The Strength of Family Ties and Community Social Capital: A Mixed-Methods Examination of Zadar, Croatia

Introduction

Scholarly literature has recognized Robert Putnam as one of the most influential figures in popularizing the concept of social capital and as having made the greatest contribution to the research in this field (Castiglione et al. 2008; Field 2008). His works revealed positive outcomes of social capital in various spheres of individual and social well-being, including economic development and quality of governance (Putnam 1993), living in safer neighborhoods, having a happier and healthier life (Putnam 2000), improving life chances, and mitigating poverty (Putnam 2016). Although, it should be acknowledged that the outcomes of social capital are not only positive, but can also be negative (see, e.g., Baycan and Öner 2023).

In Putnam’s (1993) tradition, social capital refers to trust and norms embedded within social networks that foster collaborative actions, and appears in different forms. Of particular importance for the present study is the key distinction, as set out in Putnam’s (2000) theory, between bonding and bridging social capital. Bonding social capital is formed within networks of individuals with a strong group identity, where the members of that specific group provide one another with social support. As a consequence of the strong group orientation of the members of these networks, this type of social capital excludes outsiders or may even enable in-group benefits at the cost of outsiders. Conversely, bridging social capital connects individuals in a way that transcends boundaries between groups formed on the basis of shared identities, and impacts the public sphere in a way that extends beyond specific intergroup processes. Putnam (2000, 23) acknowledges that this distinction is theoretically problematic since groups may “bond along some social dimension and bridge across others”. Thus, this distinction should not be understood in terms of “either/or”, but rather in terms of “more or less” categories. The literature also identifies “linking” social capital, a type that denotes vertical relations between individuals and power holders, such as people from the world of institutionalized politics (e.g., Grootaert and Bastelaer 2002).

In Putnam’s (2000, 21) work, family ties are seen as a “form of social capital”. In his comprehensive study of social capital in the United States, one of the measures he included was socializing with family members (Putnam 2000, 100–1). Research that addresses topics broadly similar to those covered by this article generally designates family ties as a specific type of bonding social capital (Karametou and Apostolopoulos 2010). The other type of social capital, measured using similar indicators to those used in this study as an outcome variable, namely generalized trust, association membership, charitable giving, and voting in the elections, is generally understood as bridging. Since the latter type of social capital includes cognitive processes and actions oriented toward the broader public sphere, it is sometimes also referred to as “civic” social capital, or the terms “civic” and “bridging” are used interchangeably (Kurowska and Theiss 2018; Żukowski et al. 2009). The terminology based on these theoretical elaborations will also be applied in the present study, unless the specific type of social capital is obvious from the context of the discussion.

In one of the most influential studies ever written on social capital, Making Democracy Work: Civic Traditions in Modern Italy, Putnam (1993) saw amoral familism, a set of values with a strong orientation toward improvement in the material well-being of one’s own family at its core, as the main impediment to the development of civic social capital. Later studies further examined this conceptual relationship and largely confirmed that the strength of family ties has an erosive effect on various aspects of civic social capital. This research has generally addressed one (Alesina and Giuliano 2015; Durante 2010; Ermisch and Gambetta 2010; Herreros 2015; Marè et al. 2020; Ljunge 2015; Kim and Wilcox 2013), two (Alesina and Giuliano 2014; Daniele and Geys 2016; Ljunge 2011), and very occasionally three (Alesina and Giuliano 2011) aspects of civic social capital, and has drawn insights based on quantitative data from national samples. On the other hand, mixed-methods research examining the relationship between strength of family ties and civic social capital on the local community level, including more constituents of civic social capital, is rather scarce. To the author’s knowledge, the only study conducted along similar lines is the one by Karametou and Apostolopoulos (2010), which focuses on two rural mountain areas in Greece.

This article presents the results of a sequential mixed-methods study analyzing the relationship between the strength of family ties, as a measure of bonding social capital, and four constituents of civic social capital – generalized trust, association membership, charitable giving, and voting in elections. These processes were examined in the Croatian city of Zadar. The goal of the study was to determine the role of the strength of family ties in the formation of each of the four abovementioned components of civic or bridging social capital.

Key Constituents of Social Capital

In his seminal work on the quality of governance in 20 Italian regions, Putnam (1993, 167) defined social capital as “features of social organization, such as trust, norms, and networks, that can improve the efficiency of society by facilitating coordinated actions”. These constituents of social capital operate as a composite mechanism in which they mutually support and reinforce each other. Trust is understood as a pivotal part of that mechanism, since different forms of cooperation require a certain amount of trust, while conversely, cooperation itself generates trust (Putnam 1993). Particularly important is generalized trust, a form that extends beyond the range of people one knows personally, or: trust toward people in general (Putnam 2000). Networks of engagement incorporated within civic associations also play an important role in forming social capital and are therefore considered to be a “useful barometer of community involvement” (Putnam 2000, 49). Such networks include actors with a broader range of identities, thus making it possible for trust to extend beyond the scope of family and kinship, enhancing cooperation among actors who do not form tight-knit groups of individuals already well known to one another (Putnam 1993).

