Exploring the durability of community enterprises A qualitative comparative analysis

Despite the growing attention given to community-based initiatives in the delivery of public services, little is known about their durability. This article focuses on Dutch community enterprises (CEs) as an emerging form of community-based initiatives in urban regeneration. Due to their self-organizing and (largely) voluntary character, CEs face considerable governance challenges to sustain activities over time. Based on the literature, we examine the interplay of four key conditions in relation to their durability: social capital, entrepreneurial community leadership, supportive relationships with institutional key players and a strong business model. Based on an fsQCA of 12 cases, the main conclusion is that the presence of social capital, strong entrepreneurial leadership and a strong business model is the most important configuration leading to a durable CE. Government support is not found to be a necessary or sufficient condition, but its absence is part of the explanation of non-durability.

• aiming to deliver (long-term) social benefits to local people, that is, defined populations or subgroups, by providing specific goods or services; • seeking to generate a surplus through (at least in part) engaging in trade in the marketplace or other economic activity, and reinvesting the surplus in the business or community; • locally accountable and strongly committed to involving local people and other partners in their activities, through participatory and/or democratic decision-making processes.
Many scholars perceive hybridity as the defining characteristic of CEs (and of SEs in general), because of the pursuit of a dual mission: financial sustainability and a social purpose. By spanning the boundaries of the private, public and non-profit sectors, CEs bridge institutional fields and face conflicting institutional logics (e.g., Doherty et al. 2014). Generating own income (through a business model) and simultaneously adding social value are crucial for CEs. This hybridity may cause difficult trade-offs. On the one hand, financial considerations may 'squeeze out' social goals, which is also a risk for the legitimacy of CEs. On the other hand, mainly emphasizing social goals endangers financial aspects of the business and thus threatens its survival. Various studies suggest that CEs face considerable governance challenges to keep the enterprise running and effective over time, that is, delivering particular services and social benefits to the target community (Haugh 2007;Spear et al. 2009;Bailey 2012). Key governance challenges are management, in particular recruiting CE board members with the right skills and experience, involving local people and other partners through participatory decision-making processes, and balancing the dual mission of social benefits and financial sustainability (Haugh 2007;Spear et al. 2009;Bailey 2012). Achieving sufficient staff capacity is a recurring challenge for CEs, which commonly depend on volunteers and few (if any) paid staff members.
Lastly, CEs' entrepreneurial role in public service provision can be a challenge to government-led service provision, which might result in struggles to legitimize their form of service delivery and to get support (Gofen 2015). Moreover, the questions of how representative the organization is for the wider community and to what extent CEs are able to deliver regeneration benefits pose challenges in requiring resources (e.g., community assets, funding) or acceptance from public bodies.
Hence, the hybridity of CEs poses particular governance challenges for their durability. Generally, durability refers to something continuing without failing. We define a durable community enterprise as a CE which has sufficient internal capacity to keep its business running and to realize its objectives, while simultaneously meeting community needs and having external legitimacy. Hence, durability is a broader concept than performance, a common indicator of measuring how social enterprises fare (Bagnoli and Megali 2011). Starting from our definition of CEs, we argue that durability goes beyond performance, because it not only looks at goal realization (social benefits), but also pays attention to the internal capacity (financial resources, membership, staff capacity) of CEs and to their legitimacy (public recognition, acceptance and meeting community needs). In order to assess durability, we focus on three components: (perceived) goal achievement, structural internal capacity, and external legitimacy (see also

