A Note on Capacity Utilization and Measurement of Scale Economies.

This article investigates the measurement of returns to scale (RTS) via a cost function in the presence of quasi-fixed inputs. Caves, Christensen, and Swanson (CCS) proposed estimation of a variable-cost function and derived a formula for computing RTS from it. This article extends the CCS results by (1) investigating the bias that would result from erroneously using a total-cost function for measuring RTS, (2) indicating that the CCS procedure for measuring scale economies in the presence of quasi-fixed inputs is inconsistent for the case of nonhomothetic production technology, and (3) proposing an alternative procedure that allows one to evaluate RTS at long-run equilibrium points, even when firms in the sampler are disequilibrium. The similarity between this proposed method and the "temporary" equilibrium approach for measuring productivity growth is noted as well. The empirical results on RTS are compared among the three alternstive moethods (total-cost, CCS variable-cost, and proposed long-run equilibrium.