While research on the impact of power sharing on civil conflict recurrence has burgeoned in recent years, there has been little scholarship exploring how power sharing affects the power relations between government and rebels more generally. We define power relations as the relative strength between the government and a rebel group after the signature of a peace agreement. Power relations are determined by these actors’ military capabilities and their status in the political system. Thus far, however, the factors determining how the power relations in the military and political fields will change after a civil conflict are largely unknown. We argue that power-sharing institutions transform the zero-sum logic of power relations into a positive-sum game. We test this reasoning using a unique data collection on the type of post-conflict power relations and power-sharing event data. The results show unintended effects of power-sharing institutions on the power relations of the former conflict parties, offering new avenues for future research beyond the focus on democracy or civil conflict recurrence.