Development aid and infant mortality. Micro-level evidence from Nigeria

While there is a vast literature studying the effects of official development aid (ODA) on economic growth, there are far fewer comparative studies addressing how aid affects health outcomes. Furthermore, while much attention has been paid to country-level effects of aid, there is a clear knowledge gap in the literature when it comes to systematic studies of aid effectiveness below the country-level. Addressing this gap, we undertake what we believe is the first systematic attempt to study how ODA affects infant mortality at the subnational level. We match new geographic aid data from the AidData on the precise location, type, and time frame of bilateral and multilateral aid projects in Nigeria with available georeferenced survey data from five Nigerian Demographic and Health Surveys. Using quasi-experimental approaches, with mother fixed-effects, we are able to control for a vast number of unobserved factors that may otherwise be spuriously correlated with both infant mortality and ODA. The results indicate very clearly that geographical proximity to active aid projects reduces infant mortality. Moreover, aid contributes to reduce systematic inter-group, or horizontal, inequalities in a setting where such differences loom large. In particular, we find that aid more effectively reduces infant mortality in less privileged groups like children of Muslim women, and children living in rural, and in Muslim-dominated areas. Finally, there is evidence that aid projects are established in areas that on average have lower infant mortality than non-aid locations, suggesting that there are biases resulting in aid not necessarily reaching those populations in greatest need.

Funding: We acknowledge the financial support from the Norwegian Research Council as part of the projects “Development Aid, Effectiveness, and Inequalities in Conflict-Affected Societies” (Grant number 250301) and “Armed Conflict and Maternal Health in Sub-Saharan Africa” (Grant number 193754).