Skip to main content
Open Access Publications from the University of California

UC Berkeley

UC Berkeley Electronic Theses and Dissertations bannerUC Berkeley

Rebels with a Cause: Power Sharing, Negotiated Settlements, and the Logic of Preemptive Defection


Why are some negotiated settlements successful at resolving intrastate conflict, while others fail? Are settlements involving agreement to share power more effective and, if so, why have conclusions about power sharing been so mixed? I theorize in this project that power-sharing guarantees improve the prospects for a peaceful settlement by reducing the stakes of winning post-conflict elections, thereby increasing the likelihood that rebels will willingly demobilize.

Contrary to the preeminent model of bargaining for peace in the literature, I posit that the costs of complying with a negotiated settlement are asymmetrical. For rebels, compliance means forfeiting military capacity and bargaining power. In contrast, the government never concedes its monopoly on the use of force and, therefore, always retains the option of resorting to military action. Given this, and the risk of competing against an entrenched incumbent with an electoral advantage, it is a rational strategy for rebels to resist demobilization until expectations of future benefits are sufficiently high. Power-sharing reforms can help to increase the perceived value of the payoff to rebel elites, thereby increasing willingness to demobilize, but only when they are designed to outlast elections. Institutions that expire after a transitional period, such as coalitions and governments of national unity, fail to redress the incumbency advantage or to reduce the importance of winning post-conflict elections outright, heightening incentives for rebels to preemptively defect.

The logic of my theory of preemptive defection explains why rebels would be willing to trade military capacity for long-term power sharing, and it also acknowledges the potential for rebel splintering after a settlement is signed. In the absence of power-sharing guarantees, certain elements of the rebel leadership might still be able to secure a deal that is personally beneficial. As this process of selective cooptation reveals winners and losers during the implementation period, disgruntled rebel elites have an incentive to defect unilaterally, and the likelihood of conflict recurrence is determined by their capacity to access the resources of war. Specifically, my work shows that splintering is enabled by two factors. First, the failure to sign an all-inclusive settlement means that excluded groups are more willing to form an alliance in order to facilitate continued conflict. Second, governments with a reputation for defecting on their peace agreements are easy targets for rebel leaders hoping to mobilize defection from within their own ranks.

Empirically, this project relies on a nested design. It starts with new data collection on 138 negotiated settlements to domestic armed conflict signed between 1975 and 2005. I coded each settlement for its inclusion of power-sharing provisions, with specific attention to different types of power sharing and the distinction between permanent and transitional guarantees. A binomial logistic regression analysis of conflict termination reveals strong support for the central expectations of this study. Settlements that include provisions for permanent power sharing are significantly more likely to result in peace, all else equal, while transitional arrangements have no effect. I then selected a best-fit case exhibiting change over time in order to examine the underlying mechanisms at work. Through extensive field research in Uganda, including interviews with rebel and government representatives and archival and secondary research, I sought to understand why members of the insurgent party eventually rejected four of the five settlements signed, and what conditions allowed them to continue fighting for a more favorable bargain.

My research indicates that settlements are more likely to succeed in resolving conflict where they include provisions for power sharing over the long term. Such guarantees increase the willingness of rebels to surrender their weapons, and they create more diffuse benefits and long-term time horizons throughout the leadership, which reduce the potential for splintering. These findings offer a number of contributions to the literature on conflict resolution and post-conflict power sharing. Most importantly, they show that including permanent and transitional formulas under the same umbrella concept have led to faulty and contradictory conclusions about the effectiveness of power sharing. In addition, by disaggregating the way that the benefits of a settlement are perceived across the rebel elite, this project helps to explain why many are signed that have few prospects of success, and it provides lessons for identifying such bargains before they can degenerate into renewed conflict. Finally, the theory departs from the common focus on coercive third-party enforcement of peace agreements by emphasizing the rationality of rebel decisions to comply or defect and showing that settlements are most effective where insurgents prefer compliance—namely, where the benefits offered outweigh the costs of both continued fighting and democratization. This suggests a different role for the international community in peace processes, specifically in disseminating new norms and strategies for engineering institutions, rather than investing in costly peacekeeping missions.

Main Content
For improved accessibility of PDF content, download the file to your device.
Current View