Civic engagement appears in different modes and some of these, such as political participation, are particularly relevant for assessing the quality of a community’s stock of social capital. An important form of political participation is voting, although the practice does not strictly fall under the domain of social capital, given that it is done individually. Nonetheless, it is “almost always a socially embedded act” and strongly linked to other forms of engagement, which makes it a “very useful proxy measure of social involvement” (Putnam 2000, 447n2). Similarly, this is also the case for acts of philanthropy, such as charitable giving, since such a practice demonstrates a form of solidarity and altruism toward members of society and should therefore be included in “any assessment of trends in social capital” (Putnam 2000, 116–7).

Family Ties and Civic Social Capital

When Putnam’s (1993; 2000) major works are examined with a particular focus on the relationship between the family, as a specific form of bonding social capital, and more public-oriented bridging or civic social capital, certain inconsistencies are detectable. In his study on social capital in the United States, Putnam indicated a positive role of family ties on volunteering, which is considered an aspect of civic social capital (Putnam 2000, 119–20 and 463n16). However, in his earlier study (1993), he presented these forms of social capital as less compatible, and it is this that has received greater attention among scholars who have subsequently confirmed that strength of family ties has negative effects on civic social capital.

In his study of governance in Italian regions, Putnam (1993) found that social capital in southern Italian regions was less developed than in the country’s north, something he ascribed to a widespread ethos of amoral familism in the south – a phenomenon he saw as one of the main hindrances to developing social capital. Putnam did not offer a systematic operationalization of the concept of amoral familism, instead referring to Edward Banfield’s (1958) book The Moral Basis of a Backward Society which first introduced the concept. Banfield’s study was based on predominantly qualitative data and, as with Putnam’s work, also lacked a systematic quantitative examination of the relationship between amoral familism and social capital. Nonetheless, Banfield provided elaborate and detailed explanations of how amoral familism affected certain phenomena that were later commonly incorporated into the concept of social capital. He studied a small southern Italian rural community and found that its inhabitants acted toward extrafamilial members “as if they were following this rule: Maximize the material, short-run advantage of the nuclear family; assume that all others will do likewise” (Banfield 1958, 85).

Such an ethos can be broken down into two components, one of which relates to attitudes on the trustworthiness of people outside the family circle, while the other concerns the motivational sphere of the individual (Ljunge 2015). Regarding the first component, people in the community under study saw relations with people outside their small tight-knit family circle as “potentially dangerous” (Banfield 1958, 121). A multitude of studies later examined the relationship between the strength of family ties and generalized trust, confirming an erosive effect of strong family ties on that particular aspect of social capital (Ang and Fredriksson 2017; Durante 2010; Ermisch and Gambetta 2010; Marè et al. 2020; Alesina and Giuliano 2011; 2014; see also Alesina and Giuliano 2015). Ultimately, all these studies echo Putnam’s Italian study (1993, 111–12), in which, with reference to Banfield, he detailed how in the southern regions, where an ethos of amoral familism prevails, people express lower levels of generalized trust compared to those in the north.

In Banfield’s study, the second component of the amoral familist ethos relates to the domain of actions and manifests in several social repercussions apparent in community life. First, the community is not well networked or cohesive. The scarce connections that people maintained with those outside their family often took the form of “patron-client” relations, again primarily oriented toward providing benefits for one’s own family (Banfield 1958, 100). The reason for this, apart from the aforementioned paucity of generalized trust, was the extreme poverty of the community under study. In such a milieu, socializing with people outside the family circle was seen as “potentially costly” or even as an unaffordable luxury (Banfield 1958, 121–3, 143–54). These factors, along with the general lack of moral responsibility toward people outside the family, prevented members of the community from establishing and joining civic organizations whose work was focused on public issues (Banfield 1958, 76–8). Consequently, civic associations were almost nonexistent in such communities (Banfield 1958, 16–7, 23–8). Putnam (1993, 91–2) confirmed that the southern Italian regions with their widespread ethos of amoral familism had a less-developed culture of civic associations compared to the north, while Kim and Wilcox (2013), in their study of the United States, also found a negative effect of familism on people’s participation in secular civic associations.

Individualized forms of civic engagement, such as acts of philanthropy and charitable giving, were also not customary in the poor community studied by Banfield (1958, 85–6). A commonplace belief here was the idea that the goods available to people were limited, scarce, and of a permanently fixed quantity. Such a belief system facilitates the impression that benefits provided to others will automatically be at the expense of one’s own family. In other words, in this context, charity was perceived as a luxury (Banfield 1958, 115–6).

Banfield and Putnam had different views on how amoral familism influenced political participation. Banfield (1958, 23, 86–91) noticed that amoral familists do not tend to participate in political action. The inhabitants of the community under study were not particularly interested in taking part in the elections, and this was the case despite the potential benefits of vote selling. There are three reasons why it was impossible to uphold a political machinery in such a social setting. First, because of the secrecy of the ballot, the political patron could not be sure that the “amoral voter” would actually vote as he had been paid to. Second, such machinery lacked the immediate material benefits that might have attracted voters. Lastly, the pathological distrust toward people outside the family made it very difficult to maintain any kind of permanent organization, including an organized clientelist voting system (Banfield 1958, 103).

