Overlapping Communities and the Coevolution of the Defense Cooperation-Conflict Networks

Abstract

Global networks are distinguished by the presence of communities varying in size, often overlapping to form "community brokers" with memberships across multiple communities. Utilizing Link Community Detection Algorithm, this study unveils two hierarchi- cal influences of such overlapping communities on global security networks. Firstly, the cross-network effect posits that community brokers, through their involvement in various communities, attain an informational edge, potentially mitigating inter-communal conflicts. Secondly, the network endogenous evolution effect suggests that diminishing reliance on a singular community may erode the credibility of community brokers in fostering communal security, diminishing their attractiveness as defense collaborators and intensifying rivalries among brokers. Using data on global defense cooperation and conflicts from 1990 to 2010, the study tests these hypotheses. While fixed-effect Logit models provide evidence for the cross-network effects, when incorporating the co-evolution of defense cooperation and conflict networks via Stochastic Actor-Oriented Models (SAOM), this effect has limited ability to explain interstate conflict. However, SAOM elucidates a network hierarchy overlooked by dyadic models and support the endogenous evolution effect: states with greater importance in established defense communities are less favored targets for new defense partners, and community brokers show a reduced propensity to build direct defense cooperation links with each other. This research elucidates the complex dynamics and hierarchical structures shaped by overlapping communities within global networks, offering novel insights into the intricate web of interdependencies that underpin global security dynamics.