December 2010

To determine the effect of economic shocks on civil conflict, the empirical approach must be tailored to the shocks’ persistence. I illustrate this point by revisiting Miguel, Satyanath, and Sergenti (2004). MSS argue that lower rainfall levels and negative rainfall shocks increase the probability of civil conflict in Sub-Saharan Africa over the 1979-1999 period. I find MSS’s approach and conclusion to be incorrect. For example, according to MSS’s data, lower rainfall levels and negative rainfall shocks (droughts) actually decrease the probability of civil conflict outbreak. The 1979-2009 data also reject that civil conflicts are more likely to start following lower rainfall levels or negative rainfall shocks.

JSTOR link to Miguel, Satyanath, and Sergenti’s (2004) Journal of Political Economy paper