Working Paper Series Should Micro Credit Be Integrated with Local Developing Country Health Programs? Issues and Preliminary Evidence from Latin America

This paper examines the effects of tying microenterprise credit cooperatives (or “village banks”), with health and education services, using data from two Latin American countries, Ecuador and Honduras. Credit constraints particularly affect women (microentrepreneur) borrowers. Evidence suggests that mothers’ income increases are more likely to be spent on nutrition and other health improving expenditures for children than fathers’ income increases. Thus, relaxing credit constraints for women has been viewed as a particularly powerful development policy tool; and village banking has been a popular strategy for doing so. Many micro credit institutions, such as the Grameen Bank, bundle social programs together with credit, so their effects are difficult to disentangle. Tie-ins assume that credit may not be sufficient to raise incomes, while raising family income is not sufficient to meet development goals such as ending malnutrition, reducing infant mortality, and increasing the education levels of girls. There is at least some evidence supporting these assumptions, as reviewed in the next section. But non-financial responsibilities divert the attention of financial institutions from their comparative advantage; and a banking role causes health NGOs to lose their own comparative advantage. Other costs of a tie-in approach are reviewed. To resolve the debate on tie-ins, and to understand the source of effectiveness of micro credit, experimental evidence is needed. The paper uses financial and health data collected from mothers participating in Project HOPE’s Village Health Banks, in conventional credit-only village banks, and from women not participating in either type of bank, in Ecuador and Honduras, to provide the first evidence on the effectiveness of tie-ins. Fixed effects are accounted for with 104 community dummy variables for Ecuador, and 70 for Honduras; we also control for period effects, family size, age and education of the mother, and her marital status, past child deaths, and other proxies for general wealth and health. Effects on expenditures and breastfeeding incidence of health and credit bank participation were ambiguous. However, in Honduras, health bank participation was robustly associated with reduced conditional child diarrhea probability, while in no specification was credit bank participation found to reduce the conditional probability of diarrhea. In Ecuador, results were mixed but if anything supported a larger effect of the credit only banks. For Honduras, in all specifications health bank participants have significantly higher subsequent conditional probability of cancer screening, our proxy for formal health care. A much weaker effect was found for credit only participants. Similar but less pronounced effects were found for Ecuador. There is no clear link between tie-ins and performance of the banks themselves.