| CONDITIONS FOR DURABLE COMMUNITY ENTERPRISES
This article aims to provide a better understanding of the combined impact of various conditions upon the durability of CEs. We need to move beyond a focus on 'stand-alone' conditions, towards an understanding of the interplay between various conditions. In other words, this article will analyse configurations of conditions, which are highly context-dependent. From the cross-disciplinary literature on civic and community capacity building, communitybased (social) enterprises and entrepreneurship, we detected four different, but often typified as crucial, conditions which are likely to affect the durability of CEs: strong social capital, strong entrepreneurial community leadership, supportive relationships with institutional key players and a strong business model. Each of these conditions and their interplay will be further discussed in the following subsections.
3.1 | (Access to) strong social capital (C1) Social capital generally refers to resources that are accessible through social interactions and networks, reciprocity, norms and mutual trust (Bourdieu 1986;Coleman 1990;Portes 2000;Putnam 2000). In particular, Putnam's seminal book Bowling Alone (2000) moved the spotlight from individual to community benefits from social capital (see also Portes 2000).
As social networks and connections are the main 'production mechanism' of social capital, they play a crucial role in the functioning and durability of CEs. Social capital facilitates the mobilization of resources (e.g., knowledge, information and experience), coordination of actions, and creation of safety nets that reduce risks for individual community members (Peredo and Chrisman 2006;Bailey 2012).
The literature generally distinguishes three types: bonding, bridging and linking social capital (Putnam 2000;Woolcock 2001). In the context of CEs, bonding social capital is derived from strong ties between neighbours, friends and association members. These ties help the formation of CEs and dynamics in board management and continuity (Somerville and McElwee 2011;Bailey 2012). Bridging capital refers to the CE's ties in the wider neighbourhood, ties to other networks and other community organizations. Bridging capital can generate support and a base for recruiting volunteers, potentially impacting upon the structural capacity and legitimacy of the CE (Somerville and McElwee 2011;Bailey 2012;Healey 2015;Kleinhans 2017). Linking capital refers to ties of CEs with institutional key players, which can be crucial for getting resources and support. Because the literature attaches strong importance to the attitude and (un)supportive behaviour of (external) institutional key players as counterparts of the CE, we will elaborate the latter aspect as a separate condition (see section 3.3).
Several authors conceive social capital as a necessary but not sufficient condition for CEs to sustain their activities, because in itself, social capital cannot deliver economic capital and institutional support (see Saegert 2006;Dale and Newman 2010). In building a durable CE with sufficient organizational and institutional resources, social capital has to be harnessed and transformed into specific organizational and/or institutional resources, assets and support (Chaskin 2001). Strong entrepreneurial community leadership can play a key role in this.

| Strong entrepreneurial community leadership (C2)
Community leadership is considered a key condition in the mobilization of resources and development of organizational capacities. The literature extensively describes activities and characteristics of community leaders (e.g., Selsky and Smith 1994;Maton and Salem 1995;Purdue 2001). As a condition for (durable) CEs, we distinguish three sets of leadership activities (cf. Selsky and Smith 1994): (1) community building and mobilizing, (2) building collaborative and strategic alliances between CEs and institutional power holders and (3) identifying and exploiting new entrepreneurial opportunities. These activities are not necessarily performed by one specific individual, but, more likely, by different people.
First, community leaders who can formulate a vision and inspire and motivate others (Maton and Salem 1995;Hartley and Allison 2000) are able to mobilize resources and commitment from residents (and/or volunteers) and organizations towards common goals. Second, community leaders can build strategic alliances with big players such as funding agencies, local authorities and private companies (Selsky and Smith 1994;Purdue 2001;Varady et al. 2015). Such alliances enable them to mobilize resources beyond the boundaries of their own organization.
Boundary-spanning activities and competences are particularly important in developing these alliances with institutional key players (Van Meerkerk et al. 2013;Edelenbos et al. 2018). The third characteristic of community leadership in relation to CEs concerns recognizing and exploiting entrepreneurial opportunities (Haugh 2007;Gofen 2015;Renko et al. 2015). Especially in a resource-poor environment, identifying and exploring opportunities for resource acquisition while simultaneously meeting community needs is of vital importance for CEs. Key characteristics in this regard are a risk-taking attitude, flexibility and creativity (Renko et al. 2015).

| Supportive relationships with institutional key players (C3)
Even though autonomy and self-organization are important CE characteristics (see section 2), CEs are strongly embedded in their institutional environments and dependent on the (active or passive) support of vested players (Kleinhans 2017;Edelenbos et al. 2018). In particular, CEs which interfere with institutionalized processes of service delivery or decision-making are dependent on how local government and/or professional service providers respond (Bailey 2012;Gofen 2015;Brandsen et al. 2017). Others may just rely on support for getting started or for gaining assets. Support can range from allowance, counselling and stimulation (e.g., start-up grants), to intensive collaboration and co-production, being more active forms of support (Seixas and Berkes 2009;Kleinhans and Van Ham 2017;Edelenbos et al. 2018). However, the literature is unclear on what level of support is necessary for CEs to become durable.
Although governments may consider the durability of CEs to be of strategic interest in terms of maintaining service delivery, the willingness and capability of local governments to actually provide support is not straightforward (Healey 2015). Governmental organizations are often not well-equipped to deal with new forms of co-production and collaboration with CEs (Voorberg et al. 2015;Kleinhans 2017). A boundary spanner or broker within the governmental organization often seems to be necessary to mobilize support and to coordinate and align government activities with CEs (Van Meerkerk et al. 2013), which affects how the relationship with local government evolves.

