Testing Democratic Peace Theory: A New Approach with Application to The Israeli-Palestinian Conflict∗

International relations scholars have developed many theories to describe the evolution and existence of international conflicts. One important idea is the “democratic peace,” which argues that democracies are less likely to engage in conflict with each other. However, much of the work testing democratic peace theory has been static and does not address dynamic, causal mechanisms by which democracy fosters peace. We address this shortcoming by proposing a new, disaggregated, research design for analyzing the complex causal dynamics of international conflicts. This research design uses Bayesian Vector Autoregression (BVAR) models. We then discuss how these models can be used to evaluate both in and out of sample predictions and argue that these are the real tests of the democratic peace theory and its key competitor. We present a detailed analysis of the Israeli-Palestinian conflict as a test case. Our results show that it is not just democracy per se, but election cycles that are important to the arguments of the democratic peace. Finally, using BVAR models, we demonstrate the including measures for elections can improve forecasts and inferences about the path of conflict elections convey information about intent and resolve. Prepared for presentation at the Annual Meeting of the International Studies Association, Montreal, March 17-20, 2004. This work is sponsored by the National Science Foundation under awards nos. SES0351179 and SES-0351205. The authors are solely responsible for its contents. For assistance in data compilation we thank Doug and Joe Bond; for suggestions on our research design we thank John T. Williams.