Conversely, Putnam (1993, 93) saw voting as being twofold in nature, namely it could be an act of civic-minded individuals addressing broader public issues, but it could also be a tool enabling personal gains within clientelistic networks. Here he observed differences between the southern and northern Italian regions with regards to two kinds of elections. Voting in a referendum, understood as attracting profoundly civic-minded individuals, was more common in the north. On the other hand, preferential voting in general elections, symptomatic of clientelistic politics, was more common in the southern regions with their familist ethos (Putnam 1993, 94–6).

Notwithstanding these important differences in the ways Banfield and Putnam saw the relationship between amoral familism and voting patterns, several studies have shown that the strength of family ties negatively affects the development of political culture in general. Strong family ties have been found to have a negative impact on a very broad range of political attitudes and behaviors, including a general interest in politics, volunteering for political parties, and other forms of both conventional and protest-motivated political participation (Foschi and Lauriola 2016; Alesina and Giuliano 2011; 2014; Daniele and Geys 2016).

Research Design and Data Sources

The insights presented in this study were drawn from the data collected for a larger mixed-methods PhD research project, whose main goal was to examine the determinants of social capital accumulation in the Croatian city of Zadar. Of particular relevance for the present article was the exploratory sequential part of the overall research project. The main aims of the first, qualitative phase of this part of the project were to detect concepts that could help explain differences in the levels of social capital, inform the questionnaire required for the quantitative data collection, and develop hypotheses to be tested in the subsequent quantitative phase of the research (Bryman 2006). The analyses in the qualitative part were conducted in line with Layder’s (1998) adaptive theory, which combines inductive and deductive reasoning in terms of how empirical data are connected to sociological theory. In this approach, the data analyses begin with initial coding procedures that are open rather than guided by established concepts and theories. However, later coding and analytical procedures were guided to connect empirical evidence to well-established sociological concepts and theories from the extant research on determinants of social capital.

The hypotheses from the exploratory part of the research were formed as follows. If a clear and uniform pattern, indicating an effect of the independent variable on the dependent variable, is discerned, and no counter-processes are observed in the data, the hypothesis states that the specific variable has an effect. Conversely, if the patterns are not clear, and the data indicate the presence of different processes demonstrating the possible existence of opposing effects of the independent variable on the dependent variable, the hypothesis states no effect of this specific variable. Forming hypotheses in this way and building conclusions based on these hypotheses do not contribute to sociological knowledge by simply supporting the affirmed theory if the hypothesis is rejected. But nor do insights gained in this way contribute to the literature on the addressed topic by simply rejecting a specific theoretical explanation. Rather, this way of forming hypotheses facilitates a critical reflection on the theory, highlighting the existence of internal mechanisms and providing explanations for why a certain variable does not significantly affect the dependent variable. In other words, the research contributes to an understanding of competing mechanisms that possibly coexist and mutually diminish each other’s effects, rather than simply supporting or disproving extant theoretical explanations.

The aim of the second (quantitative) phase was to test the hypotheses developed in the first (qualitative) phase. One of the concepts recognized during the qualitative phase as worth examining in the quantitative part was the strength of family ties. Accordingly, this article focuses on the relationship between the strength of family ties and civic social capital. For the purpose of this study, additional analyses were performed in line with this focus.

The data for the qualitative research phase were collected through semi-structured interviews with nine participants, sampled on the criterion of maximum variation. The interviews were conducted between December 2015 and April 2016, and each lasted from 45 minutes to two and a half hours. All participants were informed about the goals and the purposes of the research. The confidentiality of the data and anonymity of the participants was guaranteed, and the results are pseudonymized. The data for the quantitative stage of the research were collected between April and November 2018.

The total sample included 348 anonymous respondents aged 18 and above residing in the city of Zadar. In total, 150 self-administered written questionnaires were collected, with 198 being collected online. When it comes to sampling strategy, the research was based on a non-proportional quota sample (Petz et al. 2012), in which three control quotas were gender, age, and education. As indicated in Table 1, in the quantitative part that followed, the achieved sample is very close to the population parameters in terms of gender. The sample had a somewhat higher proportion of younger (18–34) and somewhat smaller proportion of middle-aged (35–64) and older (65+) respondents, as well as a higher proportion of highly educated and smaller proportion of less educated respondents. Additionally, to achieve sample stability, certain residential districts in the city are intentionally over/underrepresented, and their proportions in the sample do not correspond to their proportions in the city’s population. Missing values, ranging from 1.1 % to 1.4 % per variable, were addressed with multiple imputation procedures that have proven advantages over other practices for the treatment of missing values (Cleophas and Zwinderman 2016). The analyses were performed on a pooled dataset generated after 20 iterations. A multicriteria weighting (ranking) strategy was applied to correct for sampling bias and make sample parameters closer to the real population parameters. As a consequence of the nature of the non-proportional quota sample, the biases in the sample outlined here, as well as the treatment of missing values, the results of the quantitative phase cannot be generalized to the overall population of the city of Zadar.

