多重投資期間資本資產定價模式-台灣股票市場之實證研究

Abstract

本研究主要探討放寬傳統CAPM假設單一投資期間之多重投資期間模型：Jensen (1969)一因子對數模型、Lee (1976)模型及Lee、Wu與Wei (1990)模型，相較於傳統一因子CAPM對股價報酬跨期間變動是否更具解釋能力。另加入Fama與French (1993)之規模因子與淨值市價比因子，檢視兩因子加入後之上述模型變數間的關係是否更加清晰，以及解釋能力是否增加。研究對象為台灣上市公司股票，研究期間為2002年7月至2009年12月。實證結果顯示Jensen一因子對數模型的解釋能力略高於傳統CAPM，而Lee模型及LWW模型之整體解釋能力與Jensend模型幾無差異，故考慮投資人多重投資期間之Jensen一因子對數模型可作為傳統一因子模型的替代模型。另外，發現加入二因子後之各別模型差異仍與加入前相似，惟其解釋能力均大幅提升，顯示除了市場報酬外，規模因子與股價淨值比因子仍為解釋台灣股票市場報酬關係之重要變數。

This paper examines three alternative functional forms of the CAPM, which include the investment horizon parameter, to see if their explanatory power will increase, comparing to the original CAPM. To make the relationship of the parameters more clear, SMB and HML factors of Fama-French three-factor model are added to all of the four models. The sample period is from July 2002 to December 2009. The empirical results show that the explanatory power of the heterogeneous investment horizon functional forms are indifferent from that of the original CAPM. In addition, SMB and HML factors should be included in the models for estimating the systematic risk.