Reanalysis: Are coups good for democracy?

Abstract

Several recent studies find that coups in autocracy raise the ensuing likelihood of democratization. In a recent critique (“Are coups good for democracy?”), Derpanopoulos et al. dispute this link. This paper shows that their modeling approach, which includes fixed effects for each autocratic regime spell, suffers from severe bias. A reanalysis that applies similar models to the data without this bias recovers significant effects of coups and coup attempts on democratization, although only during the post-Cold War era.