1 Introduction

Social relationships have been studied for a long time within the field of sociology (Durkheim 1897; Tönnies 2011). However, it was the (re)introduction of the concept social capital by Bourdieu (1986) and Coleman (1988), and the subsequent studies by Putnam on the subject (1993; 2000), which brought about increased attention. Social capital became a frequently addressed factor in various types of literature such as entrepreneurship, institutional economics, labor economics, education, public health.

The majority of the studies describe social capital as value (economic or social; for the individual and/or the group) that stems from resources made available through social relationships or access to social networks (Coleman, 1990; Granovetter 1992). Social capital has mainly been synthesised into three dimensions: cognitive, relational, and structural (Nahapiet and Ghoshal 1998). The cognitive dimension stands for shared meaning and understanding between actors; the relational dimension represents trust, friendship, respect, and reciprocity developed through previous interactions, and the structural dimension refers to patterns of relationships between actors (Villenaa et al. 2011).

The importance of social capital for economic performance and development is associated with a varied list of actions, from political participation to civil society involvement. The power of social capital to enable individuals and communities of different types to get engaged in cooperative and high-trust networks is undeniable. This aspect of social capital is especially crucial for local development, as the high-trust, and cooperative networks act a reducing factor for transaction costs, mitigating regional and urban–rural disparities, in which a lack of material resources is compensated for with more immaterial and value-based cultural assets. Economies with greater social capital are argued to be more prosperous and to enjoy fast growth (Knack and Keefer 1997). High levels of interpersonal trust in thick social capital geographies are argued to allow communities to perform notably better in providing investment possibilities despite the lack of proper financial institutes, and to boost innovation and economic activities in general (Sack 1997; Knack and Keefer 1997).

Social capital is also embedded in space, which brings about a wide range of desirable outcomes related to entrepreneurial milieux at varying aggregates of geography. For example, both conceptual discussions and empirical evidence for the strong link between local social capital and entrepreneurship are available in the previous literature (Westlund and Bolton 2003; Scott 2006; Andersson and Larsson, 2016). In fact, the advantages that are enjoyed by individual actors due to their connection to such social organizations and groups are addressed by a significant body of literature.

While emphasizing the ‘bright side’ of social capital, the existing literature pays disproportionately little attention to the possible detrimental effects associated with certain aspects of social capital. It may not be reasonable to establish a dichotomy where social capital is labeled in a binary fashion as good or bad. However, there is a need to address how social capital manifests itself in highly context-dependent ways.

Against this background, this paper aims at reviewing the neglected ‘dark side’ of social capital and its potential detrimental aspects to highlight the contextual nature of social capital by way of examples from different geographies. This paper is by no means a complete review of the literature on social capital, but a humble attempt at highlighting the positive bias in its treatment in the literature. We focus on two aspects in particular. First, we address why it is essential to take the contextual nature of the social capital into account while investigating its implications for a variety of social and economic outcomes. Second, we are explicit in asserting the importance of geography in determining the context, as it is a necessary dimension that dictates to what extent social capital may be beneficial or detrimental for individuals. We also highlight three spatial considerations for the contextual nature of social capital and fairer treatment of the concept in a spatial context. These are: (1) the importance of the spatial scale at which social capital is studied; (2) the specific locality within the urban–rural continuum where social capital is considered; and (3) finally, path dependencies and the relevance of local social capital for positive and negative lock-in effects.

Section 2 of this paper addresses the conceptualisation of social capital and juxtaposes good and bad social capital with reference to the relevant literature. Section 3 offers a taxonomy of the dark side of social capital. Section 4 focuses on the spatial considerations to mitigate bias, and Sect. 5 concludes.

2 Social capital—the good and the bad

When one thinks about the term ‘social capital’, the first aspect that comes to one’s mind is a group structure. The group can be in the form of a network or hosted within an institutional domain, and it almost always entails some form of membership. Membership in such groups can be formal or informal (e.g. partisan organizations vs ethnic enclaves), and voluntary or involuntary (e.g. club membership vs kinship). Such groups can take the form of a family, a clan, a collective, and even a city or neighborhood. Possibly the second most common aspect of social capital that comes to one’s mind is, then, that such groups deliver benefits to their members that outsiders cannot enjoy or are largely restricted in enjoying.

