A Two-Step Theory and Test of Democratic Waves

Abstract

Scholars, observing clustering in transitions to democracy, argue that democratization diffuses across borders as citizens in autocracies demand the same reforms they witness in neighboring states. We disagree. The present paper asserts that the diffusion of democracy literature rests on weak theoretical foundations and does not properly test for diffusion. We advance an alternative two-step argument to explain clustering of democratization: (1) economic shocks, which are clustered spatially and temporally, induce the breakdown of authoritarian regimes; and (2) democratic diffusion, in turn, influences whether a fallen dictatorship will be replaced by a democracy or a new autocracy. Diffusion, despite playing an important role, is insufficient to explain the clustering of transitions, notably because it cannot account for the timing of the waves. Using data on 125 autocracies from 1875 to 2004, we show that economic crises trigger authoritarian breakdowns, while diffusion determines whether the new regime is democratic or authoritarian.