The Qualitative Research Phase

This section presents the results of the qualitative research phase. For each of the four social capital indicators addressed – generalized trust, charitable giving, association membership, and voting in elections – participant statements are presented that illustrate how the strength of family ties might affect those aspects. This is followed by the hypotheses that were tested in the subsequent quantitative research phase, along with explanations as to why the hypotheses were formed in that specific way.

Generalized Trust

The qualitative research phase revealed several descriptions which suggested that the strength of family ties may have a negative effect on generalized trust. First, some participants stated that they exclusively trust their family members. For instance, when Luka, a man in his early thirties, was asked whom he trusts in general, he simply and resolutely replied – “my family”. Vlatka, a woman in her mid-thirties, answered in a similar fashion, exclaiming: “I don’t trust anybody anymore! Except for my mum, husband, and older brother! Not even my friends. I would probably not confide in any of them. I mean, no one one hundred percent.”

Second, several participants believed that parents encourage their children not to trust others. Danijel, a man in his late thirties, expressed just such a view: “I mean, this is how your parents taught you: ‘Son look after yourself, otherwise you are fucked’.” Luka shared this opinion: “Well, how I think people are being raised these days, it is literally: ‘Don’t let anyone screw you up!’ Yup, that’s the family motto: ‘Don’t let anyone screw you up!’” Given that some of the participants offered quite illustrative examples of how the strength of family ties might have a negative effect on trust, and there were no statements indicating the contrary, it is reasonable to formulate the first hypothesis (H1) as follows: Strength of family ties has a significant negative effect on generalized trust.

Civic Engagement

According to the qualitative research phase, strength of family ties does not exhibit a uniform effect on all aspects of civic engagement, even though the participants made several statements based on which it would be plausible to argue for a negative effect on all forms of civic engagement examined. One such statement was Vice’s definition of the term “active citizen”. This particular interviewee made it clear that, for him, looking after your own family was the essence of being an active citizen: “Well, I would say, these days it’s very difficult to explain the term in general. Because today, we are all active, and if we are active, we are active for ourselves [smiles], I mean, for ourselves and for our family” (Vice, male, late forties).

Another participant, Luka, reflected on the challenges of developing a local civic culture with strongly public-oriented and active citizens. In his opinion, the emergence of such a culture was inhibited by the people’s strong orientation toward family and family rituals. Children, he stated, were forced by their parents to devote their time to family members, and this appears to be at the cost of participating in activities with other members of the community outside the immediate family. He illustrated this with a description of his visits back home to Zadar, during his childhood, when he was living in Vienna:

Essentially, we are a tribal people! Slavs! We celebrate family. You can see “revelry” at baptisms, “revelry” at [Catholic] confirmation, “revelry” at weddings. I came home from Vienna for four days, and I wanted to see my friends, but instead, I had to see three of my uncles. And then my uncle says: ‘Come here, let your uncle give you a little kiss.’ Active citizens shouldn’t be kissed by their uncles! I don’t have time for your Slavic revelry. (Luka, male, early thirties)

The aforementioned statements by Danijel and Luka in the context of trust indicate that parents teach their children not to be trustful toward people outside the family. But the same statements also illustrate that parents do not encourage their children to engage in cooperative behavior with members of the community beyond their immediate family. Other examples illustrate that the lack of moral responsibility toward members of the broader society is perceived as something that is generated within the family, and that family upbringing impedes the creation of more widespread altruism and solidarity. Danijel depicted today’s social world as being characterized by a battle over competing resources, a place in which an egocentric worldview dominates. In his words: “Simply everything became a jungle with behavior that was akin to pillage, and if you don’t look after yourself, you’ll perish.” More importantly, he expressed the belief that such an egocentric worldview is formed within the family:

I mean, you are taught by your parents: ‘Son, look after yourself, otherwise you’ll perish!’ And then, how are you going to learn to be an altruist and, I don’t know, give something to the community, especially wholeheartedly, and especially based on some revolutionary ideas or something. (Danijel, male, late thirties)

Echoing this statement, Marko, a man in his late seventies, explicitly stated that people refrain from becoming involved in community affairs, instead focusing on protecting the well-being of family members. His description clearly reflects the attitudes and values that are quite symptomatic of the arena of civic engagement in a community of amoral familism, as described by Banfield: “One person has so many worries that they even become an egoist. They become an egoist! For sure! Because they think only of themselves and their children. They live for themselves. They live for their close family circle and maybe for some of their friends.”

In light of the statements presented thus far, it is plausible to argue that strength of family ties has negative effects on all forms of civic engagement. However, such a claim can be made with confidence only in the context of charitable giving, since the interviews generated no statements that would indicate an opposite effect. The second hypothesis (H2) is therefore: Strength of family ties has a significant negative effect on charitable giving.

In contrast to charitable giving, the relationship between strength of family ties and other modes of civic engagement is more puzzling. The analysis revealed that strength of family ties might have a twofold effect on these types of civic engagement, because, along with the expected negative influence, counter-processes could be observed. In other words, for some people, ties with family members could hinder civic engagement, while for others such ties can stimulate participation in that same form of civic engagement.

