Should aid reward good outcomes? Optimal contracts in a repeated moral hazard model of foreign aid allocation

We consider in this paper a repeated moral hazard model where a donor, characterized both by altruistic and non altruistic motives, finances a three periods poverty eradication project. In order to model the significant problems that donors face in the actual implementation of aid programs, we assume that the elites of the recipient country, who play an important role in carrying out the project, have an incentive to divert resources from the intended use. We show that optimal aid contracts should be conditional on the previous results of the project. We distinguish however between strong conditionality where contracts are specified on the basis of the performance of the project in all periods and weak conditionality where contracts have, instead, short memory. In this case a recipient that experienced a negative performance will receive less aid in the following period, but will bear no further consequences in the future. If a donor assigns a lot of weight to the welfare of the recipient country compared to the cost of giving aid and the incentive of the elite to divert resources, an optimal aid allocation policy always implies a positive level of aid even if the project had a negative outcome in the previous period. In the opposite case, optimal contracts imply no aid after a negative performance of the project