Empirical approaches to the study of interstate conflict

This dissertation empirically examines the effect of political regime type and trade on inter-government interactions. These characteristics are said to form the basis of a "liberal peace" in which joint democracy and trade interdependence cause peace. While the empirical effects of trade are unclear, regression analyses consistently show that democratic pairs of states are less likely to engage in militarized conflict. This notion of a democratic peace suggests to democratic policymakers the peace enhancing benefits of promoting democracy worldwide. In Chapter 1 I review the theoretical and empirical basis of a democratic peace.; Existence of a democratic peace suggests something about the way countries interact economically, as one might expect that democracies are unlikely to engage each other in economic forms of conflict. In addition I suggest that current empirical evidence is insufficient to draw the conclusion that joint democracy causes peace because the rarity or absence of militarized conflict among democracies does not necessarily insure the absence of other forms of conflict. Towards these ends I examine in Chapter 2 the effect of joint democracy on the use of economic sanctions that pursue foreign policy goals. I find that joint democracy has a negative and significant effect on reducing economic conflict; In the previous analyses it was assumed, as standard in the literature, that the variables selected as controls form the "true" model that generates the data. This though is problematic in the analysis of interstate conflict because there are several theories that explain conflict and hence there are several candidate variables that researchers may select in forming their models. Further, this is important, as the choice of controls is seen to influence the magnitude and sign of coefficients in addition to their standard errors. In Chapter 3, I weaken the assumption that the researcher knows the "true" model in assuming that they only know that the true model lies within a subset of models. I apply Bayesian model averaging over this subset and find that there is strong evidence against trade having an effect on the probability of militarized conflict, while joint democracy remains a significant negative effect.