Such a dual effect can be observed in the context of membership in civic associations. As already mentioned, social environments with a widespread ethos of amoral familism are not characterized by a vibrant culture of civic associations; in the community that Banfield (1958) studied, such organizations were in fact almost nonexistent. However, the Zadar of today does not have a lot in common with that rural community of more than half a century ago. The fact that a large number of civic associations exist in Zadar does not, in itself, reveal much about the relationship between strength of family ties and membership in civic associations. More revealing when it comes to that relationship are statements made by several participants about the nature of these organizations and the reasons people join them.

In their respective studies, Banfield (1958) and Putnam (1993) generally saw civic associations as organizations that brought together civic-minded individuals and correspondingly saw the lack of such associations as a consequence of the absence of a civic-minded citizenry, or a lack of capacity to create and maintain such social networks. The participants in this study, in contrast, revealed that people’s motives for joining such civic organizations were, in some instances, far from purely civic. Some individuals suggest, for instance, that, one of the reasons was to achieve material prosperity and security. When detailing the possible motives, Lidija, a woman in her late forties, mentioned “material gains”, while Danijel stated that in his experience, “for some people, these organizations serve as a tool to secure private interests”. He added that the reasons individuals join these associations vary, and that some people are motivated by their need to “secure a salary through a certain project”. Such processes were also confirmed by Frane, a man in his mid-thirties, who said that the work he does for a civic organization is a “source of financing” for his “current life, paying bills, rent, and everything else”. Later in the interview, Frane further clarified why he had joined the civic association where he was currently employed: “Well, actually, I think in the beginning it was purely employment, right? There was a job opening. Since my last employment contract had ended after three years, I had to go to the unemployment office, and so I applied here.”

Certain participants saw civic associations as being used by individuals to acquire material benefits in corrupt ways, claiming that some of them were even founded for this purpose. Jelena, a woman in her late thirties, for instance, caustically reflected on one such association, claiming that, although the association was registered as a nongovernmental organization (NGO), it makes quite a profit: “Well, we have, I think [laughs], we have some civic associations, for instance [mentions the name of the association], which is registered as an NGO, and we all know about the activities of that organization, and how much money passes through it on a yearly basis.”

According to Danijel, a common way that individuals instrumentalize civic associations with a view to securing personal material benefits is tax evasion. He explained as follows:

Eventually it becomes a reality, and then people start doing it. I want to do something with my life. But I’m not going to start my own company, since that would mean high taxes. ‘Oh, I could do it through an NGO! I’ll launch an NGO and do my business that way.’ (Danijel, male, late thirties)

These statements correspond with findings from other studies demonstrating that the achievement of prosperity and material security are among the reasons why individuals participate in civic associations (Grødeland and Aasland 2011; Rose-Ackerman 2001; Paturyan and Gevorgyan 2014). Moreover, such material goals can be achieved by corrupt means, and even Putnam (2000, 81), in his study of social capital in the US, acknowledged that civic organizations can be “plagued by […] corruption”. In Europe’s postsocialist societies, such practices are common. Civic associations sprung up in greater numbers during the transformation from state socialism to neoliberal capitalism, which was strongly characterized by a widespread culture of clientelism and corruption (Gagyi and Ivancheva 2019; Grødeland 2006; Grødeland and Aasland 2011; Griesshaber and Geys 2012). Thus, these organizations do not necessarily attract only civic-minded and public-oriented citizens, but also individuals who want to achieve personal material gains and potentially those who want to secure material benefits for their family members.

At the same time, another process can be observed whereby ties with family members might have a positive effect on people’s inclination to join civic associations, and thus further mitigate the aforementioned negative impact of strength of family ties on this form of civic engagement. Stipe, for instance, explained that one of the reasons he joined and in fact ran the sports civic association was to involve young children in the world of sports, going on to clarify that in this way he enables his son and his son’s classmates to socialize through sports: “Well, the goal is to include youngsters, let’s say, very young people in the world of sports. […]. My son’s class, who are in their second year of elementary school, are now participating [in the sports gym] for the second year running, and that’s something I organized for them” (Stipe, male, late forties).

Vlatka, too, described a situation which indicates that ties to family members play an important role in recruiting people to join civic associations. She explained why she joined an organization that promotes urban musical culture: “Well, I joined [the organization] because my brother became a member a year earlier. Somehow, I started occasionally going to concerts with him, and eventually I decided to join” (Vlatka, female, mid-thirties).

In response to the theses deduced from the concept of amoral familism, scholars have developed a theory of moral familism. They use this theory to argue and indeed prove that, as illustrated in the examples above, relations with family members can in fact strengthen extrafamilial social capital (Prandini 2014; Donati and Prandini 2007; Żukowski et al. 2009), and in particular, play a positive role in encouraging individuals to join civic organizations (Donati and Prandini 2007; Paturyan and Gevorgyan 2014).

