| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 3.31 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
O objetivo deste trabalho é testar se o modelo CAPM de fator único, que melhor descreve a relação entre risco e retorno, é válido na bolsa de valores portuguesa, tendo para isso uma amostra de dezoito ativos pertencentes ao índice de mercado PSI-20 para o período de janeiro de 2010 a dezembro de 2014. Para tal, será necessário estimar e analisar o risco sistemático (ou beta) dos ativos. Para aumentar a precisão dos betas estimados, realizamos o nosso estudo com carteiras de ativos, e não apenas com títulos individuais.
Os resultados obtidos sugerem que, para o período em análise, e para ativos individuais, não parece existir uma relação clara entre um maior risco e um maior retorno. De acordo com os resultados obtidos relativamente à estabilidade do beta dos ativos ao longo do tempo, podemos concluir que o nível de risco sistemático não é estável. Os valores dos betas dependem do intervalo de tempo usado para os cálculos dos retornos e do número de observações usadas na análise de regressão. De forma semelhante, no caso de carteiras de ativos concluímos também que não existe uma relação entre um maior risco e um maior retorno. Temos até resultados contrários: a um retorno maior corresponde um beta menor. Por outro lado, a consideração de carteiras de ativos reduz a amplitude dos betas estimados e aumenta o poder estatístico da regressão linear.
Perante comparação entre os modelos CAPM para carteiras e para títulos individuais e as respetivas expectativas de mercado (valor real), advém que o modelo CAPM das carteiras tem uma aplicação mais vasta ao objeto de estudo do que o modelo CAPM dos ativos individuais. Os dois são validados parcialmente, isto é, para certas empresas e anos observa-se que os modelos têm uma aplicação válida.
The objective of this study is to test whether the CAPM single factor, that best describes the relationship between risk and return, is valid in the Portuguese stock exchange, using a sample of eighteen assets belonging to the PSI-20 market index for the period from January 2010 to December 2014. We will estimate and analyze the systematic risk (or beta) of the assets. To increase the accuracy of the estimated betas, we conducted our study both with portfolios of assets, and with individual titles. Results suggest that, for the period under review, and for individual assets, there does not appear to be a clear relationship between a higher risk and a higher return. According to the results obtained concerning the stability of beta over time, we are able to conclude that the level of systematic risk is not stable. Beta values depend on the time interval used for the calculation of the returns and the number of observations used in the regression analysis. Similarly, in the case of asset portfolios, we concluded that there is not a clear relationship between an increased risk and a higher return. Results are contradictory: a higher return matches a lower beta. On the other hand, the consideration of asset portfolios reduces the interval amplitude of the estimated betas and increases the statistical power of the linear regression. When comparing the CAPM models for portfolios and individual securities, and the respective market expectations (the realized values), we come to the conclusion that the CAPM model has a more reliable and adequate application to portfolios than the CAPM model of individual assets. Both are partially validated however, that is, for certain companies and years, the models yield valid results.
The objective of this study is to test whether the CAPM single factor, that best describes the relationship between risk and return, is valid in the Portuguese stock exchange, using a sample of eighteen assets belonging to the PSI-20 market index for the period from January 2010 to December 2014. We will estimate and analyze the systematic risk (or beta) of the assets. To increase the accuracy of the estimated betas, we conducted our study both with portfolios of assets, and with individual titles. Results suggest that, for the period under review, and for individual assets, there does not appear to be a clear relationship between a higher risk and a higher return. According to the results obtained concerning the stability of beta over time, we are able to conclude that the level of systematic risk is not stable. Beta values depend on the time interval used for the calculation of the returns and the number of observations used in the regression analysis. Similarly, in the case of asset portfolios, we concluded that there is not a clear relationship between an increased risk and a higher return. Results are contradictory: a higher return matches a lower beta. On the other hand, the consideration of asset portfolios reduces the interval amplitude of the estimated betas and increases the statistical power of the linear regression. When comparing the CAPM models for portfolios and individual securities, and the respective market expectations (the realized values), we come to the conclusion that the CAPM model has a more reliable and adequate application to portfolios than the CAPM model of individual assets. Both are partially validated however, that is, for certain companies and years, the models yield valid results.
Descrição
Tese de mestrado em Matemática Aplicada à Economia e Gestão, apresentada à Universidade de Lisboa, através da Faculdade de Ciências, 2016
Palavras-chave
CAPM Beta Risco sistemático PSI-20 Regressão linear Teses de mestrado - 2016
