A model-combined estimator of elasticity of scale

Abstract

Zellner and Ryu examined several popular specifications of elasticity of scale that characterizes production functions. This note goes further by combining these functional forms and proposing a model-averaged estimator of elasticity of scale. The methodology used in Bayesin Model Averaging, and the procedure is illustrated by an empirical example. Results show that the data clearly prefer some models over the others, but it can be difficult to select a model that overwhelmingly dominates the rest. This model-combined estimator of elasticity of scale is free of choosing a single specification from many competing ones. Moreover, this approach is readily applicable to comparable issues regarding production functions.