Feeding Conflict? New Data on the Impact of Humanitarian Food Aid on Civil Conflict

Abstract Anecdotal and empirical evidence suggests a link between humanitarian food aid and violent conflict, but recent empirical research produced mixed results. This research note reconciles disparate findings by showing that previous measures of humanitarian food aid (1) fail to account for differences in within-country transportation costs and (2) conflate humanitarian and nonhumanitarian food aid. I introduce a new dataset of USAID humanitarian food assistance across 103 countries between 1991 and 2019 which resolves these problems. I exclude shipping costs by using tonnage of food commodities and isolate the humanitarian portions of USAID's food assistance. I find a series of tightly estimated null relationships between humanitarian food aid and the incidence of conflict, conflict termination, and the duration of peace. These results do not change when I use these new data to measure the ease of appropriating humanitarian aid. By introducing new program-level data, this research note provides evidence on a disputed linkage and advances the literature on the unintended consequences of humanitarian assistance.