5.就風險值因素的角度發現：無論在四因子或五因子模型，風險值對股票報酬均具有顯著的解釋力，與Bali and Cakici(2004)的實證結果一致。此即顯示VaR factor將可提供股市交易者在選股操作上另一客觀的參考指標。 Since Sharpe (1964) proposed the Capital Asset Pricing Model (CAPM), investors used CAPM as a standard approach to measure the return and risk. But there are some suggest that the CAPM can not explain the return anomalies existed in the market. The multi-factor model suggested by Merton (1973) and Ross (1976) and the common factors model proposed by Fama and French (1996) argued that the beta in CAPM cannot explain well of return and risk.
The explanation power of return can not capture by beta only. The previous return pattern will affect the future expected return distribution. From the market microstructure view of point, the trading volume (or turnover rate) can also affect the change of cross section return.
The paper analyzes the Taiwan 50 index stock from July 2, 1999 to June 30, 2004. We apply EWMA approach to estimate the Value at Risk (VaR) of these stocks. The Panel Data Analysis is also uses for empirical examination. From the basis of Fama-French''s three factors model, we add the VaR and Turnover rate to form the four and five factors model to analyze the return and risk in Taiwan''s stock market.
The empirical results show that: 1.The EWMA approach can estimate VaR efficiently from the back testing result.
2.From the three factors model: The relationship between return and market factor is not significant. The MV/BE factor is significant explanation for the return. This is consistent with the Fama-French mondel. But the market value factor is not significant.
3.From the four factors model: The market factor, MV/BE, and VaR are significant. The only insignificant factor is market value.
4.From the five factors model: The turnover rate has positive effect on the return. The MV/BE is not significant. The other factors are significant explanation power of return.
5.The VaR factor is significant is both four and five factors model to explain the return. This is consistent with the Bali and Cakici (2004) finding. This shows that VaR can be another reference index for stock choosing strategy.