In this paper we deal with the measurement of technical efficiency (or X-efficiency) of institutions whose operations might be significantly affected by macroeconomic, environmental and non-controllable factors. For this purpose we introduce a four-stage DEA methodology based on the approach advocated by Fried - Schmidt - Yaisawarng (1999) and advanced by Drake - Hall - Simper (2003), respectively. The latter approach improves upon the former by employing a slacks-based DEA model (SBM) in combination with a Tobit regression approach to account for potential environmental and market influences on technical efficiency. In order to cope with the inherent dependency problem of DEA-based efficiency scores when incorporated into regression analysis we propose a Bootstrap method as suggested by Xue - Harker (1999). In so doing we attempt to overcome the dependency problem which plagues the power of standard regression analysis based on DEA data. As illustration, we apply this four-stage model to a balanced panel of data of 729 Austrian banks ranging over 1995 to 2002.
efficiency measurement data envelopment analysis slack adjustment environmental variables banking