| Strong business model (C4)
CEs seek to generate a surplus (partly) through engaging in trade in the marketplace or other economic activities, and reinvesting the surplus in the community (see section 2). A business model which can generate a steady and secure revenue stream can enhance the autonomy of the CE and thus its durability. A business model may include several components (Bailey 2012;Varady et al. 2015;Kleinhans and Van Ham 2017): • The use of assets, that is, buildings, land or other resources from which a (preferably steady) revenue can be extracted (e.g., renting out spaces for business); • Services or facilities offered to generate revenues (for an overview, see Bailey 2012). These are not by definition services enacted to achieve the social ambitions of the CE, but there may be overlap in activities organized for simultaneous financial benefit (continuity of the CE) and community benefits; • The use of external funding-subsidies, start-up grants, donations or loans. Depending on the conditions, such funding may imply a certain level of dependency on the donor.
Many CEs tend to rely on more than one of these components. The strength of the business model is partly determined by the extent to which CEs make a surplus and are not too dependent on (short-term) external funding.
However, achieving break-even is a considerable challenge for many CEs (Pearce 2003, Bailey 2012Healy 2015).

| The interplay between the conditions
While the above discussion highlights the theoretical conditions for durable CEs individually, they are likely to be related. First of all, community leadership (C2) and strong social capital (C1) are related in several ways. Through providing a vision, developing common ground and motivation, bonding capital can be enhanced and fostered into strong organization and board ties by community leaders, while bridging capital can be used and enhanced for mobilizing volunteers and attracting visitors to CE activities. Furthermore, by their boundary-spanning and entrepreneurial activities, community leaders can accumulate linking social capital and transform this capital into specific resources or collaborative projects for building a strong business model (C4) (Chaskin 2001;Van Meerkerk et al. 2013). However, in this transformation, they are likely to be dependent on the willingness and supportive attitude of institutional key players (C3). In general, entrepreneurial community leadership (C2) seems to be important to develop a sound business model (C4) that enables CEs to move beyond break-even and start making a profit. Supportive relationships with key players (C3) can be indispensable for providing CEs with specific resources that can facilitate their activities and business model, such as specific assets from which rent income can be generated. Hence, we expect that the presence of all four conditions, in conjunction, is likely to lead to a durable CE. The next section moves to our approach to examining durability and the presence or absence of different conditions in the context of CEs in the Netherlands.

| Data collection and case studies
This article combines data from three qualitative studies on CEs, that is, a panel study with in-depth interviews during the years 2012-15 (Kleinhans 2017), and two follow-up studies (2016-17) which generated a dataset of 12 case studies. This approach enabled us to select nine case studies for which we have longitudinal data of two or three measurements since 2012. The exception are three cases that started more recently. We took a purposive sample including both perceived 'success cases' (in terms of durability), and cases which are perceived as less durable.
Appendix 1 shows the main characteristics of the 12 cases.
The main data sources are semi-structured interviews, conducted with CE initiators (usually active residents), board members, other involved volunteers/residents and professionals from local governments and housing associations, who were involved in the initiatives. The analysis is based on 70 interviews: 36 recent interviews (as part of the follow-up studies, including all cases) and analysis of 34 previous interviews (covering nine cases). Interviews lasted between 45 minutes and 2 hours and were recorded and subsequently transcribed. We coded the interviews based on the four conditions and the outcome (durability) (see next section on operationalization). Moreover, we conducted direct observations (for example at CE meetings) and a document analysis, in particular of CE annual reports which provided useful information on financial and organizational performance.

