Estimating the efficiency of Brazilian electricity distribution utilities

ABSTRACT

This paper proposes a differing methodology from the Brazilian Electricity Regulatory Agency on the efficiency estimation for the Brazilian electricity distribution sector. Our proposal combines robust state-space models and stochastic frontier analysis to measure the operational cost efficiency in a panel data set from 60 Brazilian electricity distribution utilities. The modeling joins the main literature in energy economics with advanced econometric and statistic techniques in order to estimate the efficiencies. Moreover, the suggested model is able to deal with changes in the inefficiencies across time whilst the Bayesian paradigm – through Markov chain Monte Carlo techniques – facilitates the inference on all unknowns. The method enables a significant degree of flexibility in the resultant efficiencies and a complete photography about the distribution sector.

Keywords:

• Bayesian inference
• Gibbs sampler
• scale mixture of normal distributions
• state-space models
• Stochastic frontier analysis

Disclosure statement

No potential conflict of interest was reported by the author(s).