Social Networks and the Propagation of Household Shocks to Asset Prices

Abstract

When can local household disturbances move credit prices? I develop a financially weighted propagation operator on the social network. In a conditional household-intermediary equilibrium, its projection determines the household-demand component of the excess bond premium. Because responses decay at different rates across network patterns and channels, a location's relative financial importance generically changes with the horizon, so no single centrality measure summarizes exposure across horizons. Household measurements recover this projection exactly when they span independently measured demand sensitivities. Empirically, social exposure to storm-damaged counties is negatively associated with spending weeks later in counties outside the recorded footprints. In simulations on the observed network under the model's diffusion law, the operator predicts the propagated credit-premium component of held-out storms with about 30 percent less error than direct social exposure.