On the Persistence Tests of Hedge Funds’ Abnormal Returns under Different Factor Models ─ the Bayesian Approach versus the OLS Approach

• Main Authors: Wei-Tse Hsiao, 蕭瑋澤
• Other Authors: Ying-Lin Hsu
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/v4hu9m

碩士 === 國立中興大學 === 統計學研究所 === 100 === The main thrust of this study is to test the performance of hedge funds reported by Hedge Fund Research, Inc. The HFR hedge funds dataset spans a period from June 1994 to May 2011, and contains a total of 8920 hedge funds. We apply the ordinary least square method and the Bayesian approach to identify hedge funds with statistically significant abnormal returns under different factor models, and test whether the abnormal returns persist in the future. Furthermore, we compare these two methods to see which one is more powerful. We find that the Bayesian approach is more predictable than the OLS method for relatively short sample periods. Our results are robust to the selection of the cutting point dividing the out-of-sample period from the in-sample period. In addition, Fung and Hsieh’s seven-factor model performs best among all factor models investigated.