The Feasibility Study of Improving Performance of GA’s Portfolio with Tracking-Error Constraints

• Main Authors: Lin, Chen-Li, 林烝立
• Other Authors: Shih, Chun-Hsieh
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/74454895858559451781

碩士 === 樹德科技大學 === 金融保險研究所 === 95 === Variance is the critical factor to decide the portfolio performance. If we efficiently control the volatility of risky assets, we can do well in asset allocation and further make the investment policy for investors. Jorion (2003) explores the risk and return relationship of active portfolios subject to a constraint on tracking error volatility (TEV), constructed by four equity indices which included the U.S., U.K. Japan, Germany, and the Lehman Brothers U.S. Aggregate Bond Index. It shows that this method can improve the performance of the active portfolio by adding a constraint on total portfolio volatility in TEV-constrained portfolios. The paper is to explore whether changing portfolio weights by the TEV method can improve the performance on the constraint of not changing the portfolio constituents. This research constructs a risky asset portfolio related with TWSE index (TAIEX) by genetic algorithm, and then uses the TEV method to calculate portfolio weights. Finally, this thesis compares the performance between them. The empirical period is from 2005/01/03 to 2006/12/29, including 492 daily data. The result shows that The TEV-constrained Portfolio has a higher return related to GA’s Portfolio, and the correlation coefficient between portfolio and TAIEX is above 0.96. Furthermore, its performance is also better than Taiwan 50 Index.