A review and empirical comparison of Bayesian and classical approaches to inference on efficiency levels in stochastic frontier models with panel data

This paper applies a large number of models to three previously-analyzed da ta sets, and compares the point estimates and confidence intervals for tech nical efficiency levels. Classical procedures include multiple comparisons with the best, based on the fixed effects estimates; a univariate version, marginal comparisons with the best; bootstrapping of the fixed effects esti mates, and maximum likelihood given a distributional assumption. Bayesian p rocedures include a Bayesian version of the fixed effects model, and variou s Bayesian models with informative priors for efficiencies. We find that fi xed effects models generally perform poorly; there is a large payoff to dis tributional assumptions for efficiencies. We do not find much difference be tween Bayesian and classical procedures, in the sense that the classical ML E based on a distributional assumption for efficiencies gives results that are rather similar to a Bayesian analysis with the corresponding prior.