60 Pages Posted: 21 Mar 2006
Date Written: February 21, 2006
This paper shows that because of data limitations available estimates of returns to scale at the firm level are for the revenue function, not production function. Given this observation, the paper argues that, under weak assumptions, micro-level estimates of returns to scale are often inconsistent with profit maximization or imply implausibly large profits. The puzzle arises because popular estimators ignore heterogeneity and endogeneity in factor/product prices, assume perfect elasticity of factor supply curves or neglect the restrictions imposed by profit maximization (cost minimization) so that estimators are inconsistent or poorly identified. The paper argues that simple structural estimators can address these problems. Specifically, the paper proposes a full-information estimator that models the cost and the revenue functions simultaneously and accounts for unobserved heterogeneity in productivity and factor prices symmetrically. The strength of the proposed estimator is illustrated by Monte Carlo simulations and an empirical application. Finally, the paper discusses a number of implications of estimating revenue functions rather than production functions and demonstrates that the profit share in revenue is a robust non-parametric economic diagnostic for estimates of returns to scale.
Keywords: production function, identification, returns to scale, covariance structures
JEL Classification: C23, C33, D24
Suggested Citation: Suggested Citation
Gorodnichenko, Yuriy, Using Firm Optimization to Evaluate and Estimate Returns to Scale (February 21, 2006). Available at SSRN: https://ssrn.com/abstract=892142 or http://dx.doi.org/10.2139/ssrn.892142