We decompose the relationship between food aid and conflict into the channels through which food aid can affect conflict. We address questions of methodological choice and estimation techniques for empirical studies. Our review of the empirical evidence on the effect of food aid on conflict shows that none of the previous studies proposes a compelling identification strategy. While existing research shows promising approaches in terms of econometric methods, i.e., instrumental variables estimation, they have not succeeded in i) using instruments that pass the necessary tests of instrumental variable estimation and ii) identifying the channels through which food aid influences conflict. We argue that future work should contain a rigorous identification strategy with a stricter focus on the impact of food aid on conflict intensity, empirically examine the conceptual channels through which food aid affects conflict, and the need for data at a more disaggregated level to achieve both objectives.