simplified the analysis of the data, since it provided a clear overview how the different respondents evaluate performance agreements: It enabled easy inter-case as well as inter-stakeholder comparison. Some respondents provided additional information (documents) regarding the proceed of the interaction processes for the establishment of performance agreements for local housing policy.

Data analysis
For this study thematic analysis was employed using a qualitative data coding approach (resembling method of constant comparison (Strauss 1990), using a combination of a deductive and an inductive reasoning. Themes were identified using CIT (Bressers 2007(Bressers , 2009) as a framework (i.e. in a deductive manner) by referring to concepts as actor characteristics, problem context, political context, previous interactions, and governance structure. This also included a stakeholder analysis (Varvasovszky and Brugha 2000) of the three main stakeholders involved in the implementation of performance agreements in the Dutch social housing sector, with the aim to identify their main goals, motivations (access to key) resources, power, and problems with regard to the implementation of performance agreements. Data were subsequently analysed manually through reading the transcriptions until a general understanding of the content was derived. From the thematic analysis, barriers could be identified, that hamper successful implementation and functioning of the policy instrument of local performance agreements for (social) housing policy. Data analysis also included cross-case analysis comparing the three local case studies with the aim to identify commonalities, differences, and more specifically patterns (Gerring 2007). In comparing the three cases the conceptual structure of CIT was used: i.e. (1) Structural context (subdivided into local housing policy and perceived influence of the Revised Housing Act); (2) Case specific context (subdivided into financial position-as key resource characteristic-and task in relation to social housing); and (3) Interaction process (subdivided into motivation in the process, actors in the process, network relations, responsibilities in the process, and outcome of interaction).
Whereas geographical generalization of the results of cross-case analysis pertaining to three case studies is not possible the results are useful in terms of theoretical generalization.

Results
In this section first the results of the stakeholders analysis are presented (Sect. 5.1) followed by the results of the case study analysis, focusing on the practice of performance agreements in local policy making processes (Sect. 5.2), issues in the wider governance structure (Sect. 5.3), and issues with the organisational structure of the performance agreements networks (Sect. 5.4).

Stakeholder analysis
The three main stakeholders are housing associations, municipalities, and tenants' organisations. Housing associations own and let dwellings to tenants. Due to the introduction of the Revised Housing Act, housing associations are forced to emphasise on their core task: the provision of affordable housing for low-income households. Tenants of the regulated dwellings are directly influenced by the (policy) choices of the housing associations and the results of the performance agreements regarding social housing. Measures such as rent moderation, improving the energetic quality of the housing stock and adjustment of the number of regulated dwellings in a municipality directly affect the tenants. Tenants are represented by tenant's associations. Since the revised Housing Act came into force in 2015 the involvement of tenants in the process of local performance agreements has obtained a legal status. By involving tenants' organisations more actively in this process central government and into management of housing associations and the formulation of local housing policies it is assumed that social legitimacy increases (Ministry of the Interior and Kingdom Relations et al. 2015;Terlingen 2016). The responsibilities of the municipality are: linking of the investment capacity of housing associations with the local housing objectives by means of releasing funds by these organisations for social benefits. The Revised Housing Act prescribes to a larger extent than its precursor (the "BBSH" scheme) how municipalities and housing associations should cooperate. An overview of the characteristics of the three main actors is presented in the Appendix.

Performance agreements in local policy making processes
In Bodegraven-Reeuwijk, Zoetermeer and Rotterdam the local authorities developed different approaches to arrive at a set of local performance agreements. Table 2 presents an overview of the comparative case analysis. In Table 3 information is provided on the strengths and weaknesses regarding interaction processes that occurred in the three cases.
More detailed background information about the case studies can be found in the study by (Plettenburg 2018).

Issues in the wider the governance structure
Like other policies performance agreements are not implemented in a policy void. Therefore, it is important to analyse what kind of structures, processes and positions already exist in the policy environment (Hoppe 2009). The Revised Housing Act, that contains performance agreements, aims to further regulate the activities of housing associations by Table 3 Strengths and weaknesses of the interaction processes in the formulation of the performance agreements Interaction processes for establishment of performance agreements

Observed strengths Observed weaknesses
Case Bodegraven-Reeuwijk Strong involvement and close coordination of a wide variety of actors The formulation of a practical and concise housing policy document containing an action program from which the performance agreements can, almost directly, be derived Constructive relationships based on mutual trust between the municipality, the housing associations and the tenants' organisations Proper division of agreements made on the official level and the board level Still exploring the desired role and responsibilities for the tenants' organisation in the process Not taking full advantage of possible support instruments such as insight in the housing associations' financial situation

