A Monte Carlo Study of Efficiency Estimates from Frontier Models

2013

It has been argued that the deterministic frontier approach in inefficiency measurement has a major limitation as inefficiency is mixed with measurement error (statistical noise) in this approach. The result is that inefficiency is contaminated with noise. Later stochastic frontier approach improves the situation with allowing a statistical noise in the model which captures all other factors other than inefficiency. The stochastic frontier model has been used for inefficiency analysis despite its complicated form and estimation procedure. This paper introduced an extra parameter which estimates the amount of proportion that an error component shares in the observational error. An EM estimation approach is used for estimation of the model and a test procedure is developed to test the significance of presence of the error component in the observational error.

Journal of Econometrics, 2013

Traditional stochastic frontier models impose inefficient behavior on all firms in the sample of interest. If the data under investigation represent a mixture of both fully efficient and inefficient firms then off-the-shelf frontier models are statistically inadequate. We introduce the zero inefficiency stochastic frontier model which can accommodate the presence of both efficient and inefficient firms in the sample. We derive the corresponding log-likelihood function, conditional mean of inefficiency to estimate observation-specific inefficiency and discuss testing for the presence of fully efficient firms. We provide both simulated evidence as well as an empirical example which demonstrates the applicability of the proposed method.

2008 Annual Meeting July 27 29 2008 Orlando Florida, 2008

The objective of this article is to address heteroscedasticity in the stochastic frontier cost function using aggregated data and verify it using a Monte Carlo study. We find that when the translog form of a stochastic frontier cost function with aggregated data is estimated, all explanatory variables can inversely affect the variation of error terms. Our Monte Carlo study shows that heteroscedasticity is only significant in the random effect and the unexplained error term not in the inefficiency error term. Also, it does not cause biases, which is quite opposite of previous research. These are because our model is approximately defined by first order Taylor series around zero inefficiency area. But, disregarding heteroscedasticity causes the average inefficiency to be overestimated when the variation of inefficiency term dominates the other error terms.

Journal of Econometrics, 2013

Traditional stochastic frontier models impose inefficient behavior on all firms in the sample of interest. If the data under investigation represent a mixture of both fully efficient and inefficient firms then off-the-shelf frontier models are statistically inadequate. We introduce the zero inefficiency stochastic frontier model which can accommodate the presence of both efficient and inefficient firms in the sample. We derive the corresponding log-likelihood function, conditional mean of inefficiency, to estimate observation-specific inefficiency and discuss testing for the presence of fully efficient firms. We provide both simulated evidence as well as an empirical example which demonstrates the applicability of the proposed method.

European Journal of Operational Research, 2018

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Highlights  A novel efficiency effects stochastic frontier model is proposed.  Standard normal distribution function of exogenous variables is employed to capture technical efficiency.  The proposed frontier specification contains only the two-sided random noise term  Philippines rice data is used to demonstrate usefulness of the proposed model.

Economic Modelling, 2009

The purpose of this paper is to propose a simple stochastic frontier model with a non-parametric specification for covariates affecting the mean of technical inefficiency. We derive a simple two-step semiparametric estimation procedure to estimate the frontier parameters as well as the mean of the technical inefficiency. The consistency of the estimator and its asymptotic normality are shown. The proposed method is illustrated using a large panel data set of British manufacturing firms.

Empirical Economics, 2004

Efficiency estimation in stochastic frontier models typically assumes that the underlying production technology is the same for all firms. There might, however, be unobserved differences in technologies that might be inappropriately labeled as inefficiency if such variations in technology are not taken into account. We address this issue by estimating a latent class stochastic frontier model in a panel data framework. An application of the model is presented using Spanish banking data. Our results show that bank-heterogeneity can be fully controlled when a model with four classes is estimated.

Empirical Economics, 2008

In this paper, we use the local maximum likelihood (LML) method proposed by Kumbhakar et al. (J Econom, 2007) to estimate stochastic cost frontier models for a sample of 3,691 U.S. commercial banks. This method relaxes several deficiencies in the econometric estimation of frontier functions. In particular, we relax the assumption that all banks share the same production technology and provide bank-specific measures of returns to scale and cost inefficiency. The LML method is applied to estimate the cost frontiers in which a truncated normal distribution is used to model technical inefficiency. This formulation allows the cost frontier, inefficiency effects and heteroskedasticity in both noise and inefficiency components to be quite flexible.

2000

This paper applies a large number of models to three previously-analyzed data sets, and compares the point estimates and confidence intervals for technical 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 estimates; and maximum likelihood given a distributional assumption. Bayesian procedures include a Bayesian version of the fixed effects model, and various Bayesian models with informative priors for efficiencies. We find that fixed effects models generally perform poorly; there is a large payoff to distributional assumptions for efficiencies. We do not find much difference between Bayesian and classical procedures, in the sense that the classical MLE based on a distributional assumption for efficiencies gives results that are rather similar to a Bayesian analysis with the corresponding prior.

Journal of Productivity Analysis, 2007

We study the construction of confidence intervals for efficiency levels of individual firms in stochastic frontier models with panel data. The focus is on bootstrapping and related methods. We start with a survey of various versions of the bootstrap. We also propose a simple parametric alternative in which one acts as if the identity of the best firm is known. Monte Carlo simulations indicate that the parametric method works better than the percentile bootstrap, but not as well as bootstrap methods that make bias corrections. All of these methods are valid only for large time-series sample size (T), and correspondingly none of the methods yields very accurate confidence intervals except when T is large enough that the identity of the best firm is clear. We also present empirical results for two well-known data sets.