DEA-Risk Efficiency and Stochastic Dominance Efficiency of Stock Indices

In this article, the authors deal with the efficiency of world stock indices. Basically, they compare three approaches: mean-risk, data envelopment analysis (DEA), and stochastic dominance (SD) efficiency. In the DEA methodology, efficiency is defined as a weighted sum of outputs compared to a weighted sum of inputs when optimal weights are used. In DEA-risk efficiency, several risk measures and functionals which quantify the risk of the indices (var, VaR, CVaR, etc.) as DEA inputs are used. Mean gross return is considered as the only DEA output. When only one risk measure as the input and mean gross return as the output are considered, the DEA-risk efficiency is related to the mean-risk efficiency. The authors test the DEA-risk efficiency of 25 indices and they analyze the sensitivity of their results with respect to the selected inputs. Using stochastic dominance criteria, they test pairwise efficiency as well as portfolio efficiency, allowing full diversification across assets. While SD pairwise efficiency testing is performed for first-order stochastic dominance (FSD) as well as for second-order stochastic dominance (SSD), the SD portfolio efficiency test is considered only for the SSD case. The authors´ numerical analysis compares the results using two sample datasets: before- and during-crisis. The results show that SSD portfolio efficiency is the most powerful efficiency criterion, that is, it classifies only one index as efficient, while FSD (SSD) pairwise efficiency tends to be very weak. The proposed DEA-risk efficiency approach represents a compromise offering a reasonable set of efficient indices.