Dissolution of expressive and instrumental voluntary associations: Four key dimensions in explaining organizational change

This study inspects the reasons as for why voluntary associations end their activity and dissolve as legal entities. The empirical data was gathered through semi-structured interviews with former leaders of 30 organizations in Estonia and Finland. To appreciate the sector’s diversity, the analysis makes a distinction between associations with mainly expressive, and those with mainly instrumental functions. Causes for the associations’ dissolution fall into four different categories related to (1) the association as a social entity, (2) resources, (3) goals and (4) environment. Reasons belonging to the first category of internal, social reasons, such as problems of leadership, internal organization and membership recruitment were the most common ones. However, several reasons often intertwined. A comparison between associations of different types showed that expressive associations were most often threatened by causes in the first category, while they were surprisingly immune to the other ones. Instrumental associations are dependent on other actors and external circumstances, and the reasons for their dissolution are likely to be related to problems with resource acquisition, goals or environments.


Introduction
The dynamics of organizational change have always been at the heart of organizational studies, and much research deals with the mechanisms behind change. Researchers of the voluntary sector are often interested in the "best practices" and therefore focus, for good reasons, on the most successful organizations -on how and why they both survive and succeed. In this paper, we focus on something that could even be called "worst practices" and propose a way of examining the final life cycle change of associations, i.e., their dissolution. Research on processes leading to dissolution can contribute to our understanding of the strengths and weaknesses of the association form.
Voluntary associations differ in many ways from other dominant forms of organizing, such as bu-reaucracies and enterprises. Unlike these, associations' two main features are that they pursue mostly non-economic collective goals or interests of the participants, and that membership is voluntary (Bromley & Meyer 2015, p. 13). They rely mainly on the members' voluntary, non-remunerated work. In short, social and normative factors are crucial. This can make voluntary associations fragile, but sometimes also very tenacious. Some of the most short-lived organizations, but also some of the oldest ones, are in the voluntary sector (Anheier, 2005). While the social dimension is often the voluntary associations' most important resource, it is also a main explanatory for their dissolution. Research has seldom addressed this issue. We wish to point out two challenges for future research. First, the study of organizations in the nonprofit and voluntary sector has often overlooked voluntary associations in favor of other, more formalized organizational forms, such as nonprofit agencies and corporations (Harris, 1998). David Horton Smith (1997) has argued that research on non-profit organizations (NPOs) mainly focusses on organizations with paid staff, thus ignoring grassroots associations, or in his estimation, 90% of the organizations in the sector in the US. Second, the field of voluntary associations has developed toward greater fragmentation and diversity regarding, e.g., operative goals, dependency on other actors, level of formalization and professionalization. This poses serious challenges for researchers and requires new concepts to be applied.
A recent literature review on the life cycles of voluntary associations (Sundblom et al., 2017) concludes that the transition phases between life cycle stages are under-researched. This also includes the reasons for transition to the final life cycle stage, which is death. While the authors plead for studies done "on random samples of associations in some territory" (p. 964), we have confidence in a qualitative approach to comprehend the subtleties of the reasons for dissolution. While quantitative studies are able to point out the characteristics that correlate with dissolution (Sundblom, 2020;Wollebaek, 2009), a qualitative inquiry explores the actual reasons leading to organizational death.
The present article discusses the closing down of voluntary membership associations in two Northern European countries, Finland and Estonia. It is based on interviews conducted in the two countries in 2012 and 2013 with (former) members of organizations that within the past two years had terminated their legal existence, or were just about to do so. We thus define an organization's closure as the end of its legal existence as an association. Admittedly, this definition is not self-evident. As we will see, the legal termination of an association does not completely coincide with the end of its activities. Very often, it has been on the decline during a longer period, to which the termination just puts the end. In a few cases, (some of) the members continue with similar activities, but no longer as a separate association. However, we consider the legal termination of an association as a close enough indicator of its "death". 1

Theory and research on closure of organizations in the nonprofit and voluntary
sector Several theoretical models have been tested empirically in order to explain the dissolution of NPOs. Fernandez (2008) notes that resource dependency theory (Pfeffer & Salancik, 1978) is underexploited as a model for explaining closure of voluntary associations. The theory's starting point is that organizations are interdependent, i.e., depend on others for different resources, which will ultimately affect their chances of survival. The fewer the sources of resources, the higher the risk of dissolution, as organizations with a limited number of dependencies are more vulnerable to disruptions in one of them. For instance, associations are often highly dependent on public funding, the discontinuity of which can have terminal consequences. Fernandez (2008) finds support for resource dependency theory in his work on dissolved Spanish associations.
