Stochastic Data Envelopment Analysis applied to the 2015 Brazilian energy distribution benchmarking model

The Brazilian energy regulator has been applying data envelopment analysis to set operating costs for distribution system operators since 2013. In addition to data envelopment analysis, further adjustments using Bootstrap and a reference efficiency, estimated as 79%, allow companies to have cost efficiencies above their observed costs. Thus, companies are allowed to cross the efficiency frontier. Similarly, stochastic frontier models also allow companies to cross the efficiency frontier. This work proposes the use of stochastic data envelopment analysis as an alternative for estimating efficient costs, thus providing a much simpler alternative. A new estimation algorithm is proposed in which the number of companies crossing the frontier comprises one important parameter. Simulation studies provide convergence evidence of the proposed model and results using the Brazilian database show that the stochastic data envelopment analysis is a promising model for upcoming tariff review cycles.