Isolation of the Environmental Effect on Efficiency Scores: A Comparative Analysis of Two Second-Stage Correction Propositions

This study analyzes the technical efficiency of electricity distributors in Brazil using Data Envelopment Analysis (DEA) and second-stage models. Initially, efficiency scores are estimated using an input-oriented AED model. Then, the influence of environmental factors is evaluated, especially rainfall, using two approaches: Ordinary Least Squares (OOM) and the method of Simar and Wilson (2007). In addition, a procedure for adjusting the scores is proposed in order to isolate exogenous effects, allowing a more adequate comparison between the units. The results indicate high heterogeneity among the distributors and show that part of the inefficiency observed is due to environmental conditions. Second-stage methods significantly reduce the correlation between efficiency and exogenous variables, providing a more consistent measure of managerial efficiency.