As an outcome of the marketization of the nonprofit and voluntary sector, many NPOs compete with each other and with for-profit organizations for, e.g., service recipients, resources and funding. It comes, hence, as no surprise that in certain fields, market mechanisms explain many closures of voluntary orga-nizations. Among the reasons, we find failure on the market, legitimacy crisis (Hager et al., 1996;Hager, 1999), and financial vulnerability (Hager, 2001).
A frequently reported, rather different case is that NPOs also close down because of mission completion (Hager et al., 1996;Hager, 1999;Fernandez, 2008). An organization that considers its mission fulfilled might simply conclude that its activities are no longer needed. For a for-profit organization, an organizational closure is always a failure, while for a mission-lead nonprofit it might in fact signal success. After reaching its initial goals, the organization may of course also decide on another option: to pursue other goals. March of Dimes 2 is a prime example (Hager et al., 1996).
These significant findings about the closing down of NPOs may however have limited validity when voluntary associations are concerned. The object of study has mostly been the highly formalized NPOs, i.e., organizations with a high degree of standardization regarding for example policies, routines and organizational roles among members. The results are fully relevant for only a fraction of all voluntary associations. For instance, Hager et al. (1996) draw their sample out of IRS (Internal Revenue Service) tax exemption approved public charities, thereby excluding by one estimation 90% to 95% of all associations. 3 Similarly, Hager's (2001) data on arts organizations and Twombly's (2003) data on human service organizations build on IRS data sources. Outside the US, Fernandez (2008), who studies the reasons for closure among officially deregistered Spanish associations in the Madrid metropolitan area, notes that more formalized associations are probably overrepresented in his sample, "particularly those with some physical capital or bureaucratic characteristics" (Fernandez, 2008, p. 134).
A more inclusive approach is offered by Wollebaek's (2009) study on the survival of associations in one of Norway's nineteen counties. The composition of the study includes both formal (in Wollebaek's interpretation, "mechanistic") and informal ("organic") associations, and the former are proven to outlive the latter ones. Wollebaek's study finds, furthermore, that "voluntaristic" factors, i.e., factors that the associations can influence themselves through management, e.g., through their board structure and activities, are of nearly as high significance in explaining survival as are deterministic factors, such as age, size and organizational density. Speaking of the particularity of voluntary associations, the author remarks on an "absence of competition effects with regard to organizational survival" in the population he studied and that "[v]oluntary organizations seem more tenacious than other organization types" (Wollebaek, 2009, p. 279).
Comparisons of subgroups of associations seem helpful for understanding associational decline. In a quantitative study of registered voluntary associations in 49 municipalities across Finland, Sundblom (2020) finds that so called old associations, i.e., those founded before 1989, comply better with theoretical assumptions than new associations, founded between 1990 and 1997. For example, and in accordance with institutional theory, embeddedness in (organizational) society protected old associations from dissolution, while embedded new associations did not gain similar protection. For the latter, a large membership base and non-leisure activity type were the only explanatory variables that correlated significantly with survival (Sundblom, 2020, p. 255).
Obviously, studying a wide variety of associations deepens our understanding of how they change. To sum up, some major conclusions can be drawn. First, research has tended to neglect a large part of the associational field (cf. Smith, 1997Smith, , 2000. Especially associations with few members, little funding and a low degree of formalization are excluded from many studies. Consequently, the theoretical understanding of the mechanisms involved remains limited and biased. The prevailing organizational theory with its origin in the research on private companies and public agencies is obviously better suited for studying highly formalized NPOs. For research in voluntary associations, there is a need for theories that take into account the particularities of their organizational form.

