DEA environmental assessment of coal fired power plants: Methodological comparison between radial and non-radial models

This study discusses how to apply Data Environment Analysis (DEA) for environmental assessment. A unique feature of DEA environmental assessment is that it classifies outputs into desirable (good) and undesirable (bad) outputs because many firms often produce not only desirable outputs but also undesirable outputs as a result of their economic activities. A methodological difficulty of DEA applications is how to combine operational performance on desirable outputs and environmental performance on undesirable outputs in a unified treatment. Although previous DEA environmental studies have utilized mainly radial models and their extensions, this study uses a non-radial DEA model for the output unification because the non-radial model can unify the two types of outputs more easily than the radial models. This study incorporates three types of output unification in DEA environmental assessment. The first unification considers both an increase and a decrease in an input vector along with a decrease in the direction vector of undesirable outputs. This type of unification measures unified efficiency. The second unification considers a decrease in an input vector along with a decrease in the vector of undesirable outputs. This type of unification is referred to as natural disposability and measures unified efficiency under natural disposability. The third unification considers an increase in an input vector but a decrease in the vector of undesirable outputs. This type of unification is referred to as managerial disposability and measures unified efficiency under managerial disposability. All the unifications increase the vector of desirable outputs. Using the output unification under natural and managerial disposability, this study examines methodological strengths and drawbacks of the proposed non-radial approach. Moreover, using a data set on U.S. coal fired power plants, we compare methodological strengths and drawbacks of radial and non-radial models for DEA environmental assessments. The methodological comparison is important in guiding a large energy issue because policy implications depend upon a methodology(s) used for an empirical study. (C) 2012 Elsevier B.V. All rights reserved.