Working Paper: CEPR ID: DP2121
Authors: Susanto Basu; John Fernald
Abstract: This paper estimates the degree of the returns to scale for 2-digit U.S. manufacturing industries from the output-based primal and price-based dual equations implied by firms' cost-minimization problems. It seeks to reconcile the cyclical behavior of the primal and dual productivity residuals by allowing for nonconstant returns to scale and imperfect competition. We find significant differences between the estimates of the returns to scale parameter derived from the primal versus the dual equations. The existence of time-varying markups reduces the incidence of significant differences in the primal versus dual returns to scale estimates for the durable goods industries but not for the non-durable goods industries. Likewise, the presence of the quasi-fixity of capital helps to reconcile the behavior of the primal and dual productivity residuals for the durable but not for the non-durable goods industries
Keywords: returns to scale; markups; quasifixity of capital; procyclical productivity
JEL Codes: D24; E32
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Primal returns to scale estimates (C51) | Greater incidence of decreasing returns to scale (R12) |
| Dual returns to scale estimates (C51) | Constant or increasing returns (D25) |
| Time-varying markups (D43) | Reduced incidence of significant differences in returns to scale estimates for durable goods (L68) |
| Time-varying markups (D43) | No effect on returns to scale estimates for nondurable goods (C51) |
| Quasifixity of capital (E22) | Reconciles behavior of primal and dual productivity residuals for durable goods (D22) |
| Quasifixity of capital (E22) | Does not reconcile behavior of primal and dual productivity residuals for nondurable goods (D22) |