Análise comparativa dos modelos de cálculo dos retornos anormais utilizando o evento recompra de ações na Bovespa

• Ano de defesa: 2011
• Autor(a) principal: Gratz, Deborah de Souza Neves
• Orientador(a): Lima, Alvaro Vieira
• Banca de defesa: Silva, Adolfo Henrique Coutinho e , Carvalho, Frederico Antonio Azevedo de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade do Estado do Rio de Janeiro
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ciências Contábeis
• Departamento: Centro de Ciências Sociais::Faculdade de Administração e Finanças
• País: BR
• Palavras-chave em Português: Estudo de evento; Recompra de ações; Eficiência de mercado
• Palavras-chave em Inglês: Event study; Stocks Repurchase; Market Efficiency
• Área do conhecimento CNPq: CNPQ::CIENCIAS SOCIAIS APLICADAS::ADMINISTRACAO::CIENCIAS CONTABEIS
• Link de acesso: http://www.bdtd.uerj.br/handle/1/8066

The objetive of this study is to verify if there is any difference regarding the statistical significance level on the abnormal return calculation done through four statistical models used in event studies, having as the study object companies negotiated on the Bovespa Stock Exchange from March 2003 to July 2010. Considering the abnormal return calculation relevance in event studies, as well as the possible variations of results due to the calculation model of choice, this study uses an overly known theme, the stock repurchase announcement done by the own company. The methodology used in this research was quantitative, using the transversal cut study. The results demonstrated that there is difference among the significance levels found. When analyzing the graphs of the calculated models on the event window period, it was verified that the companies that repurchased the stocks done so when their shares were with abnormal accumulated negative returns and, after repurchasing, the shares had abnormal accumulated positive returns. The two models that uses the Ibovespa in its calculation formulae were recalculated, through an non-weighted Ibovespa and the conclusion pointed by the results is that the use of weighted market indexes must be avoided, giving preference to the use of portfolios composed by one share for each company that is comprised in the control portfolio. After this recalculation, it was verified that the models using the weighted market return models were graphically distant from all the other models. Lastly, the empirical evidences have demonstrated that the Brazilian capital market adjusts in a timely manner the shares of the companies that undergone the repurchase of their stocks, in line with what is expected by the efficient market hypothesis on its semi-strong form.