Last night I was writing my lecture notes on dictatorship, and realized that I didn’t have a source for a factoid I often tell my students.  Namely: Democracies and dictatorships have the same average growth rate, but dictatorships have higher variance.  The intuition: If you get a smart growth-oriented dictator, he follows the Nike strategy: “Just do it.”  If you get a sociopathic dictator, he burns his country down.  In democracies, in contrast, mediocrity and compromise reign.

But what’s the source?  It took me about fifteen minutes to track down a seemingly seminal 1991 paper by Sah in the JEP.  When I actually looked at it, however, there were no supporting empirics.  This left me wondering if my factoid was even true.  Yet eventually I found a good citation (Almeida and Ferreira, Economics and Politics 2002).  The variance of growth is 4.5 times larger in dictatorships than democracies.  There’s even a pretty graph:


I wish the data were more up-to-date – especially since the numbers for the Communist bloc seem so bogus in hindsight.  But since fixing the data to account for all the economic disasters of the Communist world would strengthen the conclusion that dictatorships have higher variance, I think I can teach this result with confidence next semester.