Does Health Aid Reduce Infant Mortality in Malawi?

Sammanfattning: In this paper we analyze whether aid projects directed to the health sector significantly reduce infant mortality in Malawi. This adds to the aid effectiveness debate, where most works are cross-country studies as opposed to our subnational approach. We use geocoded data from the Domestic Health Survey 2010 (DHS) and from the AidData initiative, which covers 80 % of aid donations to Malawi from 2002. We analyze the years 2002 to 2008 and use fixed effects and difference in differences models on district level. We also analyze infant mortality on local level, looking at aid projects in proximity of observed births. On local level we aim to overcome the selection problem of non-random aid allocation with an instrumental variable and matching. We cannot reject the hypothesis of no effect of aid to the health sector on infant mortality. We find that aid projects are not allocated randomly nor solely according to need.