Commitment and conflict

War is an inefficient outcome and therefore states ought to prefer to bargain over areas of conflict instead of fighting. However, in the anarchy of international relations there is no actor with a monopoly of power to enforce contracts between states. States then face a commitment problem when bargaining to prevent war. This dissertation explores three models where this commitment problem can lead to war. The first chapter presents a model that allows for shifts in the distribution of power which play out over an arbitrary number of time periods. This leads to a sufficient condition that implies war under a broader set of conditions than previously shown in the literature. This condition implies that preventive war may be caused by relatively slow, but persistent shifts in the distribution of power. As theorized in power transition theory, differential rates of economic growth can potentially cause war under this mechanism. Relaxing the unitary actor assumption of the first chapter, the second chapter analyzes how the domestic institutional structure of countries affects the likelihood of war. We model institutional divergence by comparing an infinitely lived dictatorship to a democracy with a replaceable leader and allow a range of leader incentives within these institutional frameworks. We show that dictators, even welfare maximizing ones, may lead to war if the initial distribution of resources is highly imbalanced whereas a democracy with a forward looking electorate is always peaceful. Yet when a democratic electorate is myopic, preventive war may result. Political parties act as a mechanism to prevent this outcome. In the third chapter, I investigate adding a third actor to the bargaining model of war. In a static setting, the model uses a notion of cooperative stability to predict balancing and bandwagoning behavior in alliance formation. When extended to a dynamic setting, changes to the system that result in alliance shifting may cause war. Additionally, alliance formation need not correspond to the static solutions, suggesting that the dynamics of power are as important as the distribution of power in alliance formation.