Aplicação do modelo alternativo de três fatores no Brasil

• Ano de defesa: 2012
• Autor(a) principal: Silva Júnior, Claudio Pilar da
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraí­ba; Brasil; Administração; Programa de Pós Graduação em Administração; UFPB
• Palavras-chave em Português: Modelos de Precificação de Ativos; Anomalias; Investimento; Asset Pricing Models; Anomalies; Investment; CNPQ::CIENCIAS SOCIAIS APLICADAS::ADMINISTRACAO
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/tede/3814

This dissertation aimed to analyze how investment and ROA are priced and whether them partially explains change stock returns in the Brazilian stock market. Initially, aimed at analyzing whether an investment and ROA premium exists. Secondly, was aimed to compare the performance of alternative three-factor model of Chen, Novy-Marx and Zhang (2010), consisting of a market risk factor, the investment and ROA factors, with the CAPM model and three-factor model of Fama and French (1993), as well as investigate the robustness of the models on commonly known stock market anomalies. To development of the study, it was used stock portfolios and to verify the performance of the model in explaining the changes of stock returns were used a set of time series regression analysis. The population consisted of all non-financial companies with stocks traded on the Bolsa de Valores de São Paulo BM&FBOVESPA, from January 1995 to June 2011.Refering to the risk factors analyzed, it was observed an average market premium of 2,303% per month. With regards to the size and book-to-market factors, it was could not find evidence of them existence in the Brazilian market, since we obtained a negative premium of 0,005% and 2,516% per month, respectively. With regards to the factors based on production, it was found for investment factor a positive and significant premium of 0,698% per month. When it comes to the ROA factor, it was obtained a positive premium of 0,263% per month, however, not statistically significant. In the analysis of portfolios formed by investment and ROA factor, it was expected that stocks with greater investment in assets tend to have lower returns than stocks with the lowest investment. This pattern can be observed, since seven of the nine portfolios formed by stocks lower investment achieved return over the portfolios formed by stocks that performed more investment in the same period, cannot rejected Hypothesis 1. Regarding the expected return, it was expected that the stock portfolios formed by high ROA submit superior returns to the returns of portfolios formed by stocks of low ROA. This pattern was observed in eight of nine portfolios formed, however, the nonexistence of a premium for the factor ROA causes the rejection of the Hypothesis 2. Comparing the three models by the adjusted R2 there was on average a superior model of Fama and French (1993) of 3.6% over the alternative model of three factors and 5.1% over the CAPM. It was observed also that the alternative model of three factors presented behavior similar of the model of Fama and French (1993) when the portfolios are sorted based on volume, momentum, leverage, EBITDA/P and PL.