Research

The positive contemporaneous comovement between aggregate inventories and sales is a well-known stylized fact that guides the assessment of models and aggregate implications of inventory behavior. This paper highlights an overlooked feature that durable input inventory movements lag sales movements by around three quarters. This lagged comovement is discernible both in the unconditional cyclical components of data and in the impulse responses to identified aggregate shocks. To assess its quantitative significance, I develop a tractable supply chain production problem that is capable of reproducing the lagged comovement. In this model, producers are required to order critical inputs from suppliers one quarter in advance and they occasionally adjust their optimal order sizes based on forecasts of their own future sales subject to information frictions. I embed the production problem into a multisector New Keynesian model with input-output relations. Following a monetary shock, relative to a counterfactual scenario in which the inventory-sales comovement is fully synchronized, the estimated model demonstrates dampened responses of aggregate output over the first year but more gradual recovery over later horizons due to the reduced sensitivity of user cost of capital with respect to real interest rate changes.

Global trade flows and supply chains adjust gradually. Disparity of empirical estimates of trade elasticity in the short and long run suggests substantive adjustment frictions in trade. We develop a tractable framework that provides microfoundations for dynamic trade adjustment. The model features staggered sourcing decisions, nests the Eaton-Kortum model as the limiting long-run scenario, and rationalizes reduced-form estimation of horizon-specific trade elasticities. We calibrate the model with horizon-specific trade elasticities and sectoral input-output relations and use it to quantify the welfare impact of the 2018 US-China trade war. Staggered sourcing decisions imply that directly applying the well-known static welfare formula to observed domestic trade shares may generate qualitatively wrong predictions for ignoring the distortions. The short-run welfare impact can be smaller than the long-run level for US but larger for China despite the same low short-run trade elasticity. Third countries such as Mexico and Vietnam may experience welfare losses in the short run but welfare gains in the long run.

How are sales of an industry affected when sales of its immediate downstream industries fluctuate? Related studies have relied on exogeneity of shocks hitting the final demand and the Leontief inverse for identification. A limitation of such an approach is that the magnitude of the resulting estimates depend on both how industries respond to shocks and where they are placed in the production network. In this paper, I instead consider the possibility of directly estimating the sales responses of supplier industries to sales changes of their immediate customer industries. To overcome the identification challenge due to shocks passing from suppliers to customers, I construct a novel downstream demand instrument at the 4-digit SIC industry level by exploiting the compositions of intermediate inputs across customer industries and conduct estimation under a shift-share design. Preliminary results based on annual data suggest modest impact of sales fluctuations among customer industries on supplier industries. I plan to explore alternative monthly data that should provide additional insights in the near future.

cModel is the workhorse model behind the policy work conducted at the Globalization and Prosperity Lab at UC San Diego. It is a general equilibrium international trade model involving 170 industries across 44 countries connected via Eaton-Korton style trade for intermediate inputs and final goods. Simulation results from the cModel has been supporting the cBriefs for addressing pressing policy issues.