The approach of this article
Our theoretical starting point aims at merging four different paradigms within the sociology of organizations. They analyze organizations as either rational or natural systems, and either as closed or open systems (Scott & Davis, 2007). 4 The four paradigms shed light on different aspects of organizations and are perfectly valid on their own. However, their focus on certain features of organizations necessarily de-emphasizes others. We therefore propose an inter-paradigmatic approach that takes into consideration either rational or natural system perspectives, combined with either closed or open system perspectives. We define organizations as: (1) social entities that apply (2) resources to achieve (3) goals in the (4) environments in which they operate. This definition brings to the fore four different, but equally important aspects of organizations, that we find relevant for explaining the survival or non-survival of organizations of different types. 5 Let us now specify what our definition means for voluntary associations and how it is helpful in explaining their dissolution: As distinct social entities, voluntary associations have characteristics that predispose them to certain challenges that will influence their prospects for survival. Importantly, associations are in principle ruled in a democratic way, and membership is non-compulsory. Furthermore, associations are chiefly run by volunteers and do not ultimately depend on paid staff to run their activities (Harris, 1998). Research on the specific challenges that voluntary associations face due to their organizational form has addressed member entry, retention and participation (see Tschirhart, 2006 for a literature overview). Other possible obstacles for the reproduction of the association are the absence of managerial rotation and low quality of governance. As in any other organization, there are both formal and informal structures. The latter are emphasized by the closed-natural systems paradigm (Scott & Davis, 2007), and include, e.g., internal strife because of dysfunctional personal relations, and operational difficulties due to conflicting goals among members.
Resources are necessary for voluntary associations, and lack of sufficient resources is in some cases expected to explain dissolution. Their amount and type needed for different activities of course varies greatly-a book circle and a volunteer fire department may serve as examples. Even the ways of obtaining resources varies; a key distinction is between external and internal ones. Papakostas (2011) points at an ongoing shift away from a membership-centered way of organizing. External resources received from other organizations have become increasingly important at the expense of internal ones, such as membership fees. This said, the dependence on external resources varies vastly across the associational field. For resource-dependent associations, a decrease in external funding can lead to dissolution, while others may overcome the change by, for example, scaling down their activities, or by mobilizing resources elsewhere. Similarly, a rise in expenditure can in some cases lead to dissolution. The importance of non-monetary resources varies as well; many associations obtain external non-monetary resources such as facilities (free of charge or at a low cost), non-monetary help, services and favors. Dissolution linked to insufficient resources can thus in practice include the loss of access to a gathering hall, or a decrease in voluntary help from non-members.
In two quite distinct situations, the goals of an association, i.e., what it is meant for, are relevant for explaining its dissolution. Both reaching the goals, and having to admit their unfeasibility can prove fatal. A key distinction in this article is that between expressive and instrumental associations, and a key difference between them is about their purposes; that is, the difference between leisure-oriented and function-oriented purposes of associations, which is defined in more detail below. While the purpose of an association can be quite straightforward and specific, e.g., to provide the platform for a self-help group, the operative goals can be many and change over time. In the case of a self-help group, operative goals could be securing outside resources and recruiting members. Associations can target objectives outside the organization or within it (as in self-help groups), or the activity can be a goal in itself (as in leisure associations).
The association's operational environment includes both risks and possibilities. Within organizational studies, the importance of environments has been emphasized by the open systems paradigm (Scott & Davis, 2007). From the point of view of this article, the most important are the institutional, organizational and local environments. As legal entities, associations have clearly defined rights and obligations, which facilitate or hinder their activities. Associations' institutional environment is, thus, also subject to policy changes. Other organizations, with which they interact, e.g. the municipality, local businesses and other nonprofit organizations, are also part of the environment. The interaction can be competitive or corporative by nature; hard competition over members, participants or markets might lead to dissolution, while cooperation around joint interests has the opposite effect. The local environment can both facilitate and hinder voluntary activity. Obviously, resource mobilization from the environment and the dependencies that come with it (Pfeffer & Salancik, 1978) have an impact on survival. On the other hand, the depopulation of a municipality will alter the operating conditions for most institutions and organizations, including local voluntary associations.
