Submitted by Sandra Pereira (srpereira@ufrrj.br) on 2016-09-20T14:34:45Z No. of bitstreams: 1 2011 - Rodrigo Alves de Melo.pdf: 647767 bytes, checksum: 56dbf52b2dc0c6c7d87ccf57e757a793 (MD5) / Made available in DSpace on 2016-09-20T14:34:45Z (GMT). No. of bitstreams: 1 2011 - Rodrigo Alves de Melo.pdf: 647767 bytes, checksum: 56dbf52b2dc0c6c7d87ccf57e757a793 (MD5) Previous issue date: 2011-03-21 / This dissertation studies the performance of brazilian mutual funds stock portfolios classified as ?multimercado macro? from april, 2005 to march, 2010 by a multicriteria manner. The research aims to evaluate: (i) the success of market timing and stock picking strategies; (ii) the performance persistence taking into account macroeconomic various periods and (iii) if persistence depends on the time period analyzed, the performance indicator used or other variables such as size, client kind or investment profile. Therefore, were calculated the risk measures beta and standard deviation, the performance measures average return and cumulative return and the stock picking measure Fama?s decomposition for six months, one year and two years and six months periods and the stock picking measure Jensen's alpha and the market timing measure proposed by Treynor and Mazuy for two years and six months periods. The statistical tools used to verify the influence of control variables on performance and whether there is an association between the results were the Mann-Whitney?s test and Spearman?s correlation coefficient. The sample exhibits survivorship bias since it includes only the mutual funds that were actives throughout the studied period. Most studied portfolios failed to outperform the market by using stock picking and market timing strategies. Significant influences of the control variables size and investment profile were found, but only concerning the market timing measure. In the performance persistence analysis, only the systematic risk beta showed persistent evidences throughout the studied period, since the managers couldn?t keep the total risk under control during the financial crisis. In general, the portfolios have been unable to persist in any of the performance indicators, except for the unauthorized-operating-leveraged portfolio group, which showed persistence in average and cumulative returns and in selectivity indicator Fama?s decomposition, but only in long-term. Finally, DEA (Data Envelopment Analysis) enabled the creation of a portfolios? efficiency ranking taking into account its cumulative returns and systematic and total risks and only one portfolio showed maximum efficiency for the studied period. / Esta disserta??o estuda o desempenho das carteiras de a??es dos fundos de investimento da categoria multimercado macro no Brasil no per?odo de abril de 2005 a mar?o de 2010 de forma multicriterial. O estudo busca avaliar: (i) o sucesso dos gestores em estrat?gias de market timing e stock picking; (ii) a persist?ncia de performance levando em considera??o per?odos macroecon?micos diversos e (iii) se a persist?ncia depende do per?odo de tempo analisado, do indicador de performance utilizado ou de outras vari?veis, como tamanho, p?blico-alvo ou perfil de investimento. Para tal foram calculadas as medidas de risco beta e desvio-padr?o, as medidas de desempenho retorno m?dio e retorno acumulado e a medida de seletividade decomposi??o de Fama para os per?odos de seis meses, um ano e dois anos e seis meses e as medidas de seletividade alpha de Jensen e de market timing proposta por Treynor e Mazuy para per?odos de dois anos e seis meses. Os instrumentos estat?sticos utilizados para verificar a influ?ncia das vari?veis de controle no desempenho e a exist?ncia ou n?o de associa??o entre os resultados foram o teste de Mann-Whitney e o coeficiente de correla??o de Spearman. A amostra apresenta vi?s de sobreviv?ncia pois compreende apenas os fundos que estiveram ativos durante todo o per?odo estudado. a maioria das carteiras de a??es estudadas n?o conseguiu superar o mercado atrav?s da utiliza??o de estrat?gias de stock picking e market timing. Foram verificadas influ?ncias significativas das vari?veis de controle tamanho e perfil de risco, mas apenas em rela??o ? medida de market timing. Na an?lise de persist?ncia de performance, somente o risco sistem?tico beta apresentou evid?ncias de persist?ncia durante todo o per?odo estudado, uma vez que os gestores n?o conseguiram manter o risco total sob controle durante a crise financeira. De um modo geral, os portf?lios n?o conseguiram apresentar persist?ncia de performance em nenhum dos indicadores de desempenho, com exce??o do grupo de portf?lios n?o autorizados a operarem alavancados, o qual apresentou persist?ncia somente no longo prazo para os indicadores de retornos m?dio e acumulado e de seletividade decomposi??o de Fama. Finalmente a t?cnica DEA (An?lise Envolt?ria de Dados) possibilitou a composi??o de um ranking da efici?ncia das carteiras levando-se em considera??o o retorno acumulado, o risco sistem?tico e o risco total das mesmas e somente uma delas apresentou m?xima efici?ncia para o per?odo estudado.