| Summary: | 碩士 === 銘傳大學 === 財務金融學系碩士班 === 99 === While using the traditional mean-variance strategy to access asset allocation, the sensitive of expects return higher than variance. Therefore, using unreasonable condition expect return to estimate new expect return will lead to expand the estimate error. Next, Chris and Barbara (2009) use the data of return of mean and variance to prove the effects of asset allocation which are influenced by the turnover and total transaction cost.
Therefore, comparing the four models of Markowize of mean-variance optimal portfolio strategy, 1/N strategy, variance timing strategy and risk timing strategy which are used in this study and adding turnover and transaction cost to model. The research sample of return of asset allocation is using 19 categories of month return, market value and B/M ratio of stock index price. The research period in this study is from January 1996 to December 2010, including 15 years and 180 months. According to Fama and French (1993), this study classifies the portfolio into five categories. In order to modify the problems which arise from traditional Markowitz theory, this study uses the four models of assets allocation strategy to analyze the portfolio and finds out the optimal weight and assets allocation of portfolio.
The major goal in this study is to find out the optimal assets allocation model. The empirical results of this study show that whether in all sample or different moving window data, the tining strategy using β which is calculated by CAPM model instead of expect return is the optimal assets allocation strategy.
|
|---|
