This study investigates the relationship between the management of development aid and violent conflict in Africa. I exploit variation in World Bank project management quality driven by the assignment of project leaders of varying ability, combined with geo-coded data on all projects linked to performance report cards. I find that better project management reduces violent conflict across sub-national aid receiving regions. Poorly-managed projects increase conflict while well-managed projects do the opposite. Project monitoring is particularly important, and management matters most in regions with a recent history of warfare and for large projects that involve the transfer of appropriable resources.
Smallest Geographic Unit: Grid cells (within countries)