The wrong suspect. An enquiry into the endogeneity of natural resource measures to civil war

This paper argues that natural resources in the past have been falsely identified as a cause of civil conflict onset. The idea that natural resources spur conflict has reached a certain degree of acceptance among scholars and policy makers with a number of quantitative studies concluding that natural resource abundant countries are more likely to experience a civil war than countries without such resources. However insights from the ‘resource curse’ literature suggest that measures of natural resources abundance conventionally used in these studies, can not be considered exogenous. Reversed causality and/or omitted political and economic variables may thus undermine the ability of these studies to prove a causal relationship between natural resource abundance and civil war. In this paper, I propose two more exogenous measures of natural resource abundance such as natural capital and subsoil capital. I identify those studies that find the strongest effect of natural resources on civil war (Collier and Hoeffler 2004, Collier Hoeffler and Rohner 2009) and replicate the exact same specifications, only changing the measure of natural resource abundance. Using the proposed more exogenous measures, I find no evidence that natural resource abundance leads to higher civil war risk. On the contrary, this paper provides some inconclusive evidence for the proposition that natural resource abundance may diminish civil war risk. Furthermore, I interact natural resource abundance with various measures of institutional quality, suggesting that the effect natural resources have on civil war might depend on quality of the institutions in the country at hand.