Data envelopment analysis for decision making unit with nonhomogeneous internal structures: An application to the banking industry

Abstract

Traditional Data Envelopment Analysis (DEA) evaluates the relative efficiency of a set of homogeneous decision making units (DMUs) regarding multiple inputs and outputs. An important implication of the DEA is dealing with applications wherein the internal structures of DMUs are known, specifically those that have a network framework. In some situations the assumption of homogeneity among the internal of DMUs is violated; for instance, when a set of universities comprises DMUs but not all of them have the same faculties. This paper proposes a DEA-based methodology to deal with the problem of evaluating the relative efficiencies of a set of DMUs whose internal structures are nonhomogeneous. It is shown that the overall efficiency of each DMU could be evaluated through two stages; in the first stage subgroup efficiency scores are derived and the second one evaluates the overall efficiency score of each DMU using a weighted average of the subgroups efficiency scores obtained in stage 1. To show the practical aspects of the newly developed model, it is applied to a set of hypothetical data-set in addition to a real data-set on bank industry.

Keywords:

• Data envelopment analysis
• efficiency evaluation
• network
• nonhomogeneous DMUs
• banking