Abstract

The non-parametric Data Envelopment Analysis (DEA) benchmarking method is frequently used by regulators to compare companies under natural monopolies. The Brazilian Energy regulator uses such an approach to define the efficient operational costs of the transmission companies. In 2019, a second-stage procedure was applied to include the effects of environmental variables on operational cost efficiency. However, only a few environmental variables were used to estimate efficient costs. This study proposes a new methodology for adjusting cost efficiencies using multiple contextual variables. Our proposal aims to adjust the input variable, using different environmental variables, before estimating the cost efficiencies. Therefore, multiple cost efficiencies are estimated, one for each environmental variable. Our proposal is based on the linear regression Analysis-of-Variance property. Results indicate that due to environmental heterogeneity of Brazilian transmission companies, cost efficiencies adjusted using multiple environmental variables are of utmost importance. Thus, our proposal manages to successfully include multiple environmental variables in the model, generating fair adjustments in the efficiency scores, and avoiding the effects of DEA modeling biases that are common in second-stage analyses.