Regarding the empirical literature on social capital, a meta-analysis by Westlund and Adam (2010) argues that the studies investigating the effect of social capital on economic performance find varying results, mainly due to the use of varying geographical aggregations, a lack of measurements, and the lack of the use of common research instruments. Among a series of studies that investigates social capital through an economic lens, Glaeser et al. (2002, p 437) listed a series of interesting findings, and highlighted the similarities between human capital and social capital:

  1. i.

    Social capital first rises and then falls with age;

  2. ii.

    Social capital declines with expected mobility;

  3. iii.

    Social capital rises in occupations with greater returns on social skills,

  4. iv.

    Social capital is higher among homeowners;

  5. v.

    Social connections fall sharply with physical distance; and

  6. vi.

    People who invest in human capital also invest in social capital.

In the absence of high-quality institutions and insurance, social capital acts as a safety net, especially in developing countries (e.g., Rosenzweig 1988; Fafchamps 1992; Townsend 1994; Fafchamps and Lund 2003). Social contracts do not always emerge based on voluntary participation (e.g. kinship), but they may still impose obligations on their members. For example, in analyzing the economic outcome of such types of social settings, Falko and Bulte (2011) have argued that extensive kinship networks are associated with lower income levels and evasive economic behavior. The presence and strength of certain types of social bonds are also seen to facilitate corruption or to lead to an inefficient allocation of resources (Callahan 2005; van Deth and Zmerli 2010). Examining the relationship between social capital and crime, Ostrom (1997) has argued: ‘Cartels and organized crime are networks of relationships that lower overall productivity while generating disproportionate benefits for a few beneficiaries’ (p.162). Even Putnam (1993), having popularised the concept of social capital, acknowledged that social capital does not necessarily always benefit the democratic processes: ‘Not all associations of the like-minded are committed to democratic goals nor organized in an egalitarian fashion’ (221). Putzel (1997) went a step further in his criticism, arguing that the work of social capital is influenced by the idealisation of social ties and communitarian thinking. The critical points of view in the literature, each of which accentuates the context-dependent nature of social capital are numerous.

Despite its ever-increasing popularity, there is still a lack of agreement on how to define and measure social capital. Each researcher and research domain promotes different aspects of social capital. Adler and Kwon (2002) have suggested leaving the definition and measurement of social capital to the researchers, ‘whether they focus on the substance, the sources, or the effects of social capital’ (p. 19).

The role of social capital for economic and political development was accentuated in the earlier works of established researchers such as North (1990) and Polanyi (1944). Similarly, Reason and the Rationalization of Society by HabermasFootnote 1 (1984) provided arguments highlighting the importance of communicative action for arriving at a consensus in society. However, it was in the early 1990s that the concept gained matchless attention. The idea of incorporating the social and institutional characteristics of varying geographies induced such attention. For example, in 1990, the World Bank identified social capital as the ‘missing link’ in development strategies.Footnote 2

3 The dark side of social capital

While most literature only considers the benefits, or in other words, the bright side of social capital, a few academics have started to refer to a potential dark side or negative consequences associated with social capital. A special issue of American Behavioral Scientist highlights the specific impact social capital may have instead of discussing the nature and/or characteristics of social capital as being good or bad. In other words, if higher social capital reduces something positively valued (e.g. social equality or tolerance) or increases something negatively valued (e.g. corruption or free-riding), then those links are considered as negative consequences of social capital (van Deth and Zmerli 2010). Such studies that look at the negative consequences refer to the other side as ‘dark sides of social capital’, ‘unsocial capital’ or ‘bad social capital’ (cf. Callahan 2005; Dekker 2004; Fiorina 1999; Hoeber Rudolph 2004; Levi 1996; Portes and Landolt 1996; Putnam 2000). Rostila (2011), for example, has discussed the shortcomings of the theoretical foundations of social capital. In their argument, it seems evident that individual and collective social capital may affect individuals, as well as social structures, negatively. Some of these potential negative consequences are the promotion of negative externalities such as unhealthy norms and behaviors (Christakis 2004; Christakis and Fowler 2007), relational strain (Rook 1984; Due et al. 1999), criminality (Gambetta 1993; Putnam 2000; Browning et al. 2004), free riding, preservation of social exclusion (Portes 1998) and inequality of resources between social groups (Lin 2000; Rostila 2007).

