Bayesian efficiency analysis through individual effects: Hospital cost frontiers

This paper develops Bayesian tools for making inferences about finn-specific inefficiencies in panel data models. We begin by establishing a Bayesian setting in which fixed and random effects models are defined. What distinguishes these classes of models is the marginal prior independence of the effects. We show how such models can be analyzed using Monte Carlo integration or Gibbs sampling. These techniques are applied to a panel of U.S. hospitals. Our empirical findings illustrate the different characteristics of both types of models, as well as the influence of the particular priors used on the finn effects.