Homogeniczny bootstrap Simara i Wilsona w estymacji miary efektywności technicznej

The main use of DEA is to calculate an estimate of the Farrell technical efficiency measure. By defining the statistical model, the stochastic version of the method is received, enabling us to derive the asymptotical form of sampling distribution of the estimate (called the DEA estimator). However, the distribution depends on unknown constants, which make deriving estimates of dispersion measures quite difficult. Thus, as an alternative, I used Simar and Wilson’s homogeneous bootstrap, a relatively simple and frequently used bootstrap procedure available in a software package for Frontier Efficiency Analysis with R (FEAR). The method is illustrated with an empirical example based on real data from the Polish energy sector.