شکل گیری سرمایه اجتماعی و دسترسی به اعتبار : مدارک و شواهد از سری لانکا

While previous studies evaluate the impact of social capital on development outcomes, there are very few empirical studies on determinants of social capital formation. We use unique long panel data from Sri Lanka to examine the mechanism of social capital formation in an imperfect credit market. We show that households facing credit constraints reduce investments in social capital. Furthermore, temporal declines in investment persistently reduce general trust, trust in villagers, and trust in business partners. While previous studies argue that social capital improves access to informal credit, we show the reverse causality. Combining these findings suggests a potential poverty trap.

Do the poorest of the poor, who are often excluded from formal credit and insurance market mechanisms, rely more on informal reciprocal arrangements through social capital than do the rich, or are they “too poor” to contribute to and benefit from the effective social safety nets generated by social capital? Using unique data from Sri Lanka, we aim to compare these two competing hypotheses empirically. More specifically, we investigate the nexus between social capital formation and accessibility to an imperfect credit market. Social capital is defined as informal forms of institutions and organizations that are based on social relationships, networks, and associations that create shared knowledge, mutual trust, social norms, and unwritten rules (Durlauf & Fafchamps, 2005). Economists and sociologists recognize the important roles played by social capital in reducing poverty and facilitating rural development (Durlauf and Fafchamps, 2005, Fafchamps, 2004, Grootaert and Van Bastelaer, 2002, Hayami, 2009, Ishise and Sawada, 2009, Knack and Keefer, 1997 and Putnam et al., 1993). Narayan and Pritchett (1999) show that villages with more social capital are more likely to enjoy better public services, adopt advanced agricultural practices, and participate in communal activities, and that these in turn increase individual income. Fafchamps and Minten (2002) find that traders with a stronger social network earn higher profits by reducing transaction costs. Higher social capital also helps solve the enforcement problem in risk-sharing arrangements, because it makes individuals creditworthy (Karlan, 2007 and Paal and Wiseman, 2011) and helps recover from negative calamities quickly (Carter & Castillo, 2005). Although a number of studies investigate these impacts of social capital, very few empirical studies establish a framework for social capital formation (Durlauf, 2002, Durlauf and Fafchamps, 2005, Miguel et al., 2006 and Möbius, 2001). Moreover, related literature typically employs cross-sectional variations to identify determinants of social capital and uses time-invariant characteristics of households and communities, such as ethnicity and demographics (Alesina and La Ferrara, 2002 and Charles and Kline, 2006). These studies, however, do not discuss how social capital is accumulated over time. As an exception, Glaeser, Laibson, and Sacerdote (2002) show that the patterns of participation in various groups, such as religious organizations and veterans’ organizations, are consistent with investment behavior into social capital in the life-cycle framework. Since Glaeser et al. (2002) employ data from the United States, they assume a perfect credit market. However, in the context of developing countries, it is likely that the credit market is imperfect, and thus analyses that fail to consider heterogeneity in access to credit markets would generate misleading policy implications. This study attempts to bridge this research gap by setting two goals. The first goal is to examine the impact on social capital investment of changes in opportunity cost caused by credit constraints. Such an analysis would be important because the opportunity costs of investments change over time (Glaeser et al., 2002) and small changes in these costs can have large effects on the equilibrium level of social capital (Möbius, 2001). To capture the degree of social capital investment, we follow Anderson, Mellor, and Milyo (2004) and employ various participation measures of social capital, such as expenditure for ceremonies, participation in community work, and participation in irrigation maintenance. The impact of credit constraints on the contribution to these activities is an empirical question.1 On the one hand, credit-constrained households facing negative income shocks may cope with them by decreasing nonfood consumption (Behrman, 1988 and Behrman and Deolalikar, 1990), and increasing labor supply (Heckman and MaCurdy, 1980, Jacoby and Skoufias, 1997, Kochar, 1999, Kochar, 2004, Morduch, 1995 and Rose, 2001). Therefore, credit-constrained individuals might spend fewer resources on social capital investments. If the declines in investment worsen bilateral relationships with other individuals in the community, it might cause a poverty trap, unless they can access credit outside the community. Lack of access to the credit market results in poor social capital, and this in turn leads to poor access to the informal credit market, as shown by Karlan (2007), thus exacerbating the credit constraints. As Möbius (2001) claims, this possibility of multiple equilibria underlines the importance of investigating the process of social capital formation. On the other hand, households with poor access to a formal credit market may constantly invest in social capital to secure access to informal credit sources, because social capital improves credit market accessibility through social enforcement and social collateral mechanisms (Besley and Coate, 1995, Karlan, 2007 and Karlan et al., 2009). Indeed, many studies have shown the importance of social networks in making available informal credit and other types of mutual insurance (Carter and Castillo, 2005, Fafchamps and Gubert, 2007, Fafchamps and Lund, 2003, Ligon et al., 2002 and Murgai et al., 2003). The second goal of this study is to investigate the persistent effects of temporal change in social capital investments—caused by credit and/or insurance market imperfections—on the level of social capital stock. Previous studies have shown the persistent impact of credit and insurance market imperfections on physical and human assets (Banerjee et al., 2010, Dercon, 2004, Hoddinott, 2006 and Quisumbing, 2006). This study is one of the first attempts to examine the long-term impacts on social capital. For this purpose, we estimate the impact of previous social capital investments on five types of stocks: general trust, trust in villagers, trust in business partners, availability of mutual assistance, and fairness. One important distinction among them is whether they measure the relationship with the entire community or bilateral relationships within the community. While trust in villagers and business partners could involve relationships with particular individuals within the community, the other three indicate general attitudes to others. Contribution to community activities may play various roles in enhancing social capital. First, it gives the community members the impression that the contributor is cooperative and wealthy enough to devote his/her resources to the community. This, in turn, establishes the reputation for trustworthiness and altruism (Posner, 1980). Second, it generates a dense social network among individuals and increases the possible punishment against selfish behavior. To preview our empirical results, we show that credit-constrained households tend to reduce their resource allocation for social capital investment, such as expenditure for ceremonies and participation in irrigation maintenance. Moreover, the negative impact of the temporal decline in social capital investment persists: those who invested in social capital less are less likely to build bilateral and general trust even five years later. The impact of social capital on the perception of fairness and availability of assistance are rather negative and ambiguous, respectively. Anderson, Mellor, et al. (2004) find correlations between the contribution in the public goods game and various trust indicators, although they are not unidirectional. In contrast to Anderson, Mellor, et al. (2004), we study the nexus between previous contributions to public goods in the real world and social capital, showing the unidirectional impact across the types of social capital. The rest of the paper is organized as follows. Section 2 first describes the study site and then discusses the dataset. Section 3 examines the short-term impacts of credit constraints on social capital investment, while Section 4 considers the persistent impacts on social capital stock. Finally, Section 5 concludes the paper.