A second analytical tool we apply is a division between expressive and instrumental organizations. The division is well known, and has been interpreted in more than one way. In current literature, an interpretation inspired by Frumkin (2002) seems to prevail. It divides between the two types of NPO based on their societal function: service providing organizations count as instrumental, while advocacy organizations are examples of expressive organizations. Here as often in current research, focus is on formal and professional organizations, while many typical forms of voluntary associations fall below the radar. We prefer to build instead on Gordon and Babchuk's (1959) classical definition, according to which such functions as service provision or advocacy both belong to the repertoire of instrumental organizations, while leisure associations can be regarded as expressive organizations. There is reason to keep in mind the broad and wide field of voluntary associations that undoubtedly includes both leisure and other activities.
Following Gordon and Babchuk's (1959) distinction, we divide the associations into two categories. First, there are those with primarily expressive functions, i.e., which offer socialization and leisurely activities. Here, the goal of the activity is the activity itself. Thus, the expressive associations provide a social platform for the participants to engage in hobby-oriented activity and to express themselves. Examples of such associations include choirs and theaters, football clubs and dog associations. The second category comprises organizations with primarily instrumental functions, i.e. advocacy or service provision. Instrumental associations "serve as social influence organizations designed to maintain or to create some normative condition or change" and "exist in order to attain goals that lie outside of the organization" (Gordon & Babchuk 1959, p. 27). Hence, associations that aim at achieving external change belong to this category. Examples of instrumental associations are trade associations, environmental associations, and homeless people's associations. The distinction between expressive and instrumental associations is based on general goals and types of activity, but it has consequences on other aspects as well. Professionalization and organization building, for example, are more typical among instrumental associations than expressive ones.
We expect that reasons for closing down will differ between associations with predominantly instrumental or expressive functions. The social aspect could be the predominant reason for closure for expressive associations; their ability to motivate members and to foster internal cooperation is imperative in order to survive. Reasons for the disbandment of associations with mainly instrumental functions will have more variation; an interplay of several reasons is likely. Especially, the well-established link between resources and environments is likely to be strong for instrumental associations that depend on external resources. Presumably, the four dimensions interplay in other ways as well. For both expressive and instrumental associations, we expect that the social core of associations is vulnerable to changes in the other three dimensions (resources, goals, and environment).

Voluntary sectors in Finland and Estonia
The empirical material to be discussed in this study refers to 30 organizations in Estonia and Finland. Such a comparative setting highlights the influence of macro-level social factors. These factors include resources available in Finland, a country with a high, and Estonia with a medium high economic standard; respectively, in a country where associations are regarded a part and parcel of a democratic society, and a country where they only are seeking their place in public life; and above all, where participation in associations is a regular and highly appreciated part of everyday life, and where it is not. The two countries represent two sides of the former Cold War East/West divide. Despite geographical proximity and frequent contacts since the 1990s, developments in the two countries differed from each other radically. In Finland, voluntary activities were since the 19 th century an important element of the country's social model. Finland, with a population of 5.5 million and approximately 105 000 registered associations 6 (Finnish Patent and Registration Office, 2020), now belongs along with Scandinavia to the countries where participation in voluntary associations is among the highest in the world. Estonia, on the other hand, lost her independence and was between 1940 and 1991 incorporated with the Soviet Union as one of the Soviet Socialist Republics. After the restoration of the country's independence, the previously almost non-existent voluntary sector was created anew. Today (2019), Estonia (population 1.3 million) has about 25,000 such registered nonprofit organizations and foundations, whose aims go beyond the immediate economic interests of their members (Estonian Statistical Council, 2020). 7 There is a clear tendency towards a decrease of the median age of NGOs (Rikmann et al., 2010). According to statistical data, almost half of Estonian NGOs have been in existence for five years or less, and both in Estonia and the two neighboring Baltic countries Latvia and Lithuania, out of ten newly founded NGOs, only three remain active also three years later (Huber, 2011). Despite this, membership in an association still is more frequent in Estonia than in most other postcommunist countries of Eastern Europe, even if it remains lower than in Western Europe and Scandinavia (Rose, 2006, p. 16).

