Monte Carlo Analysis of Estimator Performance

Abstract

For applied efficiency analysis the main intended benefit from the proposed approach is the ability to identify differences in the means of output-specific efficiencies. So the primary questions of the following Monte Carlo analysis will be whether this ability is merely a theoretical one having only asymptotic appeal or whether it is also practically useful for typical sample sizes encountered in applied efficiency analysis. This ability is ultimately owed to the joint estimation of most likely DMU-specific target output-ratios from the large set of theoretically feasible target output-ratios. For determining the latter the bias corrected version of DEA as described in section 3 of Chapter 2 will be used, while output-specific efficiency estimation proper will be affected via MC² based on the posterior distributions derived in section 2 of Chapter 7, with some missing specifications yet to be added.