Graeff (2009), building on Coleman (1990), has stated:

As there is no norm (associated with social capital) that facilitates certain actions without restricting others, the social consequences of ‘social capital norms’ are not positive in any situation. Arguably, the assessment of actions as good or bad depends on the cultural settings, the political conditions, and the social aims. But it seems worthwhile to note that the inherent ambiguity of norms due to the fact that they allow actions at the expense of constraining others suggest a ‘dark’ side of social capital as well. (p. 146)

There is hardly any consensus in the literature on whether social capital is a concept that deals with individual attributes or collectives (Portes 2000). However, according to Graeff’s argument, norms facilitate social coordination and collective problem solving, while imposing restrictions on other parties’ actions. Similarly, Coleman (1990) has stressed the collective nature of social capital, where an individual relinquishes self-interest in order to build a better collective environment, which ‘generally leads persons to work for the public good’ (p. 311). To stress our point, in this respect, the question is not whether it is feasible or reasonable to code social capital in a binary fashion as good or bad. It is whether there is a need to mitigate the idealistic approach towards the use of the concept, and instead strive for a realistic understanding of the net effects of social capital.

The level of analysis is also essential to identifying the negative and positive implications of social capital. A related discussion on the macro (e.g. groups and communities) vs micro (individual) analysis of social capital was offered by Lin (2000). The author argued that inequality in different types of social capital contributes to social inequality, e.g. economic performance and quality of life.

Likewise, de Souza Briggs (1998) introduced the concept of perverse social capital, elaborating on the dark nature of the concept through three arguments:

  1. i.

    Social capital may be beneficial for one individual – however, it may not be for the community or society as a whole;

  2. ii.

    Social capital ‘has no right or wrong to it until some judgment is made about the ends to which we put it’; and

  3. iii.

    Social capital is divided ‘along the very faulty lines that relationships, neighborhood, and social participation often are in our world’ (race and ethnicity, social class, gender, and culture) due to its nature’

Verstraete and Van Rossem (2013) have also argued that most of the literature on social capital links it to benefits for either the individual or the collective. This tendency, however, results in a normative bias, as it presumes that collective and individual interests are consistent and can be identified by researchers. They explained this with an example from Mayo (1933), who noted that informal groups could keep productivity below its maximum by using their social capital to keep the pace of work down. Moreover, they stated that groups as mafia clans or hate groups also generate substantial social capital, but that it is hard to consider this as a positive outcome for society. Dasgupta (2005) elaborated that other studies have also included extended kinship, lobbying organizations, and such hierarchical relationships as those associated with patronage (e.g. the Hindu jajmani system and the Sicilian Mafia), as well as street gangs. This implies that dense networks do not necessarily lead to general economic improvement, at least not in the long run (Gambetta 1993). The relationships can involve allocations, leaving some of the parties worse off than they would have been if they had not been stuck in that relationship. In short, even though no overt coercion may be visible, such relationships might be exploitative (Beteille 1983; Dasgupta, 2000; Ogilvie, 2003).

Against this background, we provide a brief review of the dark nature of social capital under five bullet points:

  1. i.

    Negative attributes of bonding and bridging social capital

  2. ii.

    The kinship paradox

  3. iii.

    Corruption and organized crime

  4. iv.

    Negative effects on innovation and organizational efficiency

  5. v.

    Social and political intolerance

3.1 Negative Attributes of Bonding and Bridging Social Capital

Putnam (1995), among others, has distinguished between the two different structures or types of connectedness for social capital: bridging and bonding. Bonding social capital refers to the links between a homogenous group of people with similar objectives, particularised trust, and specific norms of reciprocity (Putnam 2000). It is manifested in local groups, such as guilds, mutual-aid societies, sports clubs, and mothers' groups. Bridging, on the other hand, represents the capacity of such groups to make links with other groups that may have different views. Those heterogeneous social networks offer their members the opportunities to socialise regularly with people from different social backgrounds, which may lead to successful cooperation, trustful relationships, and reduced stereotypes. There is also ‘linking social capital’, a relatively novel concept, which describes the ability of groups to engage with external agencies in order to influence their policies or gain useful resources (Pretty 2003).

Bonding social capital primarily refers to social ties that are built based on commonality in some fundamental aspect. An ethnic or religious neighborhood organization can be an example of such social capital. Bridging social capital, on the other hand, refers to the connections that allow for interaction across different identities (e.g., across race, ethnicity, or sector). Although bridging social capital is known to be more fragile, it is argued to foster social inclusion and deliver desired outcomes both socially and economically (Baron et al. 2000).

