At Risk: Thresholds of Democratic Breakdown

• In Event: Democratic Backsliding and Authoritarian Control

Abstract

It has become an established practice in studies of comparative politics to utilize time-series datasets that code countries as democratic or autocratic according to a particular set of criteria (see Biox et. al 2012, Cheibub et. al 2010, POLITY, and VDEM). In turn, researchers often use this data to categorize whether a country was a democracy in a particular year. Yet most scholars would agree that a country that was coded as democratic for a single year before reverting back to authoritarianism was likely not a true or consolidated democracy for that short period. Is there a threshold for which once a regime transitions from a dictatorship to a democracy, it is unlikely to revert back? Using survival analysis, this paper analyzes multiple regime datasets to evaluate whether such a threshold exists. We find that once a country remains a democracy for twenty years, the probability of democratic breakdown drops significantly, and this cut point remains robust to various ways of coding “democracy”. This finding raises the question of whether scholars should consider countries under the twenty-year threshold as true democracies. It also suggests that countries that are still within the first twenty years of a transition are most at risk for democratic backsliding or breakdown.

Authors

• Benjamin Clay Helms, University of Virginia

• Jonathan Kropko

• David Leblang, University of Virginia

• Anne Meng, University of Virginia