Mediation, Peacekeeping And Civil War Peace Agreements

Abstract

The post-civil war agreement phase is vulnerable to credible commitment problems, a lack of government capacity to implement, and/or mutual vulnerability to retribution from violating the agreement. This study’s main contribution is to demonstrate the combined utility of mediation and UN peacekeeping. Mediation builds trust and confidence and works with the parties to design an efficacious agreement conducive to, among other features, tamping down post-agreement violence. Peacekeeping stems violence and facilitates the implementation of the agreement. Agreements that are mediated and followed by UN peacekeeping are expected to be more robust in terms of staving off violence. We report the effects of the mediation–peacekeeping interaction using a method correcting for a common misinterpretation of interaction terms. We test logit and hazard models using a sample of full and partial civil war peace agreements signed between 1975 and 2011. Controlling for agreement design, democracy, and income per capita, the results indicate mediation and its interaction with peacekeeping reduce the probability of renewed/continuing violence and have a positive impact on agreement duration. We also report brief case study evidence from the 1990s peace process in Guatemala.

Keywords:

• Civil war
• Civil war peace agreements
• Mediation
• UN peacekeeping operations

Acknowledgment

We gratefully acknowledge the research assistance of Matthew Barrett.

Notes

¹ http://www.un.org/en/peacekeeping/missions/past/minuguabackgr.html.

² http://cic.nyu.edu/content/jean-arnault.

³ http://web.undp.org/evaluation/documents/thematic/conflict/Guatemala.pdf.

⁴ http://www.un.org/en/peacekeeping/missions/past/minuguabackgr.html.

⁵ http://peacemaker.un.org/guatemala-firmlastingpeace96.

⁶ UN website http://peacemaker.un.org/guatemala-implementation96.

⁷ http://www.un.org/en/peacekeeping/missions/past/minugua.htm.

⁸ https://www.ictj.org/ar/node/18103.

⁹ See footnote no. 8.

¹⁰ Norton, Wang, and Ai (2004) recommend using probit when estimating an interaction term in a nonlinear setting.

¹¹ The motivations for these robustness checks were: (1) peace agreements may be more stable after a long war (measured as log of war duration in days from data in Kreutz 2010) in which neither side could prevail on the battlefield; (2) since territorial wars in the UCDP data-set are on average longer and have lower death rates, the costs associated with renewed fighting are low.