DATA ENVELOPMENT ANALYSIS and Confident-DEA

This work contributes to the Data §Envelopment Analysis (DEA) literature at three ways. §First, it extends the roots of DEA by providing an §analytical approach deriving the basic Charnes-§Cooper-Rhodes (1978) model from the Weak Axiom of §Profit Maximization (WAPM) of Firm Theory. §Second, this work provides a systematic way for §classifying the existing DEA literature by offering §a taxonomy. Finally, a theoretical contribution to §the literature, Confident-DEA approach, is proposed §involving a bilevel convex optimization model to §which a Genetic-Algorithm-based solution method is §suggested. Complementing previous DEA methodologies, §which provides single valued efficiency measures, §Confident-DEA provides a range of values for the §efficiency measures, an efficiency confidence §interval and hence the name, reflecting the §imprecision in data. Monte-Carlo simulation is used §to determine the distribution of the efficiency §measures, taking into account the distribution of §the bounded imprecise data over their corresponding §intervals. Confident-DEA is applied to predict the §efficiency of banking systems in OECD countries.