Statistical evaluation of Data Envelopment Analysis versus COLS CobbeDouglas benchmarking models for the 2011 Brazilian tariff revision Marcelo Azevedo Costa a, * , Ana Lúcia Miranda Lopes b , Giordano Bruno Braz de Pinho Matos c a Department of Industrial Engineering, Universidade Federal de Minas Gerais, Belo Horizonte, MG 31270-901, Brazil b CEPEAD, Universidade Federal de Minas Gerais, Belo Horizonte, MG 31270-901, Brazil c CEMIG, Belo Horizonte, MG, Brazil article info Article history: Available online 25 November 2014 Keywords: Benchmarking Corrected Ordinary Least Squares CobbeDouglas production function Data Envelopment Analysis abstract In 2011, the Brazilian Electricity Regulator (ANEEL) implemented a benchmarking model to evaluate the operational efficiency of power distribution utilities. The model is based on two benchmarking methods: Data Envelopment Analysis (DEA) and Corrected Ordinary Least Squares (COLS) with a Cobb Douglas production function. Although the estimated scores are highly correlated, differences between the scores are as high as 41%. For some companies differences between the efficiency scores result in substantial reduction in regulatory operational costs. We provide a detailed statistical comparison which indicates that the COLS Cobb Douglas model has major deficiencies in terms of estimating efficiency scores. © 2014 Elsevier Ltd. All rights reserved. 1. Introduction The use of frontier-based methods in the incentive regulation of power companies has grown in recent years. With an international survey performed in 40 countries on a total of 43 regulators be- tween June and October 2008, Haney and Pollitt [1] found that 51% of the companies were applying benchmarking techniques. The frontier techniques can be divided into three major groups: Data Envelopment Analysis (DEA), Corrected Least Squares (COLS) analysis and Stochastic Frontier Analysis (SFA). The authors found that in 2008, 34.8%, 13% and 8.7% of the power distribution regu- lation models were applying DEA, COLS analysis and SFA, respec- tively. They also found that 40%, 15% and 5% were applying DEA, COLS and SFA, respectively, in power transmission regulation models. Haney and Pollitt [1] report that Austria applies DEA and COLS for both power transmission and distribution utilities; Belgium applies DEA for power transmission and distribution and applies SFA and COLS in power transmission regulation models; Denmark applies COLS analysis; Portugal applies SFA; Slovenia applies DEA; Iceland, Netherlands and Norway apply DEA in power distribution and transmission models; Argentina and Brazil applied DEA for power transmission in 2008, but Brazil applied DEA and COLS models for power distribution companies in 2011. Col^ ombia began to apply DEA for power transmission in 2000 and for power distribution in 2002. Several studies have indicated the suitability of DEA in the analysis of efficiency in power-regulated sectors. For example, Edvardsen et al. [2] describe the Norwegian Electricity regulation model. The Norwegian Water Resources and Energy Directorate (NVE) was one of the first European regulators to use DEA. Bogetoft and Otto [3] describe the DEA-based incentive model applied in Germany; Hu and Wang [4] evaluate the efficiency of electric util- ities and their effects on consumer prices; Souza et al. [5] compare DEA and SFA in the measurement of the efficiency of 40 Brazilian energy distribution companies; Sarıca and Or [6] assess the effi- ciency of Turkish power plants using DEA; Vaninsky [7] uses DEA to measure the efficiency of electric power generation in the United States; Hu and Wang [4] analyze the energy efficiencies of 19 administrative regions in China for the period of 1995e2002 using DEA. A summary of the use of DEA in energy and environmental studies can be found in Ref. [8]. In September 10, 2010, the Brazilian National Electric Energy Agency began a debate with the Brazilian society regarding the rules and methodologies for defining the revenues of electricity distribution utilities for the 3rd Periodic Tariff Review Cycle (3PTRC) through public hearing 040/2010 (AP040). Through * Corresponding author. E-mail address: macosta@ufmg.br (M.A. Costa). Contents lists available at ScienceDirect Socio-Economic Planning Sciences journal homepage: www.elsevier.com/locate/seps http://dx.doi.org/10.1016/j.seps.2014.11.001 0038-0121/© 2014 Elsevier Ltd. All rights reserved. Socio-Economic Planning Sciences 49 (2015) 47e60