Prior commitments: Compatible interests versus capabilities in alliance behavior

The trade-off model of alliances offers a theory of alliance formation and duration that has significantly advanced our understanding of why states enter into alliances and why these alliances persist. However, we argue that extant tests of this model are improperly specified, greatly reducing the overall usefulness of the trade-off theory, and we therefore use this paper to bring foreign policy interests back into the model. Specifically, we argue that dyadic interest similarity best captures the motivations behind alliance formation and termination, and together with capability concentration, accurately predicts when dyads will seek alliances and what types of alliances are likely to be formed. We test our theory using a greatly expanded temporal domain (1648-2000), and our results are consistent and robust across each model.