This paper has examined the effect of social capital on household poverty using the 2001 Cameroon household survey. We rely on three indicators for social capital – network membership, decision making index and network support or solidarity– and employ alternative procedures to consistently estimate the impact of social capital on household per capita expenditure. Memberships in organizations, social support or decision making indices are choice variables implying that social capital indicators are by definition endogenously determined and depend on household specificities. We exploit the advantages of longitudinal data and community fixed effects to mitigate some of the concerns about spuriousness and reverse causality that predominate in this literature.
Our results show that, membership in associations and the indicator for decision making index are positively correlated with household per capita expenditure (i.e. poverty reducing), this being true with classical OLS estimates as well as when we control for the endogeneity and reverse causality bias. However, the indicator for network support significantly mitigate household poverty when we control for endogeneity and reverse causality bias, an indication that households with higher incomes tend to group together. Secondly, there are limited economies of scale in social capital (i.e. more than one member of the same household belonging to networks does not necessarily mean more benefits). Our analysis suggest that policy makers interested in improving the living conditions of households may be advised to consider promoting social capital as one relevant ingredient to achieve the Millennium development goals of reducing poverty by half.