| Operationalization and methods
This article applies Fuzzy Set Qualitative Comparative Analysis (fsQCA) to study how different conditions interact and lead to (non-)durability of CEs. FsQCA is a case-based method aimed at iterations between theoretical ideas and empirical evidence (Ragin 2000). QCA is based on three assumptions about causality which make the method particularly suited to study CEs. As the empirical evidence on CEs is scarce we want to unravel which particular explanations (i.e., paths or recipes) may equally result in a (non-)durable CE (this is called equifinality). Moreover, we expect that the selected conditions only have an effect in combination with the other conditions (this is called conjunctural causation). Furthermore, the state of a durable CE might have different causal explanations (a different combination of conditions) from a state of non-durability (this is called asymmetric causality) (Schneider and Wagemann 2012, p. 78). As we are interested in the interplay of certain conditions and want to provide an in-depth analysis which unravels the causal explanations, QCA fits our purposes very well. Moreover, given our limited sample size (n = 12), statistics cannot provide the consistency and strength to deliver robust results.
FsQCA is a qualitative method which urges scholars to describe both the conditions and the outcome as 'sets', that is, clusters with a very specific characteristic (e.g., strong social capital). Based on the features of a particular case, the researcher assigns to the case a specific membership score in this set. The most basic type of QCA is crispset QCA, in which the researcher has to translate all the conditions and the outcome in binary terms (0 and 1). We use a more fine-grained scale, that is, four-value fuzzy sets to do this. This provides a more nuanced assessment compared to crisp-set QCA, and helps to understand the cases and causal explanations more thoroughly. To each condition, we assigned scores ranging from 1 to 0, which indicate: • 1 = if a case is a full member of a set, a high degree of a particular condition • .67 = if a case is more in than out of a set, a moderate degree of a particular condition • .33 = if a case is more out than in the set, partial (but not full) absence of a particular condition • 0 = fully out of the set, absence of a particular condition.

| Process of calibration
In fuzzy-set QCA, the most important step after collecting data is calibration, by which data are converted into the different (fuzzy-set) values. We followed the procedure elaborated and recommended by Basurto and Speer (2012), consisting of six steps. Based on the literature, we first operationalized the four different conditions for durability. We subsequently developed an interview guideline based on this operationalization (for the follow-up study). In the third step we conducted a content analysis of the raw data (interview transcriptions and documents) and coded the interviews based on the indicators for each condition for durability (see Table 2). We summarized the information from the interviews by selecting typical statements that best reflect the qualification for each indicator. In the following step the specific values for each condition were discussed among the three authors in order to assign fuzzy-set values. In the sixth and final step, we also discussed each case intensively within the research team and assigned scores. Consequently, some indicators were slightly adapted or new ones were included to be better able to score the cases on the conditions and to do more justice to the specific context of our cases (cf. Basurto and Speer 2012). Table 1 presents the different indicators on three components of durability on which the outcome score is built.

| Calibration of conditions and outcome
Based on the conceptualization of durable CEs in section 2, the three components are: • the level of social effectiveness or goal achievement, • the organizational and financial capacity (structural internal capacity), and • legitimacy and recognition. Has sufficient and competent staff capacity (board members, employees, volunteers) to continue to run the business, to perform the activities underlying the CEs social objectives is flexible in adapting to changing needs (of the target groups) and external changes;

Legitimacy and recognition
Delivers goods / services that meet widely felt needs of local people, which is evidenced by the target groups' perceived level of legitimization for the CE; is actively supported by and receives (public) recognition from other stakeholders Table 2 shows how each of the four key conditions as discussed in the theoretical framework is translated into multiple indicators, and by which membership scores could be assigned to the cases regarding each condition.
In fsQCA, explanations are expressed in terms of necessity and sufficiency. Basically, a condition is necessary if durability cannot be achieved without it. However, a necessary condition is not always sufficient on its own. A condition-or a combination of conditions-is sufficient if it typically results in the presence of durability (Ragin 2000; Schneider and Wagemann 2010).
There are two parameters to consider whether the (combination of ) condition(s) leads to the outcome: coverage and consistency. The coverage score refers to the percentage of the outcome that is covered by the causal recipe (i.e., the empirical relevance) (Schneider and Wagemann 2010). Consistency is the proportion of the sum of the membership scores, which the cases have in the particular outcome (Vis 2012, p. 187;Schneider and Wagemann 2012).
We first analyse the necessary conditions for durability and non-durability, as this is the standard of good practice in fsQCA (Schneider and Wagemann 2010). Afterwards, we construct 'truth tables' for durability and non-durability. A truth table synthetizes how many cases adhere to a certain pattern and if they consistently show the same outcome, that is, durable or not. If a particular pattern is consistent, the researcher takes these rows into the minimization procedure: 'If two configurations differ in only one condition, but show the same outcome, this particular condition can be eliminated' (Schneider and Wagemann 2012, p. 105). By working this way we end up with 'minimized' causal paths explaining durability and non-durability. In deciding on which pattern is consistent enough, the consistency cut-off was set at 0.87 for the analysis of sufficient conditions. This threshold ensures that there are no deviant cases of consistency in kind. We set the consistency threshold at 0.90 for necessary conditions. These cutoffs are considered adequate (Schneider and Wagemann 2012).

