Apreciação dos índices de qualidade na precificação de ativos no mercado acionário brasileiro

Abstract

The central thesis of this study is that assets that present better quality indicators, such as higher operating margin, are valued by the market with a premium over the others. Thus, the main objective of this dissertation is to verify the existence of premiums for the quality factor in the Brazilian market and its effectiveness in explaining the return of the assets. This study analyzed seven quality metrics widely used by the market, all focused on profitability or operational productivity. There was moderate to strong correlations between metrics evaluated, as well as the absence of historical data, particularly between 1995 and 2007. Thus, remained on the following analyzes less correlated measures and with greater data availability. Due to the difficulty comparing quality metrics from different companies and specially industries, it was decided to limit the estimation universe of the study to industrial companies, mainly because of the number of companies and the industry representativeness. Overall, the study used all industrial stocks traded in Bovespa from January 1995 to June 2018. In order to investigate the existence of the quality factor premium, the present study used the same methodology relied by Fama and French (1992), Carhart (1997) Liu (2006), among others. This is based on the creation of two portfolios, elaborated based on the characteristics that one wishes to calculate the premium, each with the highest and lowest stipulated quality metric. The portfolios were rebalanced periodically in order to maintain their desired characteristics. The quality factor premium is then calculated as the excess return over the risk free asset of the difference between the cumulative return of the two portfolios (High Quality minus Low Quality). The means comparison test - one sample t test - was then used to analyze the statistical significance of the factor premium. The quality factor premiums, calculated based on four different metrics, proved to be robust and statistically significant at 1%. Then, the two-stage methodology of Fama and MacBeth (1973) was adopted to verify the explanatory power of the CAPM model and the proposed 2-factors (market beta and quality factor). The sample period was then divided into two subsamples, the ex ante, used to calculate the sensitivities between control portfolios and market premiums, and the ex post¸ used to verify the explanatory power of the proposed model. The ex ante subsample comprised the period from January 1995 to December 2006 while the ex post subsample was from January 2007 to June 2018. For this stage, only the ROE quality metric was maintained, given its robustness in the previous analyzes. The two-factor model, including the quality factor, proved superior to the CAPM model, presenting a 60% higher explanatory power, as measured by Adj. R2. The proposed 2-factor model was not sufficient to explain the cross sectional return in the period analyzed, but it seems to indicate an important direction for future research on the topic.