Estimação dos principais direcionadores dos custos operacionais das empresas brasileiras de transmissão de energia elétrica utilizando modelos de regressão e programação linear

Abstract

The Brazilian Electricity Regulatory Agency (Agência Nacional de Energia Elétrica - ANEEL) has published the subsidy taking document - TS 14/2022 to gather information regarding the database to be used in the regulatory operational cost benchmarking study for transmission agents. The calculation of these costs is carried out using Data Envelopment Analysis (DEA), with operational cost as input and eight variables as explanatory variables (drivers) of costs. However, due to the strong linear correlation among the explanatory variables, a traditional regression model may generate inconsistent values for the estimated coefficients. An alternative approach to address this issue involves the utilization of linear programming and inequalities constraints to estimate the model parameters. In this context, this work proposes a methodology that combines linear regression with linear programming and cross-validation techniques to evaluate the main drivers of operational costs for Brazilian electricity transmission companies. The results indicate that the linear programming model proved to be the most suitable, with a predictive coefficient of determination of 0.80, and that only five out of eight explanatory variables were identified as the primary drivers of operational costs. This suggests redundancy in the current methodology for calculating costs, which could potentially compromise efficiency estimates.