Gangless cross-evaluation in dea: An application to stock selection

This paper discusses the impact of ganging decision making units (DMUs) on the crossefficiency evaluation in data envelopment analysis (DEA). A group of DMUs are said to be gangingtogether if the minimum and the maximum cross-efficiency scores they give to all other DMUs are identical. This study demonstrates that the ganging phenomenon can significantly influence the crossefficiency evaluation in favour of some DMUs. To overcome this shortcoming, we propose a gangless cross-efficiency evaluation approach. The suggested method reduces the effect of ganging and generates a more diversified list of top performing units. An application to the Tehran stock market is used to show the benefits of gangless cross-evaluation.