The focus of this article is civil war peace agreement duration from 1989 to 2005. Recent work by Hartzell and Hoddie (2003, 2007) has argued that power-sharing provisions have a cumulative impact. In other words, the more power-sharing provisions there are built into an agreement, the greater the prospects for peace. Our basic theoretical premise is that power-sharing provisions that are costlier to government and more difficult to implement will decrease the life span of the peace agreement because of government motivations to renegotiate and rebel incentive to strike preemptively before the government does or out of frustration because of delays in implementing costly provisions. In other words, governments will abandon the agreement because it concedes too much or rebels will abandon the agreement because of delays in implementation and/or to move preemptively. We look at three forms of power-sharing provisions: military (integration of rebels into army), territorial (autonomy), and political (shared government). Civil war peace agreements can expire after being replaced by a new agreement or if at least one party abandons the agreement. Hazard models are specified controlling for democracy score at time of signing, intensity of war, GDP per capita, and type of agreement.The results indicate that the less costly concessions by government of military integration and autonomy increase the duration of peace agreements, while political power-sharing provisions have a negative though insignificant impact on duration.