What limits the efficacy of coercion?

We model a game between an authority, seeking to implement its state identity, and a parental generation, seeking to socialize a younger generation into their own identity. The authority first selects a coercion level against the non-state identity. The parental generation then chooses whether to insurrect in response to the coercion level and, if not, decides how much to invest in socializing their children into the non-state identity. In this overlapping generations model, we formalize and explore the consequences of an intrinsic negative reaction to coercion: coercion resentment. We show how coercion resentment can create an interval where coercion has negative efficacy in imposing the state identity. This causes the rational legitimacy maximizing authority to restrain its use of coercion. We then show how this inefficacy of coercion can make certain levels of coercion unimplementable without causing the non-state identity to insurrect. This causes the long-run equilibrium size of the non-state identity group to be dependent on their initial size and, thus, path dependence. We consider the validity of the model by reviewing two historical episodes: Stalin’s secularization project (1922–1953) and the Counter-Reformation in early modern France and the Holy Roman Empire (1517–1685).