Tie-Capacity Shocks and the Resilience of International Trade and Alliance Networks

Abstract

We analyze the effects of shocks on international networks. We develop an agent- based model (ABM) that allows us to: (1) examine two network formation processes that characterize international networks—preferential attachment and homophily; (2) study the properties of such networks at equilibrium; (3) induce shocks that reduce the capacity of nodes to form ties; (4) analyze how the networks react to these shocks by regenerating themselves in the post-shock period; (5) estimate network resilience—the extent to which networks retain their structural characteristics following shocks. We do this by comparing the pre- and post-shock structures of these networks at equilibrium and derive propositions about the relationship between shock type and network structure. We then test the propositions deduced from the ABM by comparing simulated networks to real-world alliance and trade networks. Our results suggest several things. First, shocks have a negative impact on nodal resilience in the ABM as well as on states’ alliance resilience, but not on trade resilience. Second, shocks have also indirect positive effects. Specifically, non-shocked actors whose alliance partners are dropped, increase their own resilience as well as the resilience of other nodes by forming new ties with other non-shocked actors, which increases the overall resilience of the network as a whole. This result matches the insights derived from the ABM. Third, at the network level, shock magnitude increases the modularity resilience of random networks, but reduces polarization and clustering resilience. This holds true for alliance modularity resilience but not for alliance polarization, which tends to increase following high-magnitude shocks. On the whole, economic shocks do not appear to have a significant effects on the resilience of trade networks. We discuss the implications of these results for theory and practice.