Publication Type: Conference Paper/Unpublished Manuscript
Abstract: In this paper we take a new look at the relationship between regime type and deadly militarized conflict among pairs of states (dyads) in the international system. With the goal of describing the general functional form of the relationship, we evaluate three hypotheses: democratic peace, regime similarity, and regime rationality. We employ both standard logit methods and a recently developed machine learning technique, a support vector machine (SVM). Logit is dependent on assumptions which limit flexibility and make it difficult to discern the appropriate functional form. SVM estimation, on the other hand, is highly flexible and appears capable of discovering a general functional form that is contingent on the values of other variables in the model. SVM results indicate that regime similarity is important in most dyadic interactions. However, for the special but important case of the most dangerous dyads (including enduring rivalries), regime rationality plays a role and the democratic peace effect is dominant. The results suggest that models of international conflict excluding distinct indicators for both political similarity and joint democracy are mis-specified and likely suffer from omitted variable bias. The importance of accounting for interactions between regime type and other correlates of war is also evident. SVMs are an especially useful complement to more conventional statistical methods. In addition to their other virtues, our SVM models are at least equivalent to logit in out-of-sample predictive power.