Economic shocks and civil conflict: A comment

Miguel, Satyanath, and Sergenti (2004) argue that lowerrainfall levels and negative rainfall shocks increase conflictrisk in Sub-Saharan Africa. This conclusion rests on theirfinding of a negative correlation between conflict in t andrainfall growth between t-1 and t-2. I argue that this findingis driven by a positive correlation between conflict in t andrainfall levels in t-2. If lower rainfall levels or negativerainfall shocks increased conflict, one might have expectedMSS s finding to reflect a negative correlation betweenconflict in t and rainfall levels in t-1. In the latest data,conflict is unrelated to rainfall.