I study the economic and asset pricing implications of variation in idiosyncratic-volatility spillovers. Theoretically, I develop a dynamic production-network model with exogenously varying intermediate-input weights, in which product-market input linkages generate spillovers in idiosyncratic stock-return volatility. The model implies that time variation in the spillover structure explains variation in aggregate consumption-growth volatility and that innovations in this structure affect innovations in the stochastic discount factor (SDF); assets exposed to these innovations therefore command risk compensation. Empirically, variation in idiosyncratic-volatility spillovers significantly predicts future aggregate volatility and carries significant prices of risk with model-consistent signs.
