Beyond Firms and Industries: Shock Propagation through Establishment- and Product-Level Supply Chains

This paper investigates how the granularity of supply-chain data affects the propagation of economic shocks through production networks. Using newly constructed establishment-level supply chains with product-level information links for Japan, we simulate disruption dynamics under alternative definitions of network nodes and input classifications. We show that defining inputs at the product level generates substantially larger propagation effects than industry-based classifications, indicating that coarse industry measures overstate input substitutability and underestimate systemic vulnerability. While establishment-level networks generally amplify shock propagation relative to firm-level networks, this effect is quantitatively modest, reflecting opposing forces of increased network complexity and greater substitution possibilities. We further demonstrate that incorporating establishment-level geographic information is critical for assessing region-specific shocks, as firm-level networks tend to overstate the impact of shocks originating in major metropolitan areas. Overall, our results highlight the importance of granular information on products, establishments, and geography for accurately evaluating supply-chain resilience and systemic risk.