IMF conditionality on corruption: the politics of good governance

Abstract

We offer a systematic analysis of IMF loan conditionality on corruption and argue that the Fund's pursuit of anti‐corruption measures is tempered by two political costs. First, because corruption is politically sensitive, the IMF risks destabilizing recipient governments and provoking crises. Second, these domestic risks translate into institutional costs: imposing such measures in strategically important countries (e.g., US allies or temporary UNSC members) can alienate key stakeholders. We posit that clear institutional guidelines reduce these institutional costs by aligning members and staff and promoting a more even application across the membership. We test these claims on a novel dataset of IMF loan conditions from 1980 to 2019, built with machine‐learning tools that distinguish direct anti‐corruption conditions from indirect ones aimed at accountability and transparency. The results strongly support the theory: following the IMF's 1997 corruption guidelines, the application of anti‐corruption conditionality has become more even across multiple measures; yet, domestic political costs continued to steer the Fund toward indirect rather than direct conditions. Beyond illuminating how the IMF addresses a core impediment to development, the study highlights institutional evolution and how IOs balance competing pressures.