The dark side of social capital has been widely associated in the context of bonding ties and thick trust, as it enables the creation of strong group identification, leading to the exclusion of outsiders. Putnam (2000) has argued that bonding social capital results in dangers like group closure, antagonism, low tolerance, and criminal networking. While supporting trust and reciprocity, the solidarity within dense networks can also lead to the exclusion of outsiders, limit the freedoms of insiders, and undermine democracy (Portes 1998). In this line of thinking, group boundaries and self-interest are fostered by organizations that pursue individual and build exclusive social capital that may be detrimental to society as a whole (Zmerli 2008).

The negative attributes of social capital can be explained by various factors rooted in the bonding-bridging dichotomy for social capital. To begin with, social capital is enhanced through the involvement of individuals (and their social skills, contacts, and personalised trust), regardless of the criminal or non-democratic orientation of the network itself. Striking examples for the bonding kind can be found among religious groups with sectarian tendencies, status groups focusing on social exclusion, the Mafia, skinheads, the Ku Klux Klan, or the Nation of Islam (cf. Levi 1996). Another factor might be the closure of small-scale organizations, which tends to inhibit any permeation of external inputs as they create bonding social capital. transaction-based trust seems to be challenging to develop in a segmented network, due to a weak capacity to engage in formal and informal associations. It might then hinder local communities, with many small-scale organizations required in order to produce collective goods (Crouch et al. 2001, 2004; McMillan and Woodruff 1999; Pérez-Díaz 2002). Moreover, in inward-looking and isolated social networks, the creation of uncivicness and intolerance is more likely, due to missing bridging (Li et al. 2003; Paxton 2002; van Deth and Zmerli 2010).

According to Browning (2009), under certain conditions, social capital may somewhat hinder the effective social control of crime. For example, firms with employees who have extensive social networks and interpersonal connections generally have stronger collective norms, and those norms may also allow a collusive tolerance of breaking such norms. While most scholars acknowledge the benefits of bridging networks, they also acknowledge the likelihood of those networks containing significant inequalities and tensions. The networks may create unequal benefits and costs for participants which is why some groups may be hesitant to join formal bridging relationships. Moreover, they may create new risks for community leaders as a result of the political climate (Rusch 2010).

Ethnic fractionalisation is another matter often addressed by the literature that tackles the role of social capital for societies. In a society in which there is a great degree of ethnic diversity, we see strong ties within the minorities and various ethnic groups. However, a significant challenge with ethnic diversity, especially in Sub-Saharan and African countries, is that ethnic fractionalisation in a society occasionally leads to ethnic conflicts, which in return delivers undesirable social and economic outcomes. In their foundational work, Easterly and Levine (1997) found that high levels of ethnic diversity are strongly linked to high black-market premiums, poor financial development, low provision of infrastructure, and low levels of education. Even in situations where ethnic diversity does not directly relate to an economic indicator, the scholars argued that it adversely affects the policies that are related to economic growth. In a similar study, Alesina et al. (2003) argued that ethnic and linguistic fractionalisation (even more than religious fractionalisation) relates to the economic success, welfare and policy quality, and quality of institutions across about 190 countries.

3.2 The kinship paradox

When we are examining the positive and negative aspects of social capital, one of the first paradoxical social forms to think about is kinship. The concept of kinship, i.e. the extended family, can be described as a collective institution that represents a fundamental principle of social organization in specific geographies such as sub-Saharan Africa and some Asian countries. Various forms of it can also be seen in some of the Islamic societies. Membership in a kinship network can be by bloodlines, marriage, or adoption. It regulates social relationships and marital customs, as well as access to resources, services and even social status. Kinship relations generally carry moral obligations regarding sharing and redistribution, which are supported by customs and norms. Hence, members of the extended family can claim assistance from their relatives in dealing with social and financial difficulties. The safety net provided by the family structure is known to create significant differences in the economic and social outcomes for individuals in different cultural domains.

The implications of norms for sharing of resources, or for decisions regarding consumption and accumulation of wealth, are essential to the economic conditions of individuals and groups. This is especially true in developing countries, where individuals often have to rely on community-based informal structures that offer them social security and decrease their exposure to risk (Rosenzweig 1988; Fafchamps 1992; Townsend 1994; Udry 1994; Fafchamps and Lund 2003). For example, the relevance of kinship in societies with weak social welfare nets and bad institutional quality can be considered one of the ultimate determinants of an individual’s status and success -holding unobserved abilities constant.

