Estimating efficiency effects in a stochastic frontier model with heteroskedastic errors

Abstract In this paper, we introduce a stochastic frontier model that incorporates efficiency effects and a heteroskedastic error structure. The mean efficiency is specified by a logistic function of the effects variables and the distribution for the one-sided random variable representing inefficiency is left unspecified. In contrast to conventional efficiency effects models, our approach does not necessitate the use of the JLMS (Jondrow in J Econom 19:233–238, 1982) transformation for computing efficiency scores. Efficiency scores are derived directly from the estimated model parameters using feasible generalized non-linear least squares. In order to illustrate the practical applicability of our proposed model, we present an empirical example using OECD electricity generation data.