Data and method
For our study, we conducted semi-structured, face-to-face interviews with key representatives (chairpersons or board members) of in total 30 associations that had recently ended their activities legally in two geographical regions in Finland and Estonia-Finland's southwestern corner with the city of Turku and its environs, and Estonia's capital Tallinn with the adjacent municipalities. Fifteen interviews were conducted in both countries by the authors in 2012 and 2013. 8 The transcripts were coded and analyzed with respect to the stated reasons for dissolution. The initial sample was drawn from registry data over legally dissolved associations with the aim of including associations from five different fields of activity: 1) Health and welfare, education; 2) Sports and exercise; 3) Politics and advocacy; 4) Local community development organizations; 5) Organizations for active cultural participation. We designed our sample to include associations from these five different fields in order to secure its internal variation. No comparison between different fields of associational activity will be presented, however; the small size of the sample does not allow such a comparison. Neither are the conclusions of our study based on any quantitative measures, but on an analytical comparison of the cases. Contact details for the organizations were found through the Internet, and former members in a leadership position were contacted either by e-mail or by telephone. In Finland, there were only two cases when the contact person was reached but refused an interview; in Estonia, approximately half of the individuals contacted refused. It seems that in Estonia, the registry data included a larger share of such organizations that had acted for merely a short period of time (or not at all). In such cases, the contact persons usually stated that they would "not have anything to tell" about a short-lived project and that anyhow, "they no longer remembered" what had happened. In other words, board members of short-lived organizations did not feel sufficiently committed to these initiatives in order to find time to make an appointment to discuss them.
Those who did respond were in both countries usually well informed about the association and about the events that led to its closure. Of course, what they presented was just one view of the events; however, it was not merely their personal view but that of the persons in charge at the time. Despite the possible existence of other views among the former rank and file members, the one that our respondents held is, in fact, the closest possible approximate for the closure's immediate reasons-after all, they were the people who made that decision.
In their heydays, the associations had a membership of 2-200 (Estonia), or 9-150 (Finland), and annual budgets ranging (in Estonia) from no income to 230,000 € (a theater), or (in Finland) from a few hundred to 40,000 € (a yoga club). Most of the associations in the sample had a sparser budget as shown by the medians: 400 € for Estonia, 1000 € for Finland; and 1000 € for the sample as a whole. The questions we asked during the interviews were about the association's purposes and activities, members, resources, its external relations, the internal organization and relationships between the members, and about the reasons for its final inactivity and closing down.
In our analysis, we focused on the stated and direct reasons for decline and closure. The reasons discussed are thus the ones directly stated by the interviewee, or, in some cases, such that became apparent during the interview-in other words, the situation as it appeared when seen from the inside.
The categorization of reasons for dissolution along four explanatory dimensions (social, resources, goals, and environment) followed a systematic and thorough reading of the transcribed interviews. As the variance of the studied associations was vast, e.g. in regard to internal structures and external environments, resources and purposes, as well as their respective dissolution narratives, we saw a need for a framework that would level out the differences in the sample. This process resulted in an analytical common ground by unveiling similarities among the associations and in the reasons that lead to their dissolution. The second framework we employ distinguishes between associations with primarily instrumental and primarily expressive functions. In our analysis, this distinction roughly follows the associations' field of activity: Most organizations involved in cultural and leisure time activities have been classified as expressive, while the others have been labeled instrumental. There are two exceptions: an Estonian theater group and a Finnish Yoga club. Both were the associations with the highest annual income in their respective countries, and both could have been characterized in their peaks as hybrid organizations, i.e. as borderline cases between voluntary associations and for-profit businesses.
The combination of these two frameworks allows us to present a total of four propositions. We are aware that generalizations retrieved from qualitative data can be problematic, e.g. as the sample size and representativeness are often insufficient. However, these propositions are not axiomatic generalizations but are offered as "moderatum generalizations", that is, as "testable propositions that might be confirmed or refuted through further evidence" (Payne & Williams, 2005, p. 297).

Results
In most cases, our respondents' narratives of their associations' decline and closure refer to problems related to more than just one of the explanatory dimensions (social, resources, goals, and environment). Table 1 shows the reason for closure, expressive or instrumental function, short description, age (in years), membership size and country. In 12 cases out of 30, the reason for closure lies in just one of the four categories. Consequently, in 18 cases the reasons for closure belong to several categories. A first glance at Table 1 shows that the social factors contribute to most dissolutions. The second most important reason, environmental change, was much more prevalent among the Estonian associations (in 10 of them, as compared to 5 Finnish associations). This can be explained by the rapid economic and social change in Estonia, but there are also more specific legal reasons that will be discussed below. Lack of resources was also more common in the Estonian associations. We may add that some of the associations continued their activity even after legal termination, either as clubs without legal status, or as parts of larger organizations. In the table, they have been marked with footnotes.