Case Zoetermeer
Constructive relationship between the municipality, the housing associations and the tenants' organisations Municipal housing policy priorities are internally coordinated with different departments The municipality is experienced with involving tenants in the implementation of local housing policy Close coordination between individual tenants' organisations Political disagreement in the municipal government coalition influences the establishment of performance agreements Ambiguity regarding responsibilities of actors involved in the process of formulating the performance agreement (which determines annual priorities for housing policy; absence of designated contact persons) The local housing policy is not reflecting a clear municipal vision regarding social housing Case Rotterdam Focus on the creation of performance agreements that are conceived in a way that is specific, measurable, achievable, relevant and time-bound, and that in the end facilitate proper monitoring and compliance of the performance agreements Weak involvement of actors in the processes regarding formulation of local housing policy One-sided establishment of annual priorities for social housing policy Individual sets of performance agreement are delivered instead of one mutual set of agreements Strained relations between the municipality and the housing associations It remains unclear what the desired roles and responsibilities for the tenants' organisation are in the negotiation process providing stricter rules and boundaries (Rijksoverheid 2015). During the formulation process several problems occur. First, there is a mismatch between the components of Housing Act and the desired cocreation and implementation of local housing policy. The Housing Act 2015 provides rules for the scope of action for housing associations. It prescribes that housing associations should focus on their core task: the provision of housing for low income households. However, strict regulations regarding the remit of housing associations could hamper the creation of tailor-made mutual solutions for local housing challenges in the negotiation process between housing associations, a municipality and tenants' organisations. In municipalities where market players (like commercial project developers, but also housing associations) do not take up the responsibility of building important facilities, it benefits municipalities when housing associations take up a broad remit, and when they are not restricted in their investment capacity via tax measures like the so-called 'landlord levies'. Due to the change in regulations change it is unclear to those involved in formulating local housing policies whether housing associations are allowed to do so.
A second issue concerns the lack of requirements set by the central government to the form and content of the mutual agreements. Although central government seeks to facilitate and ensure the co-creation of (social) housing policy and the establishment of local performance agreements for social housing policy, there are hardly any requirements that regulate the form and content of those agreements. This lack of (formal) rules is to some extent considered as 'freedom' to those involved in the negotiations because it enables the creation of a tailored local performance agreements. However, it also perceived as endangering achievement of goals once they have been set.

