At a general level the question is adressed, how to compare structural efficiencies for subgroups in the framework of data envelopment analysis. If a multiproduct technology is
employed in estimating the production possibility frontier and if the subgroups are distinguished by products, then specific problems arise. It is illustrated by Monte-Carlo
techniques that there may be considerable bias in relative structural efficiencies due to unbalanced numbers of observations for different subgroups. A resampling procedure is shown to outperform the simple estimators from data envelopment analysis in such circumstances. The method proposed is applied to austrian farming data to derive
conclusions about the relationship between certain specialization patterns and efficiency.
|Reihe||Department of Economics Working Paper Series|
- Department of Economics Working Paper Series