One option to tackle the predicament of civil war recurrence in post-conflict societies is to include power-sharing arrangements in peace agreements. The reasoning and motivation behind this political approach is twofold. Firstly, it is assumed that power-sharing institutions once established promote democracy in divided societies. Secondly, scholars argue that they generate higher chances for peace in post-conflict societies. The latter argument is supported by sufficient empirical evidence. Nevertheless, there is a vast literature concerning doubts about the effectiveness of power sharing in post-conflict societies. While there seems to be cross-sectional empirical evidence that power sharing makes peace more durable, an important question that has been neglected in a comparative perspective is, under which conditions are power-sharing agreements successful or under which conditions do they fail? Current research indicates that wrong incentives are set by the arrangements and spoiler groups might endanger the peace but most studies are short of systematic theory-building and hardly scrutinize causal mechanisms or processes. The paper will investigate what influence the in- or exclusion of spoilers in the peace process has on the stability of power-sharing agreements. It is assumed that political and territorial power-sharing bargaining results of peace negotiations set incentives for conflict parties to restart the conflict if they have not been included in the agreement. Based on a dataset of civil wars between 1944 and 2007 the hypotheses are tested with sophisticated quantitative analysis.
