The Dictator's Endgame: Theory and Empirical Analysis of Military Behavior in Authoritarian Regime Crises, 1946 - 2014

First results have been presented in two papers published in 2018. In a contribution to _Democracy and Security_, we conduct three case studies of mass protests in Burma (1988), Sudan (1985) and the German Democratic Republic (1989) to identify the microfoundations of military behavior during these exceptional situations of regime crisis. In this, we find dictators brought down by their failure to maintain the armed forces’ loyalty during the endgame through credibly committing to delivering financial and political spoils to the military leadership and threatening sufficiently costly sanctions if the military defects from the dictator’s regime coalition. In an article in the _Journal of Democracy_, we delve deeper into the descriptive patterns of military reactions to dictator’s endgames from 1944 to 2014. We find that, even though dictators’ endgames occur in occurred in every world region that has a significant number of autocracies, they are very rare and rather recent phenomena – they occurred only 40 times since 1946, and three-quarter of all endgames occurred after 1980. In 19 cases the military stayed loyal to the dictator and repressed the protests, compared to 15 instances of militaries siding with the opposition and six cases in which officers staged a military coup against the dictator. In discussing potential factors that might explain the dictator’s ability to ensure the military’s loyalty, we find that military regimes as well as civilian dictators who stack their military leadership with loyalists along ethnic lines are more likely to see their militaries cracking down on the protests. In contrast, militaries who have not engaged in gross human rights abuses before the endgame have relatively less to fear from the dictator’s fall and a potential regime change, and are thus more willing to let the and support the goals of the protesters.