A carregar...
Publicação
Performance of return models : a portfolio theoretical approach
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.08 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
O objetivo desta investigação é avaliar o impacto das assunções dos modelos geradores de retornos nas fronteira eficiente e seus portafolios. Isto foi conseguido mediante o trabalho in-sample (assim eliminando o risco de estimação, focando a investigação no risco de modelo) em ambos mercados de ativos financeiros, o Europeo e Americano, nos 7 anos anteriores, considerando ambos, o caso em que shortselling esta permitido como também o caso em que esta proibido. O processo inclui o calculo da fronteira eficiente seguindo as assunções dos modelos geradores de retornos. Em particular o Constant Correlation Model (CCM), o Single-Index Model (SIM) e o modelo de tres factores de Fama e French (1993) Multi-Factor Model (MFM). Para os dois mercados de investimentos, comparamos a fronteira eficiente gerada aplicando MVT nos dados in-sample com a fronteira eficiente dos modelos de retorno seleccionados. Mostramos que o risco do model è importante na aplicação do MVT. Sendo os erros encontrados em todos os modelos consideravel. Também, considerando o risco de modelo para o caso de shortselling proibido, o CCM mostra melhor desempenho que os modelos mais sofisticados. Por outra parte, em condiçoes de shortselling permitido, o SIM mostra melhor desempenho.
The objective of this research is evaluating the impact of the return generating models assumptions in the efficient frontier and its portfolios. This is accomplished by working with in-sample data (eliminating the estimation risk and focus on the model risk) looking at both the European and American stock markets for the past 7 years, and considering both, the case when shortselling is allowed and the case when it is forbidden. The process includes the calculation of efficient frontier under the assumptions of return generating models. In particular, we look at the Constant Correlation Model (CCM), the Single-Index Model (SIM) and the three factors Fama and French (1993) Multi-Factor Model (MFM). For both markets we compared the true efficient frontier generated from the in-sample MVT with the corresponding efficient frontiers from the return generating models. We show model risk is an important issue when applying MVT. The errors in all model are considerable. In addition, considering model risk for cases when the short-sell is not allowed, the CCM is more accurate than more sophisticated models. On the other hand, under conditions of short-sell allowed, the SIM seams to be more accurate.
The objective of this research is evaluating the impact of the return generating models assumptions in the efficient frontier and its portfolios. This is accomplished by working with in-sample data (eliminating the estimation risk and focus on the model risk) looking at both the European and American stock markets for the past 7 years, and considering both, the case when shortselling is allowed and the case when it is forbidden. The process includes the calculation of efficient frontier under the assumptions of return generating models. In particular, we look at the Constant Correlation Model (CCM), the Single-Index Model (SIM) and the three factors Fama and French (1993) Multi-Factor Model (MFM). For both markets we compared the true efficient frontier generated from the in-sample MVT with the corresponding efficient frontiers from the return generating models. We show model risk is an important issue when applying MVT. The errors in all model are considerable. In addition, considering model risk for cases when the short-sell is not allowed, the CCM is more accurate than more sophisticated models. On the other hand, under conditions of short-sell allowed, the SIM seams to be more accurate.
Descrição
Mestrado em Finanças
Palavras-chave
Risco de modelo risco de estimação composição de carteiras Model risk estimation risk portfolio composition
Contexto Educativo
Citação
Frade, Carlos Augusto Zerpa (2017). "Performance of return models : a portfolio theoretical approach". Dissertação de Mestrado, Universidade de Lisboa. Instituto Superior de Economia e Gestão.
Editora
Instituto Superior de Economia e Gestão