This paper constitutes the first stage of an analysis of the problem of aid allocation when the donor is sensitive to both needs and governance considerations and is moreover able to influence local governance through its own disciplining effort. In this first stage, we write a principal-agent model of the relationship between a donor and a single recipient country. One key and original feature is the assumed comparability between domestic and donor-imposed disciplines: the two types can be summed up to obtain an aggregate discipline. We show that, paradoxically, an (exogenous) improvement of domestic discipline may be over-compensated by the donor so that total discipline actually decreases and elite capture increases. The relationship between domestic and total disciplines may thus be non-monotonous so that no simple general testable prediction can be inferred from economic theory regarding the impact of aid even controlling for domestic governance.
