Abstract:We argue that two problems weaken the claims of those who link corruption and the exploitation of natural resources. The first is conceptual. Studies that use national level indicators of corruption fail to note that corruption comes in many fofms, at multiple levels, that may affect resource use quite differently: negatively, positively or not at all. Without a clear causal model of the mechanism by which corruption affects resources, one should treat with caution any estimated relationship between corruption and the state of natural resources. The second problem is methodological. Simple, atheoretical models linking corruption measures and natural resource use typically 4o not account for other important control variables pivotal to the relationship between huijaans and natural resources. By way of illustration of these two general concerns, we demonstrate that the findings of a recent, well-known study that posits a link between corruption and decreases in forests and elephants are not robust to simple conceptuaf and methodological refinements.