Skewness and Time-Varying Second Moments in a Nonlinear Production Network: Theory and Evidence

This paper studies asymmetry in economic activity in a multisector model with shocks to productivity and labor wedges. Complementarity across inputs—creating nonlinear intersectoral interactions—creates negative skewness. The analysis generates additional predictions—skewness is smaller at the sector than aggregate level, sector-specific shocks are unskewed, and sector centrality rises following negative shocks—and finds empirical support for them. Skewness arising out of intersector interactions helps reconcile differences in skewness at the micro and macro level. Finally, we show the model’s ability to match the data comes from the wedge shocks, rather than variation in productivity.