It's all about identification

Do civil wars affect economic growth or is it the other way around? This is a typical identification problem. Markus Bruckner and Antonio Ciccone empirically analyze this link and provide evidence that economic recessions do indeed cause civil wars. Their identification strategy uses downturns in international commodity prices as an instrument for economic downturn in sub-Saharan countries. They show that "a 10% fall in income due to falling commodity prices raises the likelihood of civil war in Sub-Saharan Africa by around 12%." This is how good econometrics can contribute to a serious and informed policy debate.

Read more here or here (if you have access to the CEPR discussion paper)!