A1245
Title: Stochastic frontier analysis with scale mixtures of normal distributions
Authors: Boris Choy - University of Sydney (Australia) [presenting]
Abstract: Stochastic frontier analysis has been widely used as an econometric model to estimate the production frontier and to measure the inefficiency of a company. In a stochastic frontier model, it has two random error terms. The random error for inefficiency is one$-$sided while the random error for the noise is two$-$sided. Different combinations of distributions for these two random errors are considered, and their performance is compared in real examples.