Anarchy, self-interest, and the duration of interstate peace

One of the main assumptions of Realist theory is that nation-states make decisions concerning peace and war based on self-interest. Nation-states exist in an anarchic international system. The presence of anarchy requires states to put their own interests ahead of any other state's interests. This fact suggests that peace is likely to be unstable, as states clash over competing interests. While scholars of international relations have examined many other aspects of Realist thought, little empirical work has been done to determine whether self-interest drives conflict behavior.; This dissertation examines two theories that have differing views on the role of self-interest. First, I examine a prominent rational model of international interactions that incorporates the realist assumption of pure self-interest, as captured by expected-utility measures. Second, I present a socio-psychological theory---hereafter referred to as "Social Conflict Theory"---that presents a different argument. Social Conflict Theory argues that actors have varying levels of genuine concern for other actors, and that peace is more likely to endure when genuine concern for the other actor is high. Nation-states that are politically and culturally similar should have higher levels of concern for each other than politically and culturally dissimilar states. Hence, similar states are prone to work out their low-level problems without recourse to military violence. Social Conflict Theory also suggests that trust is an important factor for preserving peace. I incorporate this into my model using an aspect of my statistical technique.; Unlike traditional quantitative studies of international conflict that focus on conflict initiation, I test hypotheses from both of these theoretical perspectives using a new conceptual framework that makes peace the dependent variable. Using event history analysis techniques, I examine the effects of theoretically interesting variables on the duration of interstate peace.; I run analyses on data sets on multiple datasets consisting of multiple temporal and spatial domains. The most general analysis focuses on all dyads from 1816--1984. I supplement this with analyses of all dyads from 1950--1984---a temporal period consistent with recent work on international conflict---as well as analyses of politically relevant dyads for both time periods. I find that the variables representing rational self-interest do not consistently explain the duration of peace in any of these analyses. Variables capturing general concern fare somewhat better, with political similarity being an extremely robust predictor of the duration of peace.