Social Capital and Income Generation in South Africa, 1993-98

The concept of social capital, well grounded in the sociological and anthropological literatures, is increasingly being analyzed and used by economists and other development policy practitioners (e.g., the World Bank now has a series of Web pages devoted to the topic). The entry point for many economists is Robert Putnam's research on Italian regional economic performance¹ and his subsequent work in the United States. For Putnam, "social capital refers to features of social organization such as networks, norms, and social trust that facilitate coordination and cooperation for mutual benefit."² Not all analysts view social capital as a silver bullet, however, and some even caution that it may be a dangerous concept justifying less state involvement when it is truly needed. This research contributes to the debate providing empirical evidence on the role of social capital; the goal of the paper is to determine the strength of the causal relationship, if any, between one proxy for social capital, membership in formal and informal groups, and household welfare in South Africa in the 1990s.

The hypothesized mechanisms by which social capital affects household welfare, as measured by household income, can be summarized as (1) reductions in the costs of transactions by improving information flows about new opportunities and potential shocks, improving the diffusion of innovations, and improving knowledge about the comparative performance of local government agents, (2) the promotion of consultative decisionmaking as well as collective action that minimizes negative externalities and promotes the production of public goods and (3) the fostering of time-sensitive exchanges for mutual benefit by developing norms of civic behavior, trust and reputation dissemination. Moreover, in times of crisis, this latter mechanism may serve as informal insurance.

There are a number of reasons we might expect social capital to be important for income generation in South Africa in the 1990s. First, the notion of social capital has some resonance with the traditional institution of "ubuntu," which means, "I am because you exist," and is seen as an expression of community life and collective responsibility. Years of apartheid, however, appear to have undermined this institution to some extent. Second, the flagship economic policy of the first democratic government explicitly recognizes the important role to be played by local community institutions in the implemen- tation of policy. Lastly, serious crime and politically motivated violence, while still very high post-1994, have declined somewhat in recent years. Since the 1994 national democratic elections, then, it is possible that the returns to (and therefore stocks of) social capital have risen.

Data and Methods

The research draws on a recently collected panel data set of 1,200 households interviewed in 1993 and then again in 1998 in KwaZulu-Natal, South Africa's largest province, containing one-fifth of a population of 40 million. During the mid-1980s and again in the early 1990s, there was substantial political unrest and violence in KwaZulu-Natal. As a result, it is an especially interesting place to study social capital, which may have been partly eroded during periods of unrest.

Individuals were asked about membership in 20 different specific groups in 1998 and retrospectively in 1993. The groups included financial, production, sports and music, community service, religious, and political organizations. Financial and religious organizations dominate, but membership in all types of groups grew substantially over the period. In addition, certain characteristics of the groups were reported, including whether they are of mixed gender and the household's rating of group performance. On the assumption that more mixed groups and better performing ones will augment social capital, an index is constructed combining with equal weights the number of groups, the percentage with mixed gender, and the average performance rating for each household in both 1993 and 1998.

In South Africa, household-level social capital, as measured by group membership and characteristics, has no significant impact on expenditures in 1993, but a positive and significant one in 1998.

The multivariate analysis presents "standard" expenditure functions augmented by the inclusion of measures of social capital. For each household there are two observations, one measured in 1993 and the other in 1998. The dependent variable is logarithmic per capita monthly expenditure (a proxy for income) and the controls include indicators of location, race, household size, gender of the head, age of the head, and a household dummy variable or "fixed effect." Also included are the household and community (average) social capital indices. All estimated parameters except the household fixed effect are allowed to differ between 1993 and 1998.

Findings

Household-level social capital, as measured by group membership and characteristics, has no significant impact on expenditures in 1993, but a positive and significant one in 1998. Returns to education have also increased over the period. Put a different way, those with stocks of either social or human capital in 1993 appear to have benefitted from them over time. The structural and other changes in the South African economy appear to be changing the returns to various factors, possibly indicating greater levels of efficiency. Comparing the impact of a 10 percent change in 1998 household-level social capital with that of the same change in education indicates the latter has an effect three times as large. So while the impact of social capital is strong, it does not appear to be as large as that of human capital in this setting.

An important concern in this analysis is the direction of causality between income and the measure of social capital based on group membership. While it may be that joining groups helps augment one's income, it is also possible that some groups are akin to consumption goods and thus, having more income, leads one to join more groups. The possibility of reverse causality is addressed using instrumental variables techniques; the return to social capital decreases slightly but remains positive and significant when this is done.

The formulation of an index of group membership in this paper was deliberately inclusive and did not distinguish by group function. To some extent this is a weakness in the analysis, particularly if one is interested in evaluating possible policy responses. Future work will disaggregate the index in different directions, including function and characteristics, especially gender, as well as consider other possible measures of social capital, such as family networks.

Keywords: social capital, group membership, South Africa, expenditure function, fixed effects, instrumental variables

Recent FCND Discussion Papers

• Child Health Care Demand in a Developing Country: Unconditional Estimates from the Philippines, Kelly Hallman, August 1999 DP70
• Supply Response of West African Agricultural Households: Implications of Intrahousehold Preference Heterogeneity, Lisa C. Smith and Jean-Paul Chavas, July 1999 DP69

To order a discussion paper, please fill out an online order form, email IFPRI-FCN@cgiar.org or send requests to Food Consumption and Nutrition Division, IFPRI, 2033 K Street, N.W., Washington, D.C. 20006, U.S.A.