Military Exercises and Network Effects.

Between 1980 and 2016, the percentage of states that partnered in a multinational military exercise (MME) increased twenty-fold. What explains this proliferation? Existing studies focus on the role of major powers and polarity but fail to explain exercises without great powers or the continuous growth of MME participation. I conceptualize patterns of exercises among all members of the international system as networks. Inferential network analysis shows that higher-order effects like popularity, transitivity, and memory increase the probability that states cooperate militarily. Countries with many connections have institutional knowledge and prestige to attract partners. Multinational coalitions form where mutual friends increase trust and create positive feedback. Past cooperation lowers the costs of future partnerships. Empirical analysis shows that the evolving network structure of MMEs is an emergent property driven by these interdependent processes, and that traditional explanations for security cooperation like great powers and alliances decrease in influence over time. [ABSTRACT FROM AUTHOR]