Finding the single efficient unit in data envelopment analysis with flexible measures

Efficiency evaluation is crucial for providing decision support in both target setting and competency recognition. In many practical applications, flexible measures that can be classified as either input or output often lead to inconsistent results due to distinct classifications. Conventional classification methods may identify many efficient decision making units (DMUs), which cannot provide us with the best alternative and reduces the discriminating power of associated models. In this study, we develop a novel integrated approach by identifying the single efficient DMU considering flexible measures. Specifically, our proposed approach guarantees the feasibility of the associated efficiency estimation model under relatively mild conditions. Moreover, the weight scheme and non-dominated classification result can be uniquely determined via a revised min-max computational procedure. We validate and illustrate the proposed approach by applying it to evaluate the efficiencies of Chinese commercial banks before and during the pandemic. Our findings suggest that the pandemic does have an impact on bank efficiency, and different banks respond differently to the pandemic shock. More specifically, the best-performing bank that stands out before and during the pandemic can be identified as a unique benchmark, and the others could learn from this alternative to mitigate the pandemic shock.