If these processes are reinterpreted with regards to different types of social capital, two opposing processes are discernable. The first is family ties as a form of bonding social capital preventing individuals from joining civic associations and forming bridging social capital. The second indicates a complementary and even positive spillover between the two types of social capital. Thus, while according to the interviewees’ statements, and in line with Banfield (1958) and Putnam (1993), it is reasonable to expect that strength of family ties prevents individuals from joining civic associations, the qualitative analysis revealed that concurrently, counter-processes might occur in which family ties in fact enhance the inclination of citizens to join these organizations. Given this observable dual effect, hypothesis no. 3 (H3) is formulated as follows: Strength of family ties does not have a significant effect on membership in civic associations.

Certain similarities can be detected between the dynamics related to membership in civic associations and to the last aspect of civic social capital – voting in elections. Several statements from the research participants indicate that strong orientation toward family members might generally inhibit actions oriented toward the well-being of the broader community and therefore might also exhibit a negative influence on political participation in the form of voting in elections. However, based on the opinions expressed by other interviewees, it is reasonable to assume that citizens do not necessarily participate in community political life with the aim of improving the general well-being of society at large, but rather to secure material benefits. Marko, for instance, sees citizens as reluctant to participate in political life, unless political participation enables them to improve their own material well-being:

You won’t find many of them engaging in politics […]. He isn’t interested in politics. He will only be interested if he is going to work for a municipality and therefore have some kind of benefit, to work and earn a salary […]. Nowadays everybody is struggling for material gains, to secure material benefits, to have a better life. (Marko, male, late seventies)

Other participants expressed similar views, most notably Lidija and Stipe, who both held that people often entered the political sphere to achieve material gains:

However, when I think about political authorities, any kind of political communities, be they local or national […] I believe that people participate in such communities for the wrong reasons. I mean, I don’t think all of them are like that, but a large share of people engage in politics for their own gains and interests. (Lidija, female, late forties)

Because the majority of the people enter politics to achieve personal interests, and not to improve community affairs, right? (Stipe, male, late forties)

Contemporary research has shown that broad clientelist and corrupt networks, which were nonexistent in the rural community studied by Banfield (1958), are often an indispensable part of the general societal frame in many of today’s societies. They are certainly a vital feature of Croatian political culture (Vuković 2019; Šimić Banović 2018; 2019), a political culture in which selling votes for personal benefits has a long-established tradition (e.g., Županov 2002, Vuković 2019). Of particular relevance here is a category that has long been used in Croatian public discourse but was only recently recognized and formally defined in the social science literature, and that is the concept of uhljeb. The term uhljeb, or “postsocialist homo Croaticus”, according to Šimić Banović (2018, 373–4), denotes “a person who has become a public sector employee through a nepotistic relationship or political party affiliation, normally without the required skills or qualifications for that position”, or a person employed “at a ministry, state agency, local administration body or state-owned company” by means of nepotism or party affiliation. The term is also used to designate a person from the private sector, working in a crony postsocialist capitalist economy, who utilizes “his network and party memberships” to land contracts with state institutions. The very root of the word uhljeb derives from the “pre-Slavic word for bread (xlěbъ) [hleb]”, since the person who uses networks with political actors “‘feeds’ his family while also benefiting the patron”.

Taking this specific form of social contract into consideration, in which social actors receive material security and prosperity in exchange for loyalty to the political patrons, it is possible that people with very strong family ties are divided into two distinct subtypes with regards to voting in elections. The first subtype, which can be termed “familist voting abstainers”, includes people who, owing to their strong orientation toward family well-being and lack of interest in the prosperity of the broader community, abstain from voting. The second subtype, which can be designated as “amoral familist voters”, are inclined to cast a ballot, not with the aim of improving the well-being of the community but rather to instrumentalize their right to vote for their own well-being and that of their family. It is reasonable to believe that many uhljebs belong to this category.

If these processes are placed in the framework of the different types of social capital, the interviews indicate intricate dynamics among all three types of social capital, along with two opposing processes regarding the relationship between family ties and voting in elections. The first, similar to the case of civic associations above, indicates a trade-off between bonding family ties and the bridging/civic form of social capital. In the second, individuals who act in the interests of their family, as a form of bonding social capital, simultaneously embed their act of voting within broader clientelist networks, which can be understood as a type of linking social capital. Nonetheless, their actions, in the form of voting, are also part of a broader public and civic sphere, and thus impact and facilitate as well as represent bridging or civic social capital. Therefore, strength of family ties can have a twofold effect on voting in elections. For some individuals, it can inhibit voting in elections, while potentially serving as a stimulus for others. It is therefore plausible to formulate hypothesis no. 4 (H4) as: Strength of family ties does not have a significant effect on voting in elections.

Quantitative Research Phase

Analyses and Quantitative Results

To test the hypotheses from the qualitative research phase, four regression analyses were performed. The indicators of civic social capital were used as the outcome variables, while the strength of family ties was included in the analyses as the key independent variable, along with six other independent variables. A binary logistic regression analysis was performed for the binary outcome variables – generalized trust, charitable giving, and turnout in local elections – while ordinal logistic regression was performed to test the hypothesis related to association membership.

