Abstract
We examine the ways in which the size of the governing coalition in a post–civil war state affects the durability of the peace. Previous studies relate the durability of the peace to the outcome of the civil war, the extent and forms of power-sharing arrangements, and the role of third-party security guarantors. We argue that the way conflict terminates and the power-sharing agreements between former protagonists structure the composition of governing coalition in the post–civil war state. Any settlement to civil war that broadens the size of the governing coalition should increase actors' incentives to sustain the peace rather than renew the armed conflict. Peace is more likely to fail where the governing coalition is smaller because those excluded from the governing coalition have little to lose from resuming armed rebellion. To test these propositions, we analyze data on post–civil war peace spells from 1946–2005.
Acknowledgments
A previous version of this article was presented at the 2009 annual meeting of the International Studies Association Annual Convention, New Orleans, LA. Data and replication files are available at http://dvn.iq.harvard.edu/dvn/dv/internationalinteractions.
Notes
1Calculated from the dataset to be used in this study.
2In the presidential election, the military commander who led the final battle with the LTTE received more than 60% of the votes in the Tamil strongholds (see CitationUbayasiri 2010).
3According to a news report, more than 6,500 Tamil youths, including 400 women, applied for 500 advertised police constable positions (see “Jaffna Tamils Flock to Join Sri Lanka Police,” The Independent, 30 September 2009).
4More detail on operationalization of these variables can be found in “Polity IV Project: Dataset Users' Manual” (CitationMarshall and Jaggers 2009).
5Mark J. Mullenbach and William Dixon, Third-Party Interventions in Intrastate Disputes (TPI-Intrastate Disputes) Project, available at http://faculty.uca.edu/markm/tpi_homepage.htm.
6We code only whether the power-sharing agreements were implemented or not.We do not code degree of success of implementation. For implementation of comprehensive peace agreements see ongoing research initiative “Peace Accords Matrix” at https://peaceaccords.nd.edu/.
7Because Ordered Logit is not commonly used to model panel data with ordered dependent variable, we have reported simple panel regression models.