In developing countries with weak enforcement institutions, there is implicitly a large reliance on electoral incentives to reduce corruption. However electoral discipline works well only under some conditions. In this paper we study the effect of electoral competition on corruption when uncertainty in elections is high (or accountability is low), as in many developing countries. Our theory focuses on the case of high uncertainty and shows that in this case there is a U-shaped relationship between electoral competition and corruption. We illustrate the predictions of the model with village level data on audit detected irregularities and electoral competition from India.

In this roundtable on the last day fo the ICRN Forum 2021 we welcome an eclectic mix of researchers and practitioners to discuss how the two fields can learn from each other. Academics, also those studying corruption, are often accused of being in an ivory tower, far removed from real-world problems. Practitioners and policymakers, however, face criticism of not using empirically-validated methods to tackle corruption.