However, the caveat with kinship is that membership is almost never voluntary. Entry is determined by birth or marriage, and in some cases, exit is not optional. The involuntary nature of kinship, therefore, is associated not only with positive outcomes but also with negative ones on occasion. For example, with regards to savings and sharing, Falco and Bulte (2011) have shown that households try to evade their ‘sharing obligations’ by (1) collecting durables that are non-shareable at the expense of durables that may be shared and (2) accumulating less savings in liquid assets. A striking finding in their research is that more extensive kinship networks are associated with lower incomes, as accumulation incentives are diluted. Therefore, sharing within the extended family has been argued to affect income growth negatively (Falco and Bulte 2011).

There exist complementarities and substitutions between formal and informal institutions (Naranyan-Parker 1999). Whether kinship would deliver desirable or undesirable outcomes largely depends on such complementarities and substitutions between the state and its functions and informal structures in society in any given location. We argue that kinship is useful for its members in situations where certain functions of the state fail to deliver desired outcomes for its citizens. The role of the state, then, is complemented with the safety net provided by other members of the family line or community. In this line of thinking, to what extent kinship can act as a complement defines its potential value to its members. For example, in a case where an individual is laid off and no formal framework for unemployment benefits exists, he may still get help from his kin until he is employed, constituting a safety net.

However, substituting the state’s functions with the prevailing rules and norms of kinship is where the problematic nature of social ties manifests itself. It is not uncommon to see situations where the function of the judiciary system in judging an individual’s actions and implementing their punishment is substituted with the informal norms and values of a family or tribe. So-called honor killings in the Islamic communities, or Karo-Kari practices in the rural and tribal parts of Pakistan, are tragic examples of such substitution. Although homicide is unlawful in almost all countries, in such examples, the norms dictate the killing of a female individual by a male member of the family, simply because she is assumed to bring ‘shame and dishonor’ to that family. Such norms are far less dependent on geography than they are on cultural capital. Similar incidents can, for example, be seen among the immigrant groups in Europe (Van Eck 2003). The problem is not limited to the members of such families and tribes. Individuals who share a common characteristic with those tribes and families through their ethnicity or religion are also likely to experience difficulties in their respective societies due to homogenous and unitary depiction of such incidents and norms by third parties and the media.Footnote 3

3.3 Corruption and organized crime

Previous research has contended that the norms of the social order are weakened rather than increased by social capital. The negative relationship between law-abiding attitudes (as a precondition for the stability of democratic societies) and social trust is a potential example. Having trust in other citizens’ lawful behavior is not a reliable predictor of one’s own willingness to follow the law, which is found to be true in a wide range of diverse countries, irrespective of their specific political or societal contexts (Putzel 1997; Zmerli 2010). For example, corrupt deals have a higher probability of appearing in settings where a complete network of trust exists. Otherwise, corrupt agents have to assess additional variables (e.g. the probability of completing the network with a corruptible public official) and be aware of the probability of being detected by monitoring agencies (Uribe 2014).

Similar effects of social capital can be observed when we look at the sources of organized crime. Rubio (1997) suggests that delinquency and criminality are products of a lack of social capital, but more so the result of considered choice. Negative social capital has risen alongside economic development, specifically in social capital-rich geographies. He gives an example of Antioquia (his main case study), where close-knit ties enable robust commercial trade while facilitating an increase in the formation of drug cartels. Cartels and organized crime can be defined as networks of relationships that generate proportionate benefits for a few while they lower overall productivity. Interpersonal networks of street gangs and crime syndicates can be of negative social worth even if they are beneficial to their members (Gambetta 1993). Similarly, a government system based on military command and which uses instruments of force can destroy other forms of social capital while building it’s own (Ostrom 1997). Theories about the relationship between bonding social capital and deviance also presume that strong ties between individuals are responsible for problematic group behavior, organized criminal activity, and gang membership (see, e.g. de Souza Briggs 1997, 2003; Rubio, 1997; Putnam, 2000).

3.4 Negative effects on innovation and organizational efficiency

Besides traditional inputs (R&D investments and human capital endowment) and other more traditional resource relevant characteristics of the local economy, social capital and organizational networks are highlighted as important assets for innovative activities, particularly by allowing for better collaborative work, reducing transaction costs, and also by dissemination of informal knowledge (Castells 1996; Bunnel and Coe 2001). However, it should be noted that only ‘bridging’ social capital—based on weak ties—can be identified as a key driver, while bonding ties by and large seem to have adverse effects on innovation performance (Crescenzi et al. 2013). In a number of papers, Grabher explores in detail how social networks, may operate at the organizational or at the spatial level, can present both opportunities and hindrance (see e.g. Grabher 2001; 2002; 2004; Grabher and Ibert 2006). Similarly, Florida (2002) has argued that places with high social capital generally have the least innovative performance due to network closure and insulation from external information and challenges. Relational social capital, on the other hand, only becomes relevant to the development of radical innovations when such innovations involve tacit knowledge. Otherwise, it may be an unnecessary investment of social resources and can erode the innovative capabilities. Therefore, if knowledge is minimally tacit, weak ties are argued to be preferable over strong ties (Pérez-Luño et al. 2011).

