In this chapter, I analyze data on over 300 individual members of the communist regimes in Bulgaria, Czechoslovakia, East Germany, Hungary, Poland, and Romania. I explore how an abrupt post-Stalinist transition in the wake of the Soviet dictator’s death affected elite cohesion and the relationship between ruling coalitions and their coercive subordinates. Specifically, I test whether breakdowns in elite cohesion led to more punishment of coercive agency chiefs, and their more frequent removal from office. My test of this argument exploits both variation in elite cohesion across Stalinist and post-Stalinist regimes, and variation in Soviet authority over different types of coercive agents. I analyze original data on members of communist ruling coalitions to estimate survival models of their tenures. I find that the tenures of Defense Ministers and secret police chiefs were similar under Stalinist coalitions, but secret police chiefs had significantly shorter tenures than Defense Ministers under post-Stalinist coalitions.

In this chapter, I test the effects of post-Stalinist transitions on two important measures of agency capacity: officers employed and individuals registered as secret informants by coercive agencies. I present an original cross-national dataset on officer and informant numbers for every coercive agency in communist Central and Eastern Europe from 1945 to 1989. I show that countries that experienced post-Stalinist transitions had similarly sized coercive agencies to other states before 1953, but these agencies shrank thereafter while others continued to grow. I then estimate a series of difference-in-difference models to test the effect of post-Stalinist transitions on agency size. I find that agencies under post-Stalinist regimes had significantly smaller coercive agencies after Stalin’s death. This confirms the theoretical logic laid out in Chapter 2 in a broader setting than the comparative historical analyses of Poland and East Germany in Chapters 4 and 5. Although the number of cases and coverage of data here are limited, my results suggest that the logic of elite cohesion and coercive capacity laid out in Chapter 2 is applicable to a wide range of authoritarian regimes.