Intended and unintended consequences of a maternal and child health program in rural Bangladesh: An investigation of anthropometric outcomes and intra-household spillovers

This dissertation is on the impact evaluation of a maternal and child health program in rural Bangladesh. The most important contribution of this study is the investigation of intra-household health externalities generated by the program on the non-targeted individuals living in targeted households. A theoretical model is developed to describe the mechanisms through which this spillover can occur; i.e. through an income effect, a household public good effect and a contagion effect . This program was administered in a treatment area with an equally impoverished area retained as a control area, allowing us to estimate treatment effects without the standard problems of endogenous program placement and selection. A reduced-form demand approach and data from the Matlab Health and Socio-Economic Survey of 1996 are used for the empirical analysis. This study also analyzes the impact of the maternal and child health program and other socio-economic characteristics on the health status of children. The paper also carries out a detailed investigation of the roles of mother's education and interaction between program resources and mother's human capital.;Results indicate a significant spillover on the health of the non-targeted elderly women. There is no evidence of health spillovers of the program on adult men. The externality is generated within the household and not through a community effect. Children residing in the treatment area are of higher quality in terms of health status. A significant determinant of children's health is mother's education and an important conclusion is that this program is a substitute for formal education for mothers with low schooling attainment. The treatment effect is positive and significant for girls' weight-for-age measures while boys are significantly less stunted in the treatment area as a result of the MCH program. Mother's education affects girls more than boys. However, the gender differential does not exist when the program is interacted with mother's education. To conclude, this study raises important issues for the design of public policy programs.