Using geo-referenced data on development projects by the World Bank and China, we provide a comprehensive analysis of the effect of aid on conflict using fixed effects and instrumental variables strategies. The results show that aid projects seem to reduce rather than fuel conflict, on average. Our analysis suggests that this is driven by projects in the transport and financial sectors, and through less lethal violence by governments against civilians. There are no clear differences based on ethnic fractionalization and government affiliation of a region, but some indications of spill-overs to other regions. We also find no increased likelihood of demonstrations, strikes or riots, but a higher likelihood of non-lethal government repression in areas where China is active.