| ANALYSIS AND RESULTS
In Table 3 the calibrated scores of the four conditions and the outcome for each CE are combined. The analysis was carried out with the program RStudio and the QCA(GUI) package in R (Dusa 2007;Thomann and Wittwer 2017). We analysed the paths leading to (strong) durability, as well as the recipes leading to a low level of durability. 1 Below, the results for each of these two outcomes are explained in detail. Considering the limited number of cases, anonymity of our interview respondents is guaranteed by denoting the case studies with abbreviations.

| Durable cases
The analysis of necessary conditions 2 shows that strong social capital is a necessary condition for CE to realize durability (consistency: 1, RoN 3 : 0.64, coverage: 0.81). This means that in all cases which are durable, strong social capital is present. Apparently, social contacts and networks, both within the CE and beyond, must be in place in order to successfully build a durable CE. Theoretically, this makes sense, as social capital is, according to Bourdieu (1986), a fundamental condition which can be used as a vehicle to start up movements and which can be mobilized with leadership and good ideas. Although the analysis shows that strong social capital is a necessary condition, it is not sufficient by itself as will be shown in the next section.

| Analysis of sufficient conditions
The analysis of sufficient conditions shows which configurations (i.e., combinations of conditions) lead to a durable CE. The software creates a 'truth table' (Table 4) which clusters cases adhering to the same configuration. However, their outcomes can be different. This is how the software indicates contradictions. Cases that show the same Supersubset relations reveal all possible combination necessary conditions, we only consider individual necessary conditions for their theoretical relevance. trajectory but differ in terms of their outcome are clear contradictions, and make consistency parameters drop. 4 Furthermore, the truth table shows logical remainders, possible combinations of conditions which were not observed. In this analysis we restrict ourselves to the conservative solution as we have only 12 cases. So, we explore and summarize our data rather than explain a phenomenon. Hence, we do not make any assumption about logical remainders.
The analysis of the presence of durability shows one path: having strong social capital and strong entrepreneurial leadership in combination with a strong business model (consistency: 1 PRI 5 : 1 coverage: 0.82). Of the seven durable CEs (scoring 0.67 or 1 on durability), six adhere to this recipe. In these cases we observe a stable group of core members showing strong bonding capital, but who are also embedded in broader networks of associations. These CEs are developed by individuals with a history of participation and who are highly connected with other neighbourhood associations and institutional key players. As one respondent puts it (reflecting on the community leader of ZaS): 'By his former job as chair of the neighbourhood council, he had a good position to take initiative. … He is enormously visionary … and has a tremendous ambition and drive.' Throughout the years these community leaders have built big networks and they often have experience with politics. We observe a strong interplay between social capital, entrepreneurial community leadership and the development of a strong business model. In these cases, community leaders used the bridging capital of the CE to organize community events and to recruit more volunteers and clients for their CE, thereby enhancing staff capacity and realizing more impact. As one community leader notes: 'We find volunteers and participants through existing networks, such as neighbourhood groups …, but also via open evenings. And even professionals from the welfare organization refer people to us.' Moreover, community leaders sometimes use linking capital of the CE to develop the business model (e.g., developing and focusing ideas) and to build strategic alliances. Entrepreneurial community leadership played a further important role in linking a strong business model to the social capital of the CE in order to make the business model effective (through coordinating tasks and professionalization of roles and responsibilities of When having a look at the parsimonious solution (see appendix 2), which we have to interpret with care (as the analysis is based on only 12 cases), we see that a strong business model and social capital are core conditions in explaining the durability of our cases, and entrepreneurial community leadership is a secondary condition (Fiss 2011). Although this may suggest that community leadership is less important for durability, we argue, building on the case material, that community leadership is an important condition, but plays a more indirect role: enhancing social capital and transforming social capital and ideas into projects for realizing strong business models. As leaders do not always succeed in this, we find community leadership in different combinations with social capital or a strong business model in paths explaining either durability or non-durability (see also the next section).
Contrary to our expectations, a supportive relationship with institutional key players is not part of the sufficient configuration. While the majority of the durable cases (five out of seven) do have supportive relationships with institutional key players, two (HeB and HeN) cases do not. These CEs experience a passive and reluctant attitude from the local government, but there are no outright conflicts and both cases did receive some support in the start-up phase through access to a community building for developing their enterprise. In contrast to the other durable cases, a boundary-spanning official is absent in these two cases: a public servant who serves as a broker for the initiative, facilitating government support and enhancing alignment with local government activities.

