Does World Bank aid to countries damaged by civil conflict meet its stated goals of speeding economic recovery and reducing the risk of conflict recidivism? We contend that the Bank’s success depends on its ability to bolster and signal the credibility of politicians’ commitments to peaceful politics and tailor its programs to the post-conflict environment. In the first systematic evaluation of World Bank post-conflict assistance, we estimate selection-corrected event history models of the effect of Bank programs on recovery and recurrence using an original dataset of all World Bank programs in post-conflict environments. Among key results, we find that the Bank tends to select aid recipients according to their pre-existing probability of conflict recurrence and that, once we control for this non-random selection, the Bank has no systematic effect on either conflict recurrence or economic recovery.