A Study of the Best Nested Linear Models for Hedge Funds Performance under An Out-of-sample Test

• Main Authors: Fong-Ci Jhuang, 莊豐綺
• Other Authors: 許英麟
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/78371314153063339039

碩士 === 國立中興大學 === 統計學研究所 === 101 === The aim of this study is to find the best nested linear model for each strategy from the HFR Hedge Funds database. The datasets are chosen during July 1995 to December 2012, including 210 months in total. This study improves the models in references which are highly used by researchers, including the one-factor model by Brown et al. (1999), the 7-factor model by Fund and Hsieh (2004), the MSCI Emerging Market Index by Fung and Hsieh(2001) and 11-factor model by Capocci and Hubner (2004). We employ the Calhoun OOS test to investigate if any of the factors from other models has to be added to each model so as to significantly improve its out-of-sample forecasting ability.