Uganda’s persistent poverty confronts external donors with the classic Samaritan’s Dilemma: altruism dictates generous budget-
support, yet the same funds risk being diverted through misappropriation and patronage networks. Combining a three-year panel of
World-Bank–financed projects (2019-2023), with elite interviews with Ugandan line ministries, and a donor survey covering the EU,
UK, US and Nordic agencies, we show that (i) every 1 % increase in direct budget support is associated with a 0.8 % rise in
perceived leakage; (ii) fear of reputational loss makes donors “over-monitor” but under-enforce, creating a Stackelberg game in
which the recipient government anticipates continued disbursements despite non-compliance; and (iii) the equilibrium outcome is a
low-effort, high-aid trap that neither reduces poverty nor satisfies fiduciary concerns. We identify two institutional innovations—ring-
fenced social-sector tranches and third-party escrow accounts—that credibly tie future tranches to independently verified service-
delivery indicators. Counterfactual simulations suggest these devices could cut leakage by up to 34 % and raise donor welfare by the
equivalent of 6 % of annual aid commitments. The findings extend Buchanan’s original Samaritan’s Dilemma to a multi-donor,
bureaucratic setting and offer a roadmap for reconciling the ethical imperative to help Uganda’s poor with the strategic imperative to
protect scarce aid resources.