Empirical Test of Multi-Factor Models in the Taiwan Stock Market

서지정보

· 발행기관 : 아시아.유럽미래학회
· 수록지 정보 : 유라시아연구 / 13권 / 4호 / 33 ~ 48페이지
· 저자명 : 최문섭, Seth Huang, 전예리

초록

Large number of studies have been performed to model asset returns in terms of risk factors. After the emergence of modern portfolio theory, asset pricing models have been evolved from single factor to multi-factors. Along with development of the models, validation of models with various portfolios and regions continued in wide range. However, researches in emerging Asia-Basin markets still need considerable amount of regional modeling and verification. Also, limited number of studies were conducted on industry sorted portfolio returns and more recent multi-factor models in particular. Therefore, this study focuses on validation of several market models in Taiwan stock market.
This paper applies empirical tests with several well-known multi-factor models on portfolios sorted by industries in Taiwan stock market. Sample data includes all listed stocks collected from Taiwanese Stock Exchange (TWSE) from July 2001 to December 2015. Following Moerman (2005), industry portfolio is constructed based upon TWSE’s industry group classification codes per FTSE codes. To compare effect of each model in different portfolios, capital asset pricing model (CAPM), Fama and French’s (1993) 3 factor model, and Carhart’s (1997) 4 factor model are used in the study.
In previous studies of the Taiwan stock market, it is known that market premium (MP) factor shows significant influence on return and relatively weak support to the effects of “high-minus-low” value premium (HML) and “small-minus-big” size premium (SMB) factors (Ko, 2014; Cheng, 2006). However, effect of the momentum factor (“up-minus-down.,” UMD) needs more validation. Also, as Fama and French (1997) discussed in U.S. market, returns of portfolios constructed in respect to industry sector is not very accurate when compared to other portfolio returns. Comparable results were observed in the European markets (Moerman, 2005). Industry sorted returns in emerging Asia-Basin markets were mostly done in limited extent and need further research.
Empirical results from various time period samples have shown some differences between industries exist when different factor model is tested in Taiwan market. When whole period sample is used, influence of MP factor prevails over other factors. SMB, HML and UMD factor show significance in certain industries. In 5-year basis period sample, 2001-2005 and 2006-2010 periods show results similar to whole period, but period from 2011 to 2015 deviates from formal results with significant UMD factor in number of industries. However, similar to result of Fama and French (1997), there is no evidence of clear winner on different models in all time periods in terms of model fitness. Overall, consistent with findings of Cheng (2006), MP factor has significant influence in explaining industry return. Effects of other factors are limited with weak influence.

참고자료

· 없음