Published online by Cambridge University Press: 09 February 2016
A method for predicting conflict zones in civil wars based on point process models is presented in this paper. Instead of testing the validity of specific theoretical conjectures about the determinants of violence in a causal framework, this paper builds on classic literature and a wide body of recent studies to predict conflict zones based on a series of geographic conditions. Using an innovative cross-validation design, the study shows that the quantitative research program on the micro-foundations of violence in civil conflict has crafted generalizable insights permitting out-of-sample predictions of conflict zones. The study region is delimited to ten countries in Sub-Saharan Africa that experienced full-blown insurgencies in the post-Cold War era.
Sebastian Schutte, Marie Curie Fellow, Department of Politics and Management, Zukunftskolleg, University of Konstanz, 78457 Konstanz, Germany ([email protected]). The author would like to thank Rolf Turner for having saved the day twice when the author got lost in the intricacies of the spatstat package for the R programming language. Yuri Zhukov provided very helpful feedback as a discussant for this paper at the Peace Science Society Workshop “Disaggregation in Terrorism Studies,” Philadelphia, PA, October 9, 2014. Participants of the Workshop on “Early Warning and Conflict” hosted by the Peace Research Institute Oslo also provided great advice and helpful in April 2015. Finally, the author would also like to thank the EU FP7 Marie Curie Zukunftskolleg Incoming Fellowship Program (Grant No. 291784) for financial support. Replication data and code are available both on the PSRM dataverse and the author’s personal website at http://sebastianschutte.net/. To view supplementary material for this article, please visit http://dx.doi.org/10.1017/psrm.2015.84