Abstract: Coups d’etat continue to be common around the world, often leading to changes in leaders and institutions. We examine the relationship between military spending and coups and find that (i) successful coups increase military spending by more than failed attempts, and (ii) coups are more likely when military spending as a share of GDP is relatively low. Our identification strategy deals with the problem of reverse causality between coups and military spending by exploiting the conditional independence between a coup’s outcome and the change in military spending that follows it. We interpret our results as evidence that the military may stage coups in order to increase its funding, and rule out several alternative explanations.