Tuning regression results for use in multi-stage data adjustment approach of DEA

Data envelopment analysis (DEA) has been a wildly used powerful method to measure efficiencies of decision making units (DMUs). However, DEA efficiency scores are influenced by uncontrollable factors for respective DMUs. Previous studies attempted separating such factors from DEA scores. Fried et al. (1999) proposed a multi-stage data adjustment approach using DEA and a regression model, and several studies have followed it, such as Fried et al.(2002), Avkiran and Rowlands (2008), and so forth. Firstly, we point out shortcomings of the traditional adjustment scheme for combining regression results for use in DEA in the multi-stage approach, and then we propose a new scheme for data adjustment. We demonstrate the effect of this adjustment formula using an electric utility dataset.