Modeling undesirable output with a DEA approach based on an exponential transformation: An application to measure the energy efficiency of Chinese industry

There is great interest in measuring energy efficiency of China's industry using data envelopment analysis (DEA) approaches. This paper proposes a new exponential transformation of undesirable outputs into desirable outputs as part of a DEA approach for calculating environmental efficiency by using all kinds of classic models. Different from the existing transformations, the exponential model should provide a more flexible way to deal with undesirable outputs, allowing the user to adjust the slope and distribution of the transformation function by setting a parameter. A strategy for maximizing the standard variance among the undesirable outputs after transformation is employed in this paper to determine the optimal parameter value for the exponential model. This paper empirically examines the energy performance of sectors in Chinese industry from 2010 to 2014 using the proposed DEA model and the results show that most sectors in Chinese industry have not performed well, especially the sectors concerned with energy extraction. Based on the results of the efficiency evaluation, this paper calculates how can each sector achieve efficiency improvement through energy conservation and emission reduction, and propose specific improvement target values for each sector accordingly. (C) 2019 Elsevier Ltd. All rights reserved.