| Summary: | 碩士 === 淡江大學 === 財務金融學系碩士在職專班 === 99 === The study applies the Mean-Variance Model and Mean-CVaR to construct optimal weighted portfolios comprising stocks used in the TSEC Taiwan 50 Index, Hong Kong’s Hang Seng Index, and China’s CSI 300 Index. The purpose of this paper is to review the performance portfolios and find the optimal weights. The stock selection strategies in different regions showing a significant difference between Taiwan, Hong Kong and China.
Geometric mean, arithmetic mean, Cumulative Return, Sharpe Ratio, Treynor Ratio, Jensen''s Alpha and the Information Ratio are used to review the performance portfolios in different regions. When the confidence level of 95%, M-CVaR model is superior to M-V model; when the 99% confidence level, the empirical results were inconsistent.
However, if observed in all investment regions, the number of wins by the rate of return or cumulative number of comparisons shows that the rate of return, when the confidence level of 95%, M-CVaR model is superior to M-V model, but the M-CVaR or the MV model is superior Index performance is not necessarily. If the performance from three of the total number of investment strategy choice model, when the confidence level of 95%, M-CVaR model is superior to M-V model and M-CVaR model is also better than the market benchmark index performance.
By investigating the average divergence with t-test, we analyzed whether or not the average return showed statistical significance between these two models. The results revealed that the performance of different investment strategies only evinced statistical significance from the benchmark index in the case of the average return on investment in Taiwan stocks.
In conculsion, our investigation provide investors with recommendations for their investment portfolios. If investors wish to invest in Taiwan, it is recommended that they use the M-CVaR model with a confidence level of 99%. If investment in China or Hong Kong stocks is desired, choosing an index-bound approach will be appropriate.
|
|---|