Issues with the organisational structure of the performance agreements networks
A second set of barriers concerning the implementation of performance agreements pertains to the organisational structure of the performance agreements established. Introduction of performance agreements to the local tri-partite actor network is considered troublesome. One of the goals performance agreements have in the social housing sector is to establish or strengthen cooperative, but non-permissive, actor networks. This particularly focusses on strengthening of ties between municipalities, housing associations and tenants, with the final aim of having housing associations making funds available that benefit social goals. In this view housing associations, municipalities and tenants' organisations should legally be considered as equal partners. The organisational set-up and associated deliberative negotiation processes is to a large extent left to the local actors involved. National government tries to ensure and facilitate the process of shared decision-making by means of supportive (policy) instruments. Fixed deadlines, a dispute settlement body and the provision of financial information of the housing associations are made available to facilitate and ensure the establishment of local performance agreements and commitment of housing associations. However, the present study revealed that issues may arise due to the current inter-organisational set-up. The first of these problematic issues pertains to the role of tenants' organisations in the process. Difficulties are faced regarding the role of tenants' organisations in the process of formulating local performance agreements. According to the Revised Housing Act tenants' organisations should preferably be full members in these decision-making processes. This implies that they should be able to assess the investing capacity of housing associations and thereby have sound knowledge of strategic housing management. However, involvement in this decision-making process is new to many tenants' organisations. This is also related to the fact that board of tenants' organisations are formed by and fully depend on volunteers (Terlingen 2016). These new responsibilities require additional knowledge, time and skills. The lack of knowledge about strategic housing management, financial knowhow, and limited negotiation capacity are perceived as problematic. This is illustrated by tenants' representative in the Zoetermeer case, "We are volunteers and both the municipalities and housing associations are professionals, which have access to the required expertise and time to deliver their responsibilities. The tenants' organisations can fulfil its role better when the municipality and the housing association take into account that they are a voluntary body and adjust accordingly". This was also observed in the Rotterdam case, "The network comprised of the municipality, housing associations and tenants' organisations should, theoretically, results in a proper triangular relationship. However, this is not the case. During the negotiations it is mainly about detailed figures concerning real estate development; these are abstract topics for tenants". In sum, currently tenants' organisations are not capable of fulfilling their desired role, putting achievement of one of the goals of performance agreements at risk. That is enhancing the role of tenants' organisations in social housing to ensure the social legitimacy of housing associations.
The second issue pertains to the local housing policy as the basis of the local performance agreements. According to the revised Housing Act the performance agreement(s) should be based on the local housing policy document because housing associations are required to reasonably contribute to this document. However, in practice this leads to problems in the tri-partite actor configuration. The local housing policy is often not suitable as the basis of the performance agreements for housing policy, since it offers a much broader political view regarding housing, and is not only focussing on social housing policy alone. Therefore, it often lacks direction and is therefore not considered a proper basis for the local performance agreements. Since this document does often not clearly reflect the social housing objectives, it is hard for housing associations to establish a proposal in which they convey on how to could contribute to meeting these.
A third barrier associated to the organisational structure of the tri-partite actor configuration is the operational area of housing associations. This is often not restricted to the boundaries of one jurisdiction. Housing associations that own property across multiple municipalities have to decide on how to allocate their investments. This often means that they have to prioritise investing in housing stock in one municipality over another (Severijn 2013). As such this also influences contribution to local housing policies. This problem was observed in the case of Zoetermeer where housing association Vestia (the largest of its kind in the Netherlands) was unable to guarantee making large investments because it was already committed to making investment in other municipalities in which it also owns housing property. Moreover, in the Rotterdam a representative from a housing association stated that, "The present local housing policy emphasising on so-called non-SGEI (Service of General Economic Interest) activities, which has hampered the process of establishing performance agreements." Fourth, there are ambiguities regarding actor responsibilities. Local performance agreements do not precisely prescribe how the collaboration between housing associations, municipalities and tenants' organisations should be arranged. This is left to those involved locally. In practice this led to problems and ambiguities regarding roles and responsibilities of the local actors. This was primarily related to facilitation shared decision-making processes. The Rotterdam case shows that one of the three parties-i.e. the municipality-tried to take a leading role, but did this in a non-facilitative, rather problematic way. A housing association representative states, "According to the law, the municipality is not completely at the helm and the instrument is meant to create partnerships. However, here the municipality adopted a directive role within the process of performance agreements". And a tenants' organisation representative in Rotterdam argues to regret the directive role the municipality of Rotterdam has taken in the process of the performance agreements, and laments the (right wing) government coalition, "In general the housing associations have better insights and know-how regarding the needs of the tenants. However, the position of the municipality in this process is to a large extent dependent on the type of coalition. Another coalition with left-wing (political) parties, had probably led to a completely different process". To ensure proper collaboration and coordination between the three main actors there is a need for a specific actor who facilitates the local housing policy making process. Therefore, it is important that it has been assured that the collaboration between actors is facilitated, and whether the right parties are involved with the right mandate to act, while having the required knowledge level to participate in decision making.
Finally, there is an issue with the strained relationships between municipalities and housing associations. Proper collaboration between the three main actors cannot readily be taken for granted. Although there are shared responsibilities between municipalities and housing associations, the two are very different organisations, each with their own responsibilities and (conflicting) interests (Van Kessel et al. 2017). Strained relationships-often based on conflicts previously encountered-might seriously affect the negotiation-processes and therefore hamper successful formulation and implementation of the performance agreement. A cooperative relationship between the network actors is considered essential to contribute to meeting the objectives pursued by performance agreements. This is illustrated by a quote from a municipality civil servant in the Bodegraven-Reeuwijk case: "The municipality cannot simply demand things from the housing associations. They remain autonomous parties and we are dependent on each other: the municipality needs the housing associations to achieve the social housing goals and the housing associations, on the other hand, need the municipality for land." And, "It is better when more attention is paid to the relationships between the parties and a constructive way of establishing performance agreements with each other." In the Zoetermeer case missing a strong foundation for collaboration was problematized, "The relations between the different parties in the coalition are not running smoothly; it lacks a strong foundation". Moreover, "The establishment of local performance agreements for housing policy should be a team achievement, where co-creation is key".
All five issues can potentially complicate the tri-partite negotiation process. The most important condition, observed in the case study analysis concerns a cooperative relationship between the three main actors. Decision-making process could benefit from a number of conditions that are present. In the Bodegraven-Reeuwijk case, the actors involved put a lot of effort in finding common goals and agreements. Succeeding to establish these was rooted in mutual trust that contributed to a smooth negotiation process. By contrast the Rotterdam case showed a troublesome negotiation process, where the municipality had assumed a leading role in the process of establishing performance agreements but did not put much effort in finding shared interests nor in investing in sound relations with the housing associations and tenants' organisations. This led to resistance, particularly from the housing associations involved in the negotiations.
When comparing the three cases it looks like the process of formulating performance agreements runs more smoothly in small-sized municipalities like Bodegraven-Reeuwijk than in a large, densely populated ones (i.e. Rotterdam). Differences that come to the fore pertain to actor involvement, and alignment of problem perspectives, goals and ambitions. Particularly in the last dimension smaller municipalities seem to offer conditions that benefit the establishment of performance agreements. When looking into strategies, responsibilities and resources used there are only a few differences between smaller and larger municipalities. These items are moderately present in all cases.

