Three essays in health microeconomics in developing countries: Evidence from rural Senegal

Abstract

Using a data set collected from 504 households and 18 public Primary Health Care (PHC) facilities of the Saint Louis region in Senegal, the thesis explores three important topics in the field of health microeconomics applied to developing countries, in order to improve health indicators in rural Senegal.

The first essay aims to analyse health care utilisation in an area where people live on 1.17 USD per day and where only 6% have health insurance coverage. Despite the high level of poverty, 84% of the individuals sought treatment from a qualified health provider during their last illness. A deep analysis of the context is conducted in order to explain such finding. The high rate of healthcare utilisation in the area is found to be attributable to the characteristics of the PHC facilities. Indeed, PHC facilities are highly accessible and offer good medical services at a low price. The low price of medical services in the sample explains why the demand for curative care is found to be price-inelastic. This latter result suggests that policies that will reduce the price of medical services in order to increase health care use are not likely to be effective. Then, data are used to investigate the effect of income (asset index versus expenditure index) on the level of Out-Of-Pocket (OOP) medical expenses. Results indicate that the asset index is not a good predictor of health expenditure incurred last illness, probably because farming households perceive seasonal revenues and some of the households that are classified as rich by asset may lack of liquidity at the time of the illness. An endogenous switching regression is used in order to correct the selection bias regarding the health provider visited during the last medical contact. Using the expenditure index, results suggest that a 1% increase in income increases OOP medical expenses incurred during the last medical visit by 0.48% for the patients who visited a Primary Health Care (PHC) facility and by 0.57% for those who bypassed first line facilities to visit a high-level provider. At the household level, a 1% increase in income increases monthly OOP medical expenses by 0.77%. Thus, the results suggest that health is a necessity good, which has several policy implications. First, it means that health policies should prevent poor people from having high OOP medical expenses. Second, the finding that health expenditure is involuntary stands for the exclusion of health expenditure in the computation of consumption aggregates and poverty figures. Finally, I examine how women’s bargaining power affects child nutritional status in rural Senegal. In order to correct for the potential endogeneity of women’s empowerment I use information on a mother’s ethnicity relative to that of the community she resides in and on climatic shocks to construct arguably exogenous Instrumental Variables (IV). While standard Ordinary Least Squared (OLS) suggests that if a mother has more bargaining power, her children will have a better nutritional status and Two-Stage Least Squares (2SLS) estimates indicate that the true impact is underestimated if the endogeneity of bargaining power is not taken into account. The results suggest that women’s bargaining power is a strong determinant of child nutrition as an increase in one standard deviation in the women’s bargaining power index increases the Middle-Upper Arm Circumference (MUAC) Z-score by between 0.47 and 1.01 standard deviations depending on the IV and the measure of women’s bargaining power used.