Regarding organizational efficiency, Burt (1992) mentioned cohesive ties as a factor that hinders the coordination of complex organizational tasks. Strong ties to cohesive contacts limit an actor’s autonomy to negotiate his role. Burt (1997) postulated that managers monitor information more efficiently if their contact network is rich in structural holes, and found evidence for them adding value to the organization. This argument is also in line with the findings of Tsai and Ghoshal (1998) that network betweenness has a positive relationship with the ability of an organizational unit to exchange and combine resources, adding value through product innovation. Conversely, network closure hinders organizational coordination (Gargiulo and Benassi 2000).

3.5 Social and political intolerance

In the comparison of social and political tolerance, Iglič (2010) analyzed social and political tolerance in different European countries. She found diverging results in associational involvement concerning political tolerance. In countries with fragile and/or weak democratic institutions under political threat of political extremists, associational involvement seems to matter. While close social ties generally foster social tolerance of various stigmatised groups and ethnic minorities, the extensiveness of associational involvement is negatively related to political tolerance if political stakes are very high. Moreover, close ties can be noticed consistently across the old democracies, but are missing in newer democracies, including Spain (van Deth and Zmerli 2010).

It has been shown that involvement in informal groups may also reinforce negative attitudes. In the case studies in which they were engaged in informal networks, citizens were less satisfied with democracy and felt less capable of affecting the political sphere, as they felt underrepresented (García Albacete 2010). Jan van Deth (2003) tested attitudes among members and nonmembers of voluntary associations in a large number of European democracies. Negative forms of social capital were not found to be widely spread among voluntary associations. However, any association can be understood as potentially having an undesired or unproductive form of social capital. Specifically, religious organizations were mentioned, as their members tended to be less concerned about politics. Network-based transactions may also be problematic due to their exclusive nature, since exclusivity is one of the conditions for a network to exist, and it naturally implies ‘anonymity’. While anonymity is often found in competitive markets, we do not observe it in operations of networks. If networks govern allocation mechanisms, then ‘names’ matter, and transactions are personalised. Such a situation is argued to decrease efficiencies, as resources may not move to their most productive uses (Dasgupta 2005).

Informal help and welfare-state provisions are highly interrelated (Albano and Barbera 2010). The complementarity versus substitution argument comes into play here once again when we think about the relationship between social capital and formal institutions. Ahlerup et al. (2009) have found that the effect of social capital on economic growth dependent on the quality of formal institutions. An interesting finding they have offered is that social capital almost ceases to matter when institutions are strong, and matters the most when formal institutions are weak (Ahlerup et al. 2009). There is some evidence that the willingness of citizens to help each other decreases in relation to available welfare provisions, due to some form of the ‘crowding-out’ effect (Dahlberg et al. 2012). However, accumulations of helpfulness in different institutional contexts have quite different consequences for the legitimacy of the system as a whole. Previous research has argued that people who spend their time helping others in a weak welfare state have a critical attitude towards the political system, which does not hold for the helpful people in stronger welfare states (van Deth and Sonja Zmerli 2010). Regarding one of the most prominent countries when it comes to the scale of the welfare state, Sweden, Dahlberg et al. (2012) have shown that a larger immigrant population leads to less support for redistribution in the form of preferred social benefit levels.

4 Spatial context

There exists a significant body of literature that highlights the importance of social capital as a resource for fostering local development, and at the same how (and which) geographical features lead to the formation of social capital itself. To reiterate, we do not intend to review this literature in its entirety, but rather list three issues related to spatial context that may be of particular importance when studying the potentially dark side of social capital.

4.1 Spatial scale

The first issue deserving careful treatment when assessing the good and the adverse outcomes associated with social capital in a spatial context is the spatial scale and the spatial hierarchies between different levels of geography. The importance of spatial scale for studying social capital has been previously addressed in the literature (see, e.g. Westlund and Larsson 2016). Fractal-like hierarchies across different geographical aggregations are at the heart of regional science literature. When it comes to social capital, likewise, the multilevel nature of geography should be considered when determining the good and the bad attributes of social capital for a particular outcome (e.g. local entrepreneurship rate, probability of employment, democratisation, crime, etc.).

