Optimal Asset allocation with Business cycle-Compare Classification and Regression Trees with Markov Chain
| Autor: | Yi-Chen Chung, 鍾宜臻 |
|---|---|
| Rok vydání: | 2010 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 98 Before Markowitz demonstrated the portfolio theory, investors only considered the risks and returns of individual securities, instead of considering the entire risks and returns. Markowitz (1952) demonstrated that investors make investment decision not only considering returns but also the risk factors and reducing the risk of investment by diversifying the investment objectives. Brinson (1991) demonstrated that approximately 91% of the portfolio returns would be affected by asset allocation. The stock selection strategy and timing strategy exhibits little effect on portfolio returns. Therefore, the main concept of this study is to investigate how investors conduct the asset allocation to obtain the optimal investment performance. Because numerical financial data belongs to nonlinear relationship, instead of linear, the macroeconomics variables are not suitable to assess by the linear model. The CART (Classification and Regression Trees) fits to analyze the nonlinear data. Therefore, this study adopts the CART model subsuming macroeconomics variables on asset allocation. Besides, Chen (2000) reveals that the Markov-regime switching model forecasts business cycle on asset allocation and improves the investment performance of asset allocation effectively and apparently. This study also applies the Markov-regime switching model to macroeconomic variables to establish asset allocation. Moreover, for the aspect of business cycle, this study not only applies the CART model but also the Markov-regime switching model to asset allocation. Furthermore, this study will compare the difference between CART and Markov-regime switching model to assist investors to select the best asset allocation model of investment performance. The empirical results in the study indicate that considering a business cycle and applying CART model on asset allocation could assess and improve the investing performance of CART model; considering a business cycle and applying Markov-regime switching model on asset allocation could assess and improve the investing performance of Markov-regime switching model. Comparing the investment performance of Classification and Regression Trees and Markov-regime switching model, the investment performance of the CART is better than that of the MS switching model. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|