Radial and non-radial decompositions of profit change: an application

Grifell-Tatje and Lovell (1999) made a contribution to the literature by decomposing profit change in extended radial DEA framework into six mutually exclusive components. Their approach appears to suffer from two shortcomings. First, radial models do not achieve full efficiency when slacks are present, and therefore, the contributions of each of these components are grossly underestimated. Second, evaluations of these components, using base-period prices as weights, can be potentially misleading. To circumvent over these shortcomings, we, first, introduce non-radial DEA models, and second, provide strong theoretical argument in favor of either current-period prices/average price of both periods as weights to be used to value the contributions of each of these components. The Indian banking sector is taken as a case study to illustrate the radial and non-radial decompositions of profit change so as to empirically examine the role of competition on profit change as well as its six mutually exclusive components. Our broad empirical results are as follows: First, radial and non-radial models yield diametrical opposite sign on the contributions of various components. Second, the increasing efficiency change trend in all ownership groups after 2002 indicates an affirmative gesture about the effect of the reform process on the performance of the Indian banking sector. Third, despite the fact that nationalized banks are the oldest banks, they do not reflect their learning experience in their output-and resource allocation behaviors.