Dissolution in expressive associations
A key finding is that among the expressive associations, social reasons were reported as decisive or partial causes for dissolution in all but two cases. In our material, the only exceptions are the two Estonian associations, a choir and an arts and crafts association, which in fact continued their activity even after the legal termination. Expressive associations appear to be extremely tenacious when they face other than social difficulties. This leads us to the following proposition: Proposition 1: Expressive associations are inherently tenacious and mostly close down because of social (internal) reasons.
The most notable reasons for closure in this social category are members' inactivity, failed recruitment, internal strife, and lack of motivation among leaders to carry on with the activity. Consider the case with a Finnish local theater in a small town where the chairman/producer's moving away from the locality caused the inactivity of the association since no replacement was found:

Local theater, Finland, 23 August 2013 R: Everything in its time, it is rather a heavy business to lead such a large activity. […] In many contexts I have seen that everything may depend on just one individual. Meaning that there is one active person who says what should be done, and s/he will need a team around her/him who accepts it and cooperates. You can do nothing alone. But if you lack the person whom people
trust and who has ideas and leadership capacity, then the activity has no sparkle.
The respondent represents a leader-dependent association, in which only one person possesses much of the relevant knowledge and leadership needed for the activity. In a similar vein, the loss of special competence, as in the case of the inactivity of playwrights in the role-playing association, can lead to termination. Hence, our second proposition: Proposition 2: Expressive associations that depend on special competence and/or knowledge for their activities are threatened by the loss of the person(s) who possess the necessary cultural capital.
In most expressive associations, the main reason for closure was related to a failure to recruit new active members and leaders. While both expressive and instrumental associations suffered from this, it was more prevalent in expressive associations. There was a loss of members because of transitions between life cycle stages, mainly graduation, parenthood, ageing and death. As an example, the activity decreased in a Finnish live action role-playing association when the members graduated from university, started working and became parents. The relationship between membership and members' life-cycle transitions has previously been studied by Rotolo (2000).
Most other problems are about resources and the environment. As to resources, expressive organizations tend to be internally funded, and thus depend on membership fees (and possible other fees linked with the activity). For example, an Estonian cultural hobby club closed down as its aging members could not afford to pay even a small membership fee. Another example is a Finnish association that organized gymnastics classes for women and children. The development towards reduced income from membership fees, and finally dissolution, began from difficulties of recruiting gymnastics instructors, who worked for a small remuneration. These instructors, who were teenagers and young adults, preferred instructing classes in commercial studios where the compensation was better. One of the goals of the association was to offer exercise to women and children at a low cost, but they could no longer afford paying the instructors. The classes became less frequent, which led to a downward spiral. The example also shows the voluntary association's vulnerability when competing with for-profit actors.
As we mentioned earlier, also the legal environment can affect the survival of associations. Our interviews in Estonia witnessed the effects of one particularly dramatic example of this. The Associations Act (Mittetulundusühingute seadus) was passed and came into force in 1996, and by 2020, it has been modified 36 times. One important new regulation came into force in 2010, according to which all associations came under the legal obligation of submitting their annual budget reports electronically to the taxation authorities. A failure to submit the annual budget report will result in the deletion of the association from the national registry, and thus, effectively, end its legal existence. The new regulation was intended to clear the registry from inactive associations and encourage their professionalization, but at the same time, it discourages small-scale activity typical of small associations with expressive functions. In the present study, there were two associations whose representatives claimed that the organization was erased from the registry and lost its legal status only because of the new reporting requirements. Both organizations-a choir founded in the late 1980s, and an association that had offered art courses for people with mental disabilities since 1998-were at the time of the interviews fully active, but no longer legally registered. The social dimension of organizing is of critical importance for expressive associations. The fostered togetherness holds the expressive associations intact but when the social equilibrium fails, the mechanisms that would provide a safety net are non-existent. Some expressive associations depend greatly on people with special skills (such as choir leaders), and replacing them might prove difficult. The renewal of members and actives sometimes caused overbearing challenges. The strong cohesion that often goes together with expressive activity may prove a challenge for the recruitment of new members, as it also raises the threshold of joining the associations as a new member.

