Estimation and inference in cross-sectional, stochastic frontier models
(2005) , 33 pages
Files
STATdp0524.pdf
Open Access - Adobe PDF
- 227.55 KB
Details
- Authors
- Abstract
- Conventional approaches for inference about efficiency in parametric stochastic frontier (PSF) models are based on percentiles of the estimated distribution of the one-sided error term, conditional on the composite error, rather than on the sampling distribution of the inefficiency estimator. When used as confidence intervals, these have extraordinarily poor coverage properties that do not improve with increasing sample size. We present a bootstrap method that gives confidence interval estimates based on sampling distributions, and which have good coverage properties that improve with sample size. In addition, researchers who estimate PSF models typically reject models, samples, or both when residuals have skewness in the “wrong” direction, i.e., in a direction that would seem to indicate absence of inefficiency. We show that correctly specified models can generate samples with “wrongly” skewed residuals, even when the variance of the inefficiency process is nonzero. Our bootstrap method provides useful information about inefficiency and model parameters irrespective of whether residuals have the skewness in the desired direction. We also find that a commonly-used Wald test used to test the existence of inefficiency has catastrophic size properties; likelihood-ratio and bootstrap tests are shown in Monte Carlo experiments to perform well both in terms of size and power.
- Affiliations
University of Texas Department of Economics
Citations
Simar, L., & Wilson, P. (2005). Estimation and inference in cross-sectional, stochastic frontier models (STAT Discussion Papers 0524). https://hdl.handle.net/2078.5/32854