Um estudo de outputs indesejáveis em DEA com aplicação no setor de distribuição de energia elétrica

Abstract

This study discusses the treatment of undesirable outputs in Data Envelopment Analysis (DEA), through a literature on the subject, being presented and discussed some approaches to the treatment of such undesirable variables. In this sense, the study makes comparisons between these approaches, explaining the pros and cons of each. Are tested four different approaches in the Case Study, which evaluates the efficiency of twenty electricity distributors in Brazil. It is considered for calculating the efficiency of these companies a set of variables, including two undesirable outputs. In order to increase the discrimination of efficient DMUs of the problem, the Reversed Frontier is used for the four approaches to treatment of undesirable outputs, comparing the new results. Finally, some comments are presented for each approach that was tested in the Case Study.