How production networks amplify shocks

In a broad class of macroeconomic models with production networks, it is difficult to discern which set of structural relationships between sectors amplify shocks and shape aggregate outcomes. As a remedy, we provide a formula that sidesteps this issue by considering linkages in isolation, thereby quantifying the macroeconomic significance of specific intersectoral relationships. In an application, we specialize our framework to derive a closed-form expression for network spillovers in efficient economies, where network spillovers are defined as the effect of shocks on GDP due to propagation to other sectors. Empirically, we find significant fluctuations in network spillovers for 43 countries between 2000 and 2014, suggesting this channel to be a key driver of macroeconomic outcomes. In a second application, we quantify the gains of having different hypothetical input-output structures, keeping the final expenditure shares of goods and services the same. We find the United States’ growth rate would have been almost 20 percent higher per year had its input-output architecture been identical to China’s.