| Non-durable cases
The absence of a strong business model is a necessary condition for non-durability (consistency: 0.93, RoN: 0.87, coverage: 0.81). Cases that are not durable do not have a fully-fledged business model or it is weakly developed. In these cases revenues, assets, services or external funding that generate income are absent or weakly developed. This confirms previous research on CEs that developing financial sustainability is very challenging (Bailey 2012). Again, after analysing necessary conditions, we create a 'truth table' (Table 5).
The analysis of sufficient conditions shows three paths that explain the non-durability of CEs (see paths 2-4, Table 6). Again, we limit ourselves to the conservative solution.
Path 2 shows the absence of all four conditions, leading to non-durability. Two cases show this particular trajectory (HeG and HaB). Both cases are struggling in every respect. Not only do they lack entrepreneurial leadership and (access to) social capital, they also grapple with problematic or weakly developed relationships with institutional key players. Moreover, their business model is unable to generate any significant income.
The presence of all conditions except for social capital also leads to non-durability. Earlier, we observed that social capital is key and necessary to understand durability. One case (EmH) shows this configuration. This case shows strong fluctuations in the core group of active members, including several changes of key positions in the board 6 The~sign indicates a low level or the absence of a condition. CAPITAL letters indicate the presence of a condition.
VAN MEERKERK ET AL.
(financial manager). There were struggles in running the CE, and also concerning dealing with professional (paid) and volunteer staff. As a result, the bonding capital in this case scores low. However, this CE is connected to a few other community actors, scoring moderately on the bridging capital, and they have succeeded in keep the business running so far. Their advisory board includes highly experienced people with strong links to key institutional players (e.g., local government, housing association). This linking capital played an important role during the 'management crisis' of the CE's executive board. Although this CE developed quite a strong business model with different sources of income (delivering paid services and renting spaces to social entrepreneurs and community associations) and received support from the local government in several ways, it is not (yet) durable. It does not make a positive return on investments and lacks sufficient staff capacity. It is uncertain whether the CE will be sustained, although the bonding capital seems to be growing now.
Lastly, we found that the presence of leadership and social capital, in combination with the absence of a business model and a supportive relationship is a path leading to non-durability. An example concerns case SiB, which is a well-organized CE that builds on bonding and bridging capital and has shown entrepreneurial leadership, but has been unable to develop their business model, predominantly due to conflicts with the local government. We observe a strong interplay between a (non)supportive relationship with local government and the (non)development of a strong business model.
The absence of strong social capital or the absence of a strong business model and the absence of a supportive relationship with key institutional players are core conditions (see appendix 2) explaining non-durability. Again, we have to interpret this with care given the limited number of cases. In Table 6 all paths leading to non-durability and durability are summarized.

