A goal programming-discriminant function approach to the estimation of an empirical production function based on DEA results

This paper describes a new frontier estimation procedure which relies on results obtained from Data Envelopment Analysis (DEA). The paper reviews some of the earlier works in this area and points out potential difficulties with them. It further suggests ways to validate such developments. A procedure is constructed on the basis of a Goal Programming (GP)-Discriminant Function model developed in stages during the 1980s. A numerical example is used to illustrate the proposed procedure. Then, an extensive simulation in which the GP-based procedure is compared to three regression-based techniques is presented. The simulation results clearly indicate the superiority of the proposed technique over the regression alternatives.