Connectivity and contagion: How industry networks shape the transmission of shocks in global value chains

We study how disruptions propagate through global value chains (GVCs) and how network structure conditions their effects. Using Eora country–industry data for 189 countries and 26 industries (1996–2015), shocks are defined by a distribution-based rule; exposure counts disrupted upstream/downstream partners, and redundancy counts non-disrupted pre-existing links. We estimate fixed-effects impact models—where robustness is smaller contemporaneous deterioration conditional on exposure—and a grouped-time complementary log–log hazard—where resilience is faster return to the pre-shock level. Two findings emerge. First, exposure propagates, depressing output, value added, and exports, and slowing recovery. Second, redundancy mitigates these effects, offsetting the impact and raising the recovery hazard. As an illustration, high-technology activities appear less sensitive on impact and recover faster than low-technology ones, an advantage explained by higher redundancy despite deeper connections. These results suggest that strengthening robustness and resilience requires deepening GVC integration through strategic redundant linkages, rather than retreating from openness.