Dissolution in instrumental associations
While the social aspects of organizing are well represented among the causes for closure of instrumental associations, present in 15 out of 21 cases, the primary reasons are nevertheless to be found among the other three dimensions. The social reasons (such as failed renewal of membership, possibility of merger with an affiliated organization, or members' life-cycle transitions) appear as mere accompaniments, or coeffects of other, more basic reasons for closure. Overall, the range of reasons for closure is more complex among instrumental associations than among the expressive ones. In all but four cases, the reasons are multidimensional.
In the category of instrumental associations closed down because of changes in their environments, we find altogether 12 cases. Some of them are the ones that accomplished their goals in their local environment; others were affected in a negative way by changes in the local environment, e.g. competition, loss of membership-base due to depopulation or gentrification, and regulatory changes. By way of example, two Finnish local development associations were active within a small locality, one comprised of a village and the other of ten small villages. Both localities underwent crucial structural changes, such as reduction of municipal services because of depopulation. The local elementary schools, which had served as the villagers' gathering places were closed down. In both cases, the association had tried, and for a while succeeded, to maintain the school.
In Estonia, two associations closed down because new organizations with similar activities and goals emerged. In the case of an ethnic minority's cultural association, aging and passive membership along with the emergence of a competitor finally caused it to dissolve. In the case of an association that promoted bicycle riding, the goal was partially met as cycling became more popular, and they could even claim some success in influencing city planning. However, no new coherent goals emerged either, at the same time as a new larger association competed for members. Eventually, the founders lost interest in the association. A First Aid association was founded in the mid-1990s to take over the functions of a former governmental agency. Soon, it found itself working without sufficient funding, in tough competition with a private enterprise, and with most of the governmental agency's former assets privatized by individuals outside the organization.
Resource dependency was also a challenge for some instrumental associations. In four cases, including the First Aid association depicted above, access to external resources disappeared. Another risk is the rise of costs to an unbearable level. One example is a commercially successful theater in Estonia that could no longer afford the rent after the owner of the facilities raised it over their paying capabilities.
In one case, quite the opposite became a problem for the continuation as an association: A yoga club decided to become a for-profit organization since its classes and courses were so popular that their profits grew to a point that they risked losing their tax benefits. It had however not paid salaries to the instructors but offered compensation in the form of further education. This shows how voluntary associations can provide a low threshold platform for aspiring for-profit organizations. Some of the instrumental associations interviewed were in practice merged with their umbrella organizations. This is an option available for organizations within hierarchical structures. In this way, members of the dissolved association can be encouraged to join the umbrella organization and participate in its activities, while the old association's remaining monetary resources will continue to benefit the same cause. Merging non-functioning entities with functioning ones is part of what Papakostas (2011) calls the rationalization of civil society.
Not all dissolutions are to be considered failures-on the contrary, in some cases they are results of success. Six associations, all with primarily instrumental functions, were dissolved after having attained their stated goals. Two Finnish associations worked towards concrete goals: A night patrol that sought to prevent disturbance in a neighborhood; and a local development group that aspired to protect an old building from being torn down to be replaced by a department store. After having achieved their purposes, and not finding a new agenda to follow, both associations closed down. When the focus is only on the local environment, setting new goals can prove difficult after the initial ones are achieved. In a similar manner, two Estonian activist organizations saw the fulfillment of some of their initial goals. One was the above-mentioned association for the promotion of bicycle riding; and the other an environmental organization founded by university students that upheld an environmental website with information about alternative and sustainable technology. At the same time as the goals of spreading information about environmentalism were accomplished, the founders graduated and moved away from the university town. Finally, when the website became chargeable for the association, the board began the liquidation processes. As predicted, resources, environments and goals are more important for explaining dissolution in instrumental associations than in expressive ones, while strictly or even primarily social reasons are clearly marginal.
Proposition 3: Instrumental associations tend to close down because of reasons having to do with resources, goals or environments, and are thus vulnerable to external change.
Instrumental associations are, by their very character, more engaged outwards, actively creating connections and dependencies with other actors. Problems in the internal life of the associations are less prevalent, probably in part due to a higher degree of formalization, which provides a safeguard from internal problems caused by power strife or avoidance of responsibilities. On the other hand, failure to see the importance of social issues might hinder seeking solutions to the problems faced. At one point however, difficulties may arise that visibly endanger the organization.
