Efficiency Estimation of U.S. Commercial Banking : A Stochastic Frontier Approach Using Gibbs Sampling

Banking cost or X-efficiency is dependent upon the frontier analysis method used to measure the efficient frontier. Parametric methods require estimation of a composite error model where the bank's efficiency parameter is a portion of the bank's deviation from the cost frontier of the banking cost function. In this article Bayesian-based Markov chain Monte Carlo (MCMC) methods, specifically the Gibbs sampler supplemented by data augmentation, are used for the first time to estimate the cost efficiency of a sample of U.S. commercial banks. Sampling-based computational methods are shown to provide a straightforward and reasonable approach to determining bank cost efficiency. Additionally, such new analytical capabilities provide summary statistics for statistical testing previously beyond the scope of classical methods