How do governmental divisions within an authoritarian state influence protest outcomes? In this article, the authors propose two divergent mechanisms—"co-optation" and "coercion"—to capture the relationships between types of protest target and the violence that results from protest events. The "co-optation" hypothesis proposes that protests against judicial and security branches will be more likely to become violent compared to those against the administrative ones because protesters anticipate no substantial economic return from judicial and security branches that do not have financial resources at hand to distribute. The "coercion" hypothesis proposes that protests targeting judicial and security branches pose a lower risk of mass violence than those targeting administrative ones due to the public's fear of violent crackdowns by judicial or military branches that control the state's coercive means. Analysing a unique data set of protest events in China between 2006 and 2017, the authors find that protests involving administrative divisions are significantly less likely to turn violent when compared to those opposing nongovernmental targets, while protests targeting judicial or security divisions are significantly more likely to involve mass violence. The findings suggest that protest violence in authoritarian regimes is associated with the organisational divisions within an authoritarian government, and the explanation of the relations focuses on whether the branches have the co-optation capacity to allocate substantial economic resources instead of whether the branches control the coercive forces to intimidate the public.