Discussion
This study contributes to the academic body of literature regarding the functioning of performance agreements or performance measurement systems in the public sector. The way performance agreements are introduced in the Dutch social housing sector following the Revised Housing Act 2015 is a bit unusual, though. Therefore it is fairly difficult to reflect on the list of propositions presented in Table 1 (Sect. 2.1). First, in the present case performance the focus is on a set of local actors who negotiate about formulating local performance agreements. It is not about actual implementation (yet) in contrary to most studies in the academic literature (De Boer et al. 2015;James 2004;Jongbloed et al. 2019;Young 2005). So, the focus is on throughput and to some extent output, but not on outcome performance. Moreover, there is special attention to (societal) legitimacy and empowering the role of tenants' organisations and municipalities in this process in particular, and social housing policy at large. Second, whereas most performance agreement systems are based on rewarding (De Boer et al. 2015) this element appears to be missing out in the present case where monitoring and enforcement is left to local actors, and is basically hardly regulated nor incentivized. From policy effectiveness and legitimacy perspectives this is potentially troublesome because nothing has been arranged in case eventual performance by housing associations is low. Moreover, no financial penalty can be inflicted in case of insufficient performance. Third, whereas performance agreements are expected to increase in perceived accountability, transparency and credibility (Jongbloed et al. 2019) the present study showed that this does not hold in all cases. At least in the Rotterdam case this claim could not be confirmed. Fourth, whereas the present study revealed some cases of assumed increase in administrative costs following the introduction to the Revised Housing Act it is difficult to assign this to performance agreements alone. It is rather related to introduction of a wider set of actions that result from the Revised Housing Act.
In 2019, the Dutch government evaluated the Revised Housing Act (Ministry of the Interior and Kingdom Relations 2019). With regard to performance agreements, it was concluded that municipalities and tenants' organisations have more influence over the activities of housing associations, thereby increasing the legitimacy of housing associations in their area of activity. However, a significant proportion of housing associations believe that they are being over-questioned by municipalities in terms of their contribution to the local housing situation, and that the municipal housing visions are regularly not concrete enough to make agreements about, or are not appropriate to the local housing situation. Many parties also indicate that they consider the annual cycle of performance agreements as too intensive. There is a call for more commitment to multi-annual agreements. Specifically with regard to the position of tenants' organisations, the evaluation shows that the majority of all parties involved view the added value of the involvement of tenants in making performance agreements. While tenants' organisations are generally satisfied with the influence they have, the evaluation shows that it is important that tenant organisations are further supported in the practical implementation of their role, for example by support from professionals, knowledge and training of tenants active in the tenant organisation, and finding additional ways to engage their constituencies (Ministry of the Interior and Kingdom Relations 2019).
One of the conclusions of the evaluation was that although a stronger position of municipalities and tenants has been achieved in recent years, the collaboration between the three parties can be developed further. One of the policy intentions arising from the evaluation is therefore that a process should be started to further develop collaboration between municipalities, housing associations and tenants' organisations, for example with support for making good 'cooperation agreements' prior to making actual performance agreements (Ibid.). Although the national government's evaluation of the Revised Housing Act is much broader and more general than the case studies presented in this paper, many of the findings are consistent. An important difference is that where the national government evaluation points to the need to invest in making proper collaboration and sound process agreements, the present study specifically highlights the need to invest in good relations between the parties involved.
In line with Lewis (2015) we argue that performance measurement systems in the public sector can be considered as a social structure between individuals and institutions. The present study illuminated the importance of focusing on the social interaction processes between municipalities, housing associations and tenants in order to derive at local performance agreements as a compromise. The Contextual Interaction Theory (Bressers 2007(Bressers , 2009), perceiving policy implementation as a set of interactions between actors involved, demonstrated to be appropriate to analyse and assess the social dynamics and its influence on the shaping of performance agreements for local housing policy.
Like performance agreement systems in other sectors the system in the present study suffers from (information and capacity) asymmetry (Young 2005). The current system favours housing associations, because they have the capacity, professional knowledge and management systems on board, whereas municipalities-especially the smaller ones-and to a larger extent tenants' organisations do not. The latter two organisation have a big disadvantage going into the performance agreement formulation negotiations. In line with the claim by Jongbloed et al. (2019) that implementation of performance agreements depends a great deal on situational context, the present study shows that introduction of local performance agreements in the social housing sector runs more smoothly in small-sized municipalities than in large-sized complex urban cities. Additionally, the study provided insight in how the governance of social housing works in The Netherlands. Housing associations in the Netherlands are governed in these performance measurement systems and how attempts have been made to ensure the legitimacy of these organisations (Cornforth et al. 2015;Ebrahim et al. 2014;Raeymaeckers et al. 2017). Alternatively, these insights could also be applied by stimulating research addressing how performance agreements or other types of performance measurement tools influence other types of hybrid organisations, preferably also from a legitimacy perspective.
The present study revealed a few new issues that are relevant to studying performance agreements in (semi-)public service delivery. They are legitimacy in (strained) relationships between parties that have to negotiate the formulation of performance agreements, distrust between these actors, lack of knowledge and capacities in these processes, lack of prioritysetting, having a short term focus instead of focusing on strategic, longer term goals, response to the introduction of performance agreements as a new policy instrument by some of the negotiating actors (including struggles and ambiguity to master new responsibility that come with the introduction), and refraining from using possible support to the full extent to get more information from other negotiation parties (i.e. housing associations).