In order to understand whether social capital in a particular area leads to desirable or undesirable social and economic outcomes, it is not sufficient to rely on administrative boundaries, which imposes an additional challenge onto the existing set of challenges with data collection. The geographical extent of social networks is not necessarily constrained by administrative boundaries, and these networks often do not operate on a single geographical level. The interaction between the different levels of geography, for example, the neighborhood and the city, must be understood well when examining the importance of social capital. This particular issue has been highlighted by Brondizio et al. (2009) in the context of resource management and the interdependencies of ecological and social systems. Similar to our afore-mentioned arguments, these scholars have discussed the importance of the interplay between bonding and bridging networks, which often operate at different geographical aggregations (once again, e.g., neighborhood vs city). Actions that are linked to the nature of social capital at one geographical aggregation may generate negative externalities that affect others in other geographies. The flow of such negative externalities may be from the micro-geography to the meso and the macro, or vice versa. Research on social capital may result in overattributing specific desirable outcomes to social capital, or discounting the potential harm that may be caused by a particular type of social capital (i) if the social networks related to the outcome are not observed at the right geographical level, and (ii) when the interdependencies between different social networks (operating at different levels within the same geographical hierarchy) are not understood well. An in-depth discussion on this matter with the example of ethnic enclaves can be found in Klaesson and Öner (2020), where the authors have shown how the same measures for ethnic enclaves and segregation are associated with entirely different labor market outcomes when measured at different spatial scales, changing the interpretation of how productive it is to be located within an ethnic enclave. Hence the potential policy discussion.

4.2 Urban–rural context

The second issue we would like to raise concerning the bright and dark nature of social capital is the commonly used urban–rural dichotomy in regional science literature. There is a considerable amount of work available to those who are interested in the importance of social capital in rural places (see, e.g. Westlund and Kobayashi 2013; Westlund et al. 2014; Li et al. 2019). One of the arguments in this line of literature is that the absence of formal institutions and tangible resources in rural places can be substituted by social capital, and social capital can stimulate local development in those places even though there are limited benefits from agglomeration. However, there is relatively limited work on the net effects of social capital when agglomeration benefits are taken into account. Johansson (2016) has reviewed the relationship between the two elegantly using the concept of social infrastructure and the imperfections social infrastructure (particularly in developing countries) that may lead to allocation of resources away from productive use. By imposing exclusion, such imperfections may also lead to a shortage of competition. The exclusionary nature of social networks is once again particularly prominent in the context of rural places, where these networks are often bonding.

With reference to Tönnies (2011), the thesis put forward in this contribution is that rural places may be able to compensate for lack of diversity by developing patterns of a community culture (Gemeinschaft), whereas urban regions evolve in the direction of system-wide institutions (Gesellschaft) with rich opportunities instead of tight personal bonds between community members. (Johansson, 2016, p. 26)

Johansson’s point is particularly important when considering the absolute benefits associated with social capital in rural places versus the relative benefits. It is true that in most cases, one can observe that a thick social network and a high-trust social infrastructure in rural places, which can be capitalised on for survival against exogenous shocks, develop economic activities compatible with the physical geography for further development. In this way, they partially substitute for the lack of access to large regions and agglomeration benefits associated with them. However, the very existence of these thick network and bonding ties may also be the reason why an area remains (or incrementally becomes) rural when we control for the first nature geographical features. As Johansson (2016) has stated, ’The advantage of an urban agglomeration is not primarily that it is large. The advantage is that there are more things to combine and abundant experiences in the art of introducing novel combinations’ (p. 49).

In short, the type of social capital and the social infrastructure together with which it should be studied differ dramatically along the urban–rural continuum. The variation in its nature is not limited to but is related to, potential negative externalities associated with it, which at times may be hidden by the fact that it serves as a resource in the absence of other (and potentially more desirable) institutional and tangible resources in rural places.

