Recent development in skew-normal based stochastic frontier analysis

Abstract Recently, a toolkit of stochastic frontier models using Skew-Normal distributions was developed for efficiency analysis. The Skew-Normal based stochastic frontier model (SN-SFM), a model based on Skew-Normal distribution, provides a solution to rectify the issue of incorrect skewness that poses a challenge for the conventional frontier model. This paper primarily focuses on reviewing the SN-SFM and discusses the expectation conditional maximization algorithm which tackles challenges in parameter estimation. A dataset from a U.S. electric utility is employed to illustrate the advantages of the SN-SFM with the Bayesian estimation approach.