Stability of the beta coefficient of electric energy industry

Abstract

This article is aimed at evaluating the stability of the beta coefficient over time, which is the usual risk indicator of stocks. The research work embraces common (ON) and preferred (PN) stocks of eighteen Brazilian corporations of the energy industry, which are classified as governmental or private ones, and refers to the period from January 1999 to June 2008. Therefore, there are four classes of stocks: governmental common, governmental preferred, private common, and private preferred. A three-stage analysis is developed. In the first stage, a regression analysis is performed to estimate beta coefficients according to the Market Model on the basis of consecutive sixty month periods, with a lag of six months between them. In view of the fact that estimates of betas show an oscillating pattern, additive and multiplicative dummy variables are introduced in the second stage, and new regression analysis is performed by means of the stepwise procedure. In the third stage, a Panel Data Analysis is developed initially by means of the stepwise procedure and then by means of the enter procedure. The results of the first stage show a clear oscillating pattern for the four classes of stocks. In the second stage statistical evidence of the instability of the betas for the four groups of stocks is found, and more unstable beta’s estimates are observed for governmental corporations. The results from the Panel Data Analysis confirm the beta’s capacity to differentiate portfolios and to capture changes in the economic and financial scenario.

Key words: risk, stocks, beta coefficient, Brazilian energy industry.