This study examined the process of social capital formation in the framework of an imperfect credit market. If the credit market is less developed, negative shocks significantly change household behavior such as time allocated to various activities. We showed that credit-constrained households had to reduce their resource allocation for social capital investment. It was also found that households that had previously contributed to public goods suffered less from a lower level of trust. These findings imply that the persistent effect of temporal declines in social capital investment due to credit/insurance market imperfections may be a cause of the heterogeneous level of social capital among community members. While previous studies argue that those who are less trustworthy are more likely to suffer from poor access to informal credit (Karlan, 2007), this study shows the opposite direction of causality: poor credit accessibility causes lower trust. The combination of these findings and the positive correlation between trust and trustworthiness (Glaeser, Laibson, Scheinkman, & Soutter, 1999), together may suggest the possibility of a poverty trap: credit constraints cause households to make low investments in social capital; since households with poor social capital stock suffer from low trust toward business partners, villagers, and other individuals, this possibly leads to poor access to informal credit. Thus, households that have been credit-constrained could be even more vulnerable to risks, because the availability of credit can function as effective insurance against shocks (Eswaran & Kotwal, 1989). This negative cycle could spill over into the broader community in at least two ways. First, the absence of a member from a community meeting may reduce the benefits derived from the meeting for the participants. This, in turn, could be a disincentive for other villagers to attend the meetings. Second, this study showed that households facing constraints are less likely to participate in the maintenance of communal canals. This directly diminishes the productivity of irrigation infrastructure, again causing credit constraints. A possible policy implication derived from these arguments may be as follows: the improvement of the credit market is essential to reduce poverty and enhance social capital. In particular, if there indeed exist multiple equilibria including a “high” equilibrium and “low” equilibrium—that is, a poverty trap—an intensive intervention or a “big push” to relax binding credit constraints would be required.