Supply Chain Management, Endogenous Production Networks, and Aggregate Fluctuations

More on Le Xu

Abstract: I develop and estimate a multi-industry real business cycle model with production networks, in which firm producers choose both quantities and the number of input supplier firms in each industry. This extensive margin adjustment incurs a supply chain management cost which is monotonic in supplier number. The supply chain management cost, together with the return to variety in intermediate input composition, implies a positive correlation between a customer firm/industry's supplier number in a supplier industry and its input expenditure on that supplier industry. Using US firm-level supply chain relationship data, I confirm this positive correlation at both the industry and firm levels. The extensive margin adjustment has two effects on productivity shock transmission and aggregate output fluctuations: (1) It amplifies shocks through sales and input expenditures; and (2) it reallocates shocks' effects among supplier industries. The aggregate effect depends on the elasticity of substitution between different industries' goods as inputs, the curvature of the supply chain management cost function, and industry productivity processes. I estimate the model using indirect inference while simultaneously calibrating the industry productivity processes which are unrelated to the extensive margin adjustment. Simulations show that the extensive margin adjustment generates a 57.6% larger real GDP fluctuation compared to a model without it.