Abstract
Droughts are associated with several societal ills, especially in developing economies that rely on rainfed agriculture. Recently, researchers have begun to examine the effect of droughts on the risk of Intimate-Partner Violence (IPV), but so far this work has led to inconclusive results. For example, two large recent studies analyzed comparable data from multiple sub-Saharan African countries and drew opposite conclusions. We attempt to resolve this apparent paradox by replicating previous analyses with the largest data set yet assembled to study drought and IPV. Integrating the methods of previous studies and taking particular care to control for spatial autocorrelation, we find little association between drought and most forms of IPV, although we do find evidence of associations between drought and women's partners exhibiting controlling behaviors. Moreover, we do not find significant heterogeneous effects based on wealth, employment, household drinking water sources, or urban-rural locality.