Proposition 4: Internal difficulties may precipitate the ending of activities in instrumental associations that face challenges with regard to their function, while internally strong associations are more likely to seek a solution to the problems and continue their activities.

Discussion and conclusions
In this article, we set out to answer why voluntary associations end their activities. We presented the results of an empirical study conducted in Finland and Estonia that included 30 associations varying as to their functions and to their degree of formal organization. We introduced four analytical dimensions pertinent to the study of organizational change and eventual end of the life of an association. We call them social entity, resources, goals and environments. Although challenges faced by the association as a social entity (difficulties in its internal organization or membership) were present in most cases of dissolution, they were in some of them triggered by other problems, which finally led to frustration and inability to motivate activity. Our analytical distinction between expressive and instrumental associations helped us to identify some crucial differences within the vast and heterogeneous field of voluntary associations. Above all, it showed the expressive associations' relative immunity to problems outside the field of the social. In contrast, the main reasons for cessation for the instrumental associations were achievement of the initial goals, resource insufficiency, and changes in the environment. Associations that dissolved for other reasons than the social were in fact all such whose activity was primarily focused on influencing other parts of society, i.e., associations with a primarily instrumental function.
All the four major paradigms of organizational theory (Scott & Davis, 2007), i.e., the combination of rational or natural perspectives with closed or open-system perspectives, proved relevant for our study. However, we find that much of previous research has unduly neglected the natural-closed system perspective, which highlights informal relationships in an organization. Previous studies on the dissolution of voluntary organizations have mainly dealt with highly formalized associations with instrumental functions. This focus has obscured the social dimension of organizing.
There were some differences between the two countries: The Estonian associations were more often than their Finnish counterparts affected by environmental change. The economic, legal and political development during the past two decades has been much faster in Estonia than in Finland, meaning that the associations' operating environment has been less predictable. In both countries, and indeed globally, recent social changes require an ongoing rationalization of organizational activities as the social world becomes more formalized (Bromley & Meyer, 2015). This could prove crucial especially for small-scale expressive associations that are not ready to make the necessary adjustments. Digitalization has, on the other hand, made information and services increasingly available for the purposes of associational development.
A managerial implication of our study is that expressive associations should make sure not to neglect organization building. A more professionally led organization with a clear distribution of tasks and a strategy for the mobilization of members and resources will be fit to meet various external demands and eventual internal difficulties, such as leader changes and conflicts of opinion. Instrumental associations should, on the other hand, besides keeping up with their scheduled activity as best they can, foster internal cohesion, as this will probably help getting through challenges on the other fronts. They can learn something from the relative tenacity of expressive organizations.

Endnotes
1 For a discussion about the concept of organizational death, see Hager et al., 1996. 2 March of Dimes was founded in 1938 to eradicate polio. It completed its mission as a vaccine was developed in the 1950s. Since then the organization has changed mission to prevent birth defects and later on to promote healthy pregnancies. 3 Based on Smith's (2000) estimation that only 5-10% of grassroots associations in the United States are IRS registered. 4 The rational system perspective highlights goal specificity and the importance of the formal structure.
The natural system perspective, on the other hand, sees similarity of organizations with other social collectives and acknowledges the possibility of goal complexity among participants, emphasizing the organization's informal structure. The closed system perspective focuses on the internal relations and processes in organizations. The open system perspective stresses the organization as a part of the wider environment and its relations with other organizational actors and institutions (Scott & Davis, 2007). 5 Our model resembles the one suggested by Helmig et al. (2014), which is a four-dimensional model for explaining the success and failure of NPOs, although their focus on efficacy and performance limits its suitability for associations over the entire spectrum. 6 An association can also function without formal registration, and the number of nonregistered ones is unknown. An educated guess would be, that around a fourth, or an additional 30,000 of the Finnish associations active today are not formally registered. 7 The total number of non-profit organizations in the registry was 40,000, but possibly around a third of them were condominiums, garage associations, and the like. 8 The respondents were promised anonymity, but they were aware of the fact that the association they represent might be recognized by an informed reader. This is in accordance with our universities' guidelines for ethical research practice.