Table 2 presents the results for the indicator of generalized trust. Three variables prove important in predicting the outcome of this particular indicator of civic social capital. First, it is predicted by the level of educational attainment. Those with high school education, as well as those with a university degree or above, have higher levels of trust when compared to those with elementary education or less. Religiosity also plays an important role in predicting trust. Those who consider religion to be very important in their lives score significantly higher on this indicator compared to those who express the lowest level of importance of religion. Lastly, trust is predicted by the residential area. Those who live in the peripheral zone of detached family houses report significantly higher levels of trust than residents of the central city area, composed of multistorey buildings. According to the results of the regression analysis, the strength of family ties does not have a significant effect in predicting generalized trust. This does not support hypothesis no. 1 (H1) drawn from the qualitative phase of the research.

The second indicator of civic social capital is association membership. According to the results of the ordinal logistic regression presented in Table 3, falling into a higher category on this indicator is significantly predicted by three variables.

Those who belong to the third age group (65 years and above) score significantly lower on this indicator than those in the youngest age group (18–34 years). Those with the highest level of religiosity score higher compared to those who expressed the lowest level. Living in certain residential areas also plays an important role for membership in civic associations. Those who live in the peripheral zone of detached family houses score lower when compared to the residents of apartment blocks in the central city district. The strength of family ties does not significantly affect association membership, and such results are in line with hypothesis no. 2 (H2) derived in the qualitative research phase.

The three variables also predict the third indicator of civic social capital – charitable giving (Table 4). Respondents who reported religion as being the highest level of importance in their daily lives are significantly more inclined to donate money to charity compared to those who express the lowest level of religiosity. Income also has an important role in predicting acts of philanthropy, and the two higher-income groups are more likely to score higher on this indicator of social capital when compared to the lowest income group. Strength of family ties, in accordance with hypothesis no. 3 (H3) derived from the qualitative research phase, significantly and negatively predicts charitable giving.

The last indicator of civic social capital used is voting in local elections. According to the results presented in Table 5, the first variable that predicts this indicator is gender, as females are significantly less likely to vote than males.

Those who belong to the two older age groups are more likely to vote compared to the youngest age group. Religiosity is also a significant positive predictor of voter turnout, and in this case all three groups that express higher religiosity are more inclined to vote than those who consider religion to be unimportant in their daily lives. Lastly, as expected in the exploratory part of the research (H4), the strength of family ties has no significant impact on this indicator of social capital.

Conclusion and Discussion

The results presented in this article are in stark contrast with the majority of existing studies, which have documented that a strong orientation toward the well-being of one’s own family negatively affects the formation of civic social capital. The lack of effect of the strength of family ties on most of the indicators of civic social capital used here can best be interpreted by considering the nature of the institutions that operate within a given local context, but also the nature of some of the constituents of civic or bridging social capital.

Perhaps the best way of explaining the lack of effect of strength of family ties on generalized trust is to consider how the nature of institutions affects the dynamics between those two phenomena. The participants in the qualitative research phase portrayed institutions as highly corrupt and, given the findings of numerous studies on Croatian political culture (e.g., Vuković 2019; Županov 2002; Šimić Banović 2018, 2019), their views are often not far from the truth. The research conducted by Herreros (2015) is particularly valuable in this context, as he examined the relationship between trust in family members and generalized trust. He found that in states where government effectiveness is high, trust in family members has a positive effect on generalized trust, while intrafamilial trust has no effect on generalized trust in states with low government effectiveness. The low efficacy and corruption of state institutions resulted in a progressive decline in the number of actors deemed trustworthy within broader society. In these circumstances, people find it irrational to trust extrafamilial social actors, an attitude that is adopted regardless of the differences in the amount of trust they have in members of their own family (Herreros 2015). It is plausible to assume that the mechanism that is behind the lack of effect of trust in family members on generalized trust in states with stronger corruption and weak institutions also applies to other cognitive categories associated with relationships between family members. In other words, in a social milieu where state institutions are corrupt, while people often encounter untrustworthy actors outside their family, people find it irrational to trust people in general, and such an attitude is adopted irrespective of the differences in the strength of ties that exist within the family.

The results regarding the three indicators of civic engagement show that strong family ties have a negative influence only on charitable giving. Citizens with stronger family ties refrain only from aspects of engagement that do not provide material returns, while in the context of other indicators, the strength of family ties has no substantial effect. Association membership is not significantly predicted by the strength of family ties and, in line with the results of the qualitative research phase, there are two processes that are particularly important in diminishing the negative effect of strong orientation toward family well-being on this aspect of social capital. The first is a process that, in the broadest terms, resonates with the theory of moral familism, according to which strong relations with family members do not hinder extrafamilial social capital, but rather facilitate its formation (Prandini 2014; Donati and Prandini 2007; Żukowski et al. 2009).

