| 摘要: | 本研究主要以Fama and French(2015)與Hou, Xue, Zhang(2015)分別提出的因子訂價模型，探討其解釋台灣上市櫃公司股票報酬率的有效性。透過兩模型對於因子定義之差異，以Hou, Xue, Zhang(2015)與Fama and French(2015)研究方法，使用公司季財務報表公告數字建構規模因子、盈利能力因子、投資水平等因子，分別測試台灣上市櫃公司全體樣本與不含金融業樣本兩種樣本，並透過迴歸分析法，以alpha值顯著個數做為標準判斷模型之間有效性之優劣。另外，為使研究更趨於完善，本研究亦加入Fama-French-Carhart四因子模型與資本資產訂價模型(Capital Asset Pricing Model，以下簡稱CAPM)，同時納入學界在系統性風險、生產者理論、實證結果觀察三大研究方向，探討其在台灣股市中有效性之差異。
本研究蒐集1990年至2019年間台灣上市櫃公司之財務報告數字，以alpha值顯著個數、GRS test等方法進行分析。研究結果顯示：第一，不論在樣本是否含金融業，以及是否使用更新後財報公告日樣本，CAPM模型均是表現最佳的資產訂價模型，與Fama and French(2015)和Hou Xue and Zhang(2015)兩篇文獻的結論相異。第二，在全樣本時間區段中，Fama-French五因子模型在含金融業與不含金融業的樣本中表現均僅次於CAPM模型，優於Hou六因子模型。第三，在加入淨值市價比因子(HML)後，Hou Xue and Zhang六因子模型和Fama-French五因子模型均能提升模型解釋能力，與Fama and French(2015)的結果相異。
;In this paper, I study the effectiveness on models from Fama and French(2015) and Hou, Xue, Zhang(2015) in Taiwanese listed company respectively. Based on the difference in the definition of the profitability factor, I separately examine two samples, which are all listed firms and all listed companies excluding financial firms in order to study the connection between factors and stock returns in two models. Using multiple regression analysis, I rank models based on the amount of significant alpha in 74 portfolios. Moreover, for the completeness of the research, I also consider Fama-French-Carhart four factors model and CAPM.
The research period is from 1990 to 2019 and my findings are as follow. First, no matter using which announcement date version or containing financial firms or not, CAPM performs the best, which is contrary to both Fama and French(2015) and Hou, Xue, Zhang(2015). Second, during all-sample period, the Fama-French five factors model ranks after CAPM, but it is superior to the Hou et al. six factors model. Third, after adding HML factor, the performance of the Fama-French five factors model and the Hou et al. six factors model are better, which is opposite to the conclusion of Fama and French(2015). |