| DISCUSSION AND CONCLUSIONS
Citizen-generated service delivery has gained increasing attention in public management theory and practice (e.g., Gofen 2015;Edelenbos and Van Meerkerk 2016;Brandsen et al. 2017;Kleinhans 2017). This article has examined the durability of such citizen-generated services by examining 12 Dutch community enterprises (CEs). Based on the literature, four conditions for durability of CEs have been identified. By applying fsQCA, we systematically summarize which configurations of conditions lead to a (non)durable CE.
The first key finding is that a high level of social capital (bonding, bridging, and linking capital) is a necessary condition for durability. In line with the importance of social capital as established in the civic capacity and community development literature (Chaskin 2001;Somerville and McElwee 2011;McKeever et al. 2014), our results indicate that durable CEs show productive links between CE members, to other resident networks, business and institutions in and beyond neighbourhoods in which CEs are active.
However, social capital is not sufficient by itself (cf. Saegert 2006;Dale and Newman 2010). The analysis shows that the combination of strong social capital, strong entrepreneurial leadership and a strong business model is the most important recipe for durability. Of the seven 'success' cases, six adhere to this pattern. We found a strong interplay between these conditions. CE leaders use their bonding and bridging capital to organize various activities and community events which not only serve the CE's key objectives but also attempt to recruit more volunteers to strengthen their capacity. Linking capital is used to develop ideas, create strategic alliances and seek opportunities for developing the business model. Moreover, the evidence shows that entrepreneurial community leadership is important for the development of a strong business model. The finding that a strong business model and strong social capital can be considered core conditions suggests that strong entrepreneurial community leadership plays a more indirect role: strengthening social capital and transforming ideas and social capital into specific projects leading to a strong business model.
However, there are cases in which entrepreneurial community leadership does not translate into durability. In these cases, the absence of social capital and a lack of support from key institutional players explains the non-durability. In these configurations entrepreneurial leadership is hampered in effectuating the business model for the CE as social capital is needed to recruit volunteers for different CE projects, and running business activities or some assets (e.g., empty buildings) cannot be acquired or exploited in an effective way without a minimum level of support from key institutional players. Particularly in the start-up phase, institutional support (e.g., in the form of start-up capital or handing over community assets below market price) seems to be important for CEs to develop themselves (cf. Haugh 2007). However, more longitudinal research is needed to monitor the evolution of CEs and to examine which (combinations of ) conditions are crucial in explaining growth or decline.
Fourth, the results are not decisive on the role of government support. It is clear that a supportive relationship can benefit the development of a strong business model of CEs, but this is found not to be a necessary condition for durability. Simultaneously, a total lack of support or a conflicting relationship can be devastating as shown by the paths leading to non-durability. In fact, 11 out of 12 CEs (with case ArM being the exception) are quite dependent on local government in various ways. If a local government had not provided an empty building, deployed a 'boundaryspanning' civil servant or other forms of support, most of the durable CEs in our sample would probably not have achieved their current level of durability. In the non-durable cases, the local government did not actively support the initiative or withdrew support early in the process. Even without any form or withholding of support, Dutch CEs continue to be dependent on local governments in many ways, ranging from building regulations to asking approval for having unemployed people (on social benefits) as volunteers in the business (e.g., Kleinhans 2017). This suggests that future research should make more specific distinctions between various types of government support and dependency to better capture the role and impact of the relationship with government.
Community enterprises are a relatively new phenomenon in the governance of urban regeneration, especially in the Dutch context. They try to realize autonomy and (financial) independence, while being locally accountable and committed to deliver benefits for the local community (Haugh 2007;Bailey 2012;Kleinhans 2017). However, this independence seems to be rather an exception in current Dutch practice. If policy intends to encourage the growth of CEs, local governments and housing associations can invest in providing more opportunities for taking over community assets as a form of starting capital by which CEs can develop a business model (Kleinhans and Van Ham 2017). Moreover, this offers opportunities for citizens to retain assets or to improve value for the local community in times of austerity, especially since CEs are considered to have a good understanding of community needs (Bailey 2012). At the same time, community asset transfer raises several questions for local government and CEs: for example, regarding managing financial risks, public accountability and liability. Local governments can have good reasons to be reluctant to hand over community assets, for example, if CEs cannot 'prove' that they are strongly embedded in the neighbourhood and/or have a viable business case. Acknowledging the literature on asset transfer (in particular in the UK context), more research is needed to answer the above-mentioned questions for the Dutch context. Furthermore, our findings indicate that governments should carefully take into account social capital and entrepreneurial community leadership in neighbourhoods when considering handing over community assets. An adaptive and contingent approach in community (asset) management is therefore recommended.
Although the analysis offered insights into 12 case studies, this study has several limitations. While fsQCA enabled us to unravel configurations of conditions which matter for durability, it was impossible to go deeper into the relative strength or contribution of single conditions. Large-scale (variable-oriented) quantitative research on community-based initiatives could shed more light on the role of government support, which is not found to be part of the sufficient configuration, but can affect the durability and performance of community-based initiatives. Furthermore, longitudinal research designs could clarify the role of government support in different stages of a CE's development.
In the wake of the substantial rise in community-based activities, future research could reveal whether our findings are more widely applicable, including to other examples of community-based initiatives. The focus on Dutch CEs contributed to the internal validity of our research as it generated a relatively high level of comparability between the institutional contexts of the cases. Cross-country comparative research would, in particular, enable analysis of the role of the institutional context (policy and governance traditions) in the emergence and evolution of CEs. Path 1: HIGHCAPITAL*BUSINESSMODEL (cons. 1, PRI: 1, coverage raw: 0.86), n = 6.
Paths leading to non-durability.