Power-sharing arrangements are a common mode of conflict management in the aftermath of civil war. Prior work suggests that such arrangements lessen the likelihood of conflict recurrence while having deleterious effects on the quality of governance, particularly with regard to public goods and other social welfare provision. Power-sharing arrangements also vary widely in nature in terms of the degree of flexibility they afford to identitarian coalition construction (i.e. corporate vs. liberal arrangements), the branches of government in which they mandate the sharing of power, and the other institutional arrangements with which they are paired (e.g. decentralization, federalism, and/or foreign guarantorship). In this paper, we will use data collected on post-civil war power-sharing and public goods provision collected between 1945 and 2007 to assess which varieties of power-sharing produce better post-conflict governance. We will test several core hypotheses, including (1) that split executive arrangements lead to institutional deadlock inhibiting public goods provision and (2) that fiscal decentralization facilitates public goods provision, particularly with regard to local public goods. Our findings aim to complicate the power-sharing debate by describing which specific institutional arrangements are shown to prevent conflict recurrence while not prohibitively undermining elite collective action.