4.3 Path dependence

The third and final matter we would like to raise in relation to the dark side of social capital in a spatial context is the path dependencies that emerge when the importance of social capital for local development is considered. We briefly touched upon this issue above when discussing rural social capital. Path dependencies—their formation as well as potential path-breaking events—are covered extensively by a large body of evolutionary economic geography literature. We would like to pay attention to one particular approach within this strand of literature that deals with lock-in effects. If we depart from the notion that social capital and social infrastructure are spatially sticky, and that they are essential resources for local development, we also need to consider to what extent their effects are temporal and whether their positive effects are linear or not. As Martin and Sunley (2006) have put forward, the basic idea behind lock-in is that there exists an:

observed tendency for the geographical structure of the economy to exhibit historical ‘quasi-fixity’—for urban and regional patterns of economic specialization and uneven development to be self-reinforcing and self-reproducing over quite long spans of time, even extending and lingering into periods when they no longer appear to be wholly consistent with the economic conditions and circumstances of the day, which is, of course, precisely why a path dependence perspective is particularly relevant to an understanding of how an economic landscape has come to be what it is. (p. 414)

This notion is quite similar to the idea of cumulative causation presented by the economists Myrdal and Kaldor, who have argued that a change in an institution will lead to changes in other institutions and then there would be a feedback effect through which this chain reaction would take a circular form. When applied to local areas, such as neighborhoods, cities, or regions, this self-reinforcing and self-reproducing mechanism implies that the failure or success of a local economy is path-dependent, and social capital –alongside other cultural institutions—is one local element related to such path-dependency that changes incrementally or remains fixed mainly over time.

The fact that local social capital is somewhat fixed over a long period has an important implication for local development. Essentially, taking a snapshot of the social infrastructure and local social networks to relate their importance to local development may be misleading. What initially can come across as a positive resource for regional/local development, hence leading to a positive lock-in effect, may have a quadratic form, where at one point the benefits from thick/strong ties in the local network can peak. Thereafter, we may observe that the very same social infrastructure leads to a negative lock-in effect. In high-trust areas with bonding ties, positive externalities may emerge thanks to easing with embeddedness and deliver higher relative economic performance. However, Martin and Sunley (2006) have shown that the very same social infrastructure may at one point start inducing inflexibility that hinders innovation and reduces competition, leading to a negative lock-in effect.

5 Concluding remarks

The domain in which the importance of social capital is debated is no longer limited to the academic world. Today policymakers at the international, national, and local levels often take social capital as one of the resources to be utilised for economic development and the well-being of individuals in societies. For example, in 2009, a European Union report on social capital and EU cohesion policy stated: ‘Communities that have inherited a substantial stock of social capital are more apt to cooperate voluntarily on EU policy initiatives: that is, they are less likely to see a conflict between the pursuit of collective goods and the goal of maximizing individual or familistic group wealth’.

Similarly, OECD’s 2010 report on innovation strategies listed social capital among other factors as an essential challenge for low-income countries and stated: “’Low-income countries face specific challenges for making innovation the source of economic development, such as poor framework conditions and low human and social capital’ (p.22). The problem with such generalisations is that they depart from a simple and inaccurate assumption: social capital is a good asset at all times. We see an overemphasis on the benefits associated with the concept, where the context-dependent nature of social capital or varying types of social capital remains peripheral to the debate. Both empirical and conceptual research, however, argue for treating the concept carefully. Researchers carry the largest share of responsibility, and there is a need for a consensus on the definition of varying types of social capital. Research should aim to identify social capital’s direct effects, preferably in empirical settings, where causal relationships can be established. We hope this paper provides a useful review for the social capital researchers considering the geography-, time-, and context-dependent nature of the concept in their research.

Why is it essential to address the contextual nature of social capital? Issues such as empowerment of disadvantaged minorities or communities and the formation of strong civil society are central to the social capital debate. The concept has been gaining ever-increasing attention for its potential to bring about a wide range of positive economic and political outcomes. Such attention is not only evident in the academic world, but also the policy-making both at the local and national levels. There is a great risk that overemphasising the benefits of social capital, or overlooking the possible shortcomings of it, may lead to policy designs that are misaddressed, inefficient, or even spawn negative results in some instances.

How does geography come into play? To a great extent, social capital can be argued to be spatially sticky. It is place-dependent, and its nature of varies according to the spatial resolution we look at. Likewise, the cultural and institutional sources of the ‘dark side of social capital’ often originate from a limited geographical area, which is sometimes a city, a neighborhood, or a region. Thus, one of the contexts social capital is dependent on is space. Social capital is produced and reproduced in space, but also determines how space itself is used by economic actors. In this review, we strive to articulate the issues related to the dark side of social capital in a spatial context and hope to contribute to existing efforts to build a more balanced narrative around the importance of social capital.