A multi-parametric method for bias correction of DEA efficiency estimators

This paper emphasises the sensitivity of the data envelopment analysis (DEA) efficiency estimators to sampling variations of the production frontier and dimensionality of the production set. It has been proven that DEA yields asymptotic unbiased estimates. A DEA (smoothed) bootstrap method is widely being applied to address the inaccuracy of DEA estimators. The combination of DEA and a modified bootstrap expression enhances the statistical properties of DEA estimators without overcoming the inherent limitations of each of the two methods. This paper provides a non-resampling multi-parametric method to deal with the sensitivity of DEA estimators. The new method is applied to scaled samples, and the bias-corrected efficiency estimators are compared against their population counterparts. A comparative analysis among the standard bootstrap, the smoothed bootstrap, and the new method shows that the new method's estimations provide a better fit to the population than the estimations of the standard and smoothed bootstrap.