Abstract
This paper analyzes how economic development affects the utility of initiating interstate wars. Economic development affects the utility of war through multiple channels. Some of these channels works through the economy of the potential initiator, some through that of the target under consideration. This generates complex, non-linear and contingent effects. We argue that earlier research on development and interstate war has failed to sufficiently distinguish between the impacts on the initiator and the target states, respectively. We replicate Gartzke's (2007) study, extend it by distinguishing between aggressor and target states, and point to some interpretational problems in his model. We construct a formal model that incorporates several channels through which both aggressor's and target's level of economic development affect the utility of war. The model yields specific propositions showing that the aggregate effects from aggressor and target development level are non-linear, non-monotonic and contingent on several factors. These effects are investigated empirically through a statistical estimation of the theoretical model. We obtain several interesting results. Among others, we find a strong, negative interaction between aggressor and target development level on probability of war, which means that two rich countries seldom fight each other. Moreover, we corroborate Gartzke's finding of a strong, positive interaction between aggressor development level and distance to the target, implying that rich countries are generally also less likely to attack their neighbors than poor are. Finally, we find that large, poor countries are attractive targets for many different aggressor types.