As presented in the qualitative section, providing services to family members can serve as a stimulus for participation in civic associations. Additionally, family relations can play an important role in recruiting new members for civic associations. In other words, spillover of family social capital into organizational networks can occur (Donati and Prandini 2007; Paturyan and Gevorgyan 2014). The second process indicates that it is not only civic-minded and altruistic individuals who work for civic associations. For certain people, these organizations can serve as a vehicle for achieving personal material goals (Jacobsson and Korolczuk 2017; Jacobsson 2017; Grødeland and Aasland 2011; Rose Ackerman 2001; Paturyan and Gevorgyan 2014), and these material goals can even be achieved by corrupt means (Griesshaber and Geys 2012). Therefore, some of the people that are strongly oriented toward the material well-being of their own family may abstain from participating in civic associations. On the other hand, there are people who engage in the work of civic associations for the very same reason: working at an association can serve as a tool to achieve material gains for one’s own family. Similar processes were found in Poland – another European postsocialist setting (Jacobsson and Korolczuk 2017; Jacobsson 2017). During socialism, Polish citizens experienced economic hardship and this triggered a strong orientation toward family and small tight-knit informal groups, which in turn provided social and economic support (Jacobsson 2017). These structures of informality proved quite persistent during the transformation from the socialist to the neoliberal sociopolitical setting, and beyond. Moreover, they became a vital component of civic associations. This is why, in Poland, civic engagement should not necessarily be seen as standing in contrast to the domain of family and tight-knit informal groups and/or economic activity of individuals. Indeed, not only did these spheres prove not to be mutually exclusive, they even turn out to be complementary and reciprocal (Jacobsson 2017).

Lastly, also in line with the expectation based on the qualitative analysis, voting in elections is not significantly predicted by strength of family ties. This finding can best be explained by considering the twofold effect of family ties. Some amoral familists – the familist voting abstainers – probably see voting as an act oriented toward the public sphere, a sphere from which they generally withdraw due to their strong orientation toward family well-being. For other familists – the familist amoral voters – voting is an act that secures material gains in a political realm characterized by extensive corrupt and clientelistic networks. These findings are similar to those from Poland (Jacobsson 2017) in terms of how an individual’s concern for family affairs, economic activity, and actions that impact the broader society are structured.

At the same time, however, these findings imply the existence of quite different theoretically relevant processes. In the Polish example, individuals engaged in civic activism and geared their behavior toward the public sphere, while simultaneously achieving benefits for their family and their small, strongly connected informal groups. What distinguishes the processes observed in this research from those in Poland (Jacobsson 2017) is the fact that extensive clientelist networks within a voting system, in contrast, were formed by a much wider range of individuals, largely unknown to one another, who did not represent tight-knit groups. In this sense, the dynamics observed in Zadar when it came to voting reflect a complex interplay between all three types of social capital. On the one hand, the act of voting impacts the broader public sphere and therefore represents a form of bridging or civic social capital. Simultaneously, such an act is embedded in large clientelist networks related to the sphere of institutionalized politics, which can also be understood as a form of linking social capital. At the same time, however, the act of voting can enable benefits for the individual and their family members who are connected as a form of bonding social capital.

The overall results presented in this study demonstrate that the relationship between family ties and the individual’s ties with the broader society should not be understood in terms of a simple trade-off between these two types of social relationship. That said, a strong orientation toward family well-being certainly holds some relevance for the accumulation of civic social capital, as best seen in the context of charitable giving. The study shows, however, that the relationship is more intricate and complex than earlier studies indicated.

When it comes to civic associations and voting in elections, future research on social capital would be well advised to reevaluate these indicators in Europe’s former socialist countries. This is not merely a methodological issue related to the relatively simple and quite common measures of social capital in postsocialist settings. In fact, the complexities revealed in this study show that the direct application of the theory of civic social capital to this specific milieu demands caution. The motives for joining civic associations and voting which are undoubtedly part of the public sphere are often far from civic. What remains unclear is whether people’s actions capture the basic premises of social capital theory and whether they reflect their public-minded spirit.

To address these methodological complexities future empirical research should use structural equation modeling or other research strategies that examine additional mediating or moderating variables. Moreover, it might even be useful to rethink the indicators used to measure civic social capital, since local contexts differ from a sociohistorical and cultural perspective, meaning that some commonly used indicators can be a sign of both civic mindedness and orientation toward personal or family benefits.

To conclude, there are some limitations to this study that should be mentioned. First, the time delay between the two phases of data collection may have compromised the triangulation of the respective findings. The quantitative results may have been influenced by factors and social dynamics on the local, national, and global levels that occurred between the data collection phases. Therefore, the explanations that were based on the qualitative phase and used in the quantitative part of the research might not be fully plausible. Future research could verify this by examining the stability of the relationship between family ties and civic social capital in a longitudinal perspective. Such an approach would reveal whether the patterns observed in this study persist over time. Furthermore, the theoretical models used in the quantitative part should include variables that have proven relevant in other theories on the accumulation of civic social capital.

The relationship between family ties and civic social capital might also be influenced by economic ruptures, large scale migration, armed conflicts, natural disasters, pandemics, or changing institutional policies, for example. The inclusion of such real-life factors is relevant for any study in terms of the implications for contemporary policymaking. Such factors must be taken into account, as the processes observed in the qualitative part of a study might be of less importance than other factors that have gained relevance in the meantime. Additionally, future research should include comparative perspectives between cities with a socialist past but also those without. Given the multitude of different variables that affect the nature and formation of family ties, it would be useful to examine whether the patterns observed in this study may be specifically because of Zadar’s socialist Yugoslav past, or whether other mechanisms are more important.