Conclusion
The present study analyzed the introduction of performance agreements in the Dutch social housing sector following the revision of the Housing Act in 2015. The main research question was: How are local performance agreements implemented targeting increased societal legitimacy in local social housing policy making, and what are its strengths and weaknesses in three selected cases in the Netherlands? To answer this question an embedded case study analysis was performed, including a document study, expert interviews and a comparative analysis of three local housing case studies.
Whereas performance agreements usually target direct output and outcome performance of implementation processes, the performance agreements analysed in the present study also target multi-stakeholder co-creation, throughput and output performance. Results reveal two types of weaknesses: those related to the governance regime and context, and those related to inter-organisational structure and stakeholder interaction. The former pertains to a mismatch between other components of the Revised Housing Act and the desired co-creation and implementation of local housing policy, and a lack of requirements set to form and content of agreements. The latter pertains to the role of the tenants' organisations in the process, the local housing policy as the basis of the local performance agreements, allocation of resources by housing associations that own properties in multiple jurisdictions, ambiguities regarding actor responsibilities, and strained relationships between municipalities and housing associations.
All the issues mentioned were found to directly influence decision-making process regarding the establishment of performance agreements for (social) housing policy, and hamper sound implementation of performance agreements as a policy instrument. If issues are not properly addressed performance agreements will fall short of meeting their intended targets, i.e. fostering multi-actor co-creation and gaining increased legitimacy of social housing policy formulation and implementation (via increased levels of stakeholder commitment and compliance).
The most important condition for proper implementation of the policy instrument of performance agreements and the co-creation of local (social) housing policy is a cooperative relationship between the tri-partite network stakeholders. Therefore, municipalities, housing associations and tenants' organisations will have to put efforts in investing in good relationships and creating partnerships. Thereby shared goals should be identified. However, simultaneously the differences in interests and responsibilities should be taken into account. Central government could try to find ways to accommodate this process, offering support policy that goes along with the implementation of performance agreements. The case studies analysed also shows that implementation of performance agreements is more difficult in large-scale cities in which local administration has to deal with the issues and complexity of dense urban areas.
Limitations to the study pertain to geographical scope and the number of case studies. Whereas studying three cases is useful in exploratory and descriptive research into introduction of performance agreements in the social housing sector, more structured research is required to assess goal achievement and effectiveness of this policy instrument. It is recommended to conduct additional research in finding suitable approaches on how tri-partite collaboration between housing associations, municipalities and tenants' organisations can be established and facilitated, and how issues that hampered implementation of performance agreements in the social housing sector can be overcome. For example, by assessing which policies or interventions are potentially useful to support the tri-partite decision-making process. Future research may be undertaken into issues concerning the role of the tenants' organisations in the process, especially from a legitimacy perspective. Finally, research analysing the implementation of performance agreements and other policy instruments that follow from the New Public Management paradigm in the social housing sectors would be recommended. This also holds for studies in other countries than the Netherlands. Lack of organisational and professional capacity Lack of (professional, technical, legal) knowledge Poor negotiation capacity Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.