Modelo estocástico de análise envoltória de dados aplicado ao benchmarking das empresas brasileiras de distribuição de energia elétrica no ano de 2015

Abstract

Due to the natural monopoly existing in the Brazilian electric sector, the elements that guarantee the protection of consumers against measures that disadvantage them are used, disadvantages such as the exorbitant high prices. One of these mechanisms is the existence of a regulatory agency, called the Agência Nacional de Energia Elétrica, ANEEL. Among its various functions, this agency carries out the analysis of parameters of electric energy production by the 61 companies throughout the national territory, called DSO (Distribution Service Operator). Currently, the parameters are used for benchmarking the DSO's, according to the Data Envelopment Analysis (DEA) model, where efficiency is obtained during the tariff review period. The current model also has an ad-hoc procedure, where bootstrap is used, allowing some DSOs to surpass the efficiency frontier. A new method is proposed, using the benchmarking process called Stochastic Data Envelopment Analysis, SDEA, which proposes a parametric equation for the efficiency frontier, and which naturally allows the DSO to cross the frontier, without the need for additional procedures. In the applied methodology, a new SDEA model was proposed, and simulations demonstrated convergence between the current model used by the regulatory agency and the proposed model, showing that SDEA can be a good option for benchmarking in the next tariff cycle.