Measuring super-efficiency in Data Envelopment Analysis

Data envelopment analysis (DEA) is a non-parametric oriented method for measuring relative efficiency of a homogenous set of decision making units (DMUs). DEA is originally proposed by Charnes, Cooper and Rhodes in 1978. Two basic DEA models are CCR (model with constant return to scale) and BCC (model with variable return to scale). All other models are variations, an upgrade or improvement of these two basic models. Data envelopment analysis is a very convenient method for assessing the relative efficiency of nonprofit units because it allows the use of multiple inputs and outputs without necessity of their functional dependence. Main shortage of the basic models is that efficient DMUs cannot be discriminated because they form efficient frontier (envelope) on which every DMU has unity score and because of that they cannot be compared between themselves. One possible way for ranking efficient DMUs is by using super-efficiency.