Modeling risk concerns and returns preferences in performance appraisal: An application to global equity markets

A technique used to assess relative performance in a multiple input–output framework is data envelopment analysis (DEA). In basic DEA models, an entitymay show its best performanceby selecting input and output factor weights different from those selected by the other entities in thesample. Hence, when usingbasic DEAmodels, divergence of weighting schemesacross the assessedentitiescannot beruled out.Weightingimbalanceisanother issueencounteredinthe application of basic DEA models. The assignment of an extremely low or zero weight to an input or an output factor implies that it is disregarded in performance appraisal.We appraise equity market performance using the Assurance Region Global (ARG)–DEA model where weighting divergence may be eliminated while controlling weighting imbalance. We show that riskconcernsandreturnspreferencescan bemodeled intheARG–DEA modelthroughthe bounds on the virtual input and virtual output ratios. Different combinations of risk concerns and returns preferences assess equity market performance under different risk-adjusted return scenarios and thereby allow sensitivity analysis of performance. ã 2014 Elsevier B.V. All rights reserved.