| Summary: | 碩士 === 國立屏東科技大學 === 企業管理系碩士班 === 92 === This paper tries to use a grey forecasting model GM(1,1) on improving the classical capital asset pricing model (CAPM). And we want to eliminate the paradox of the minus intercept and slope appearing in the empirical research we sample from the MSCI World Index. We try to establish a more stable and correct forecasting model by using the GM(1,1).
Using the Theil’s U in the empirical study, the GM(1,1) decrease 57.63% error than the classical moving average model. All of the 22 countries’ variances of estimated betas of the MSCI World Index are obviously decreased.
Besides, another part of this paper focuses on the seasonality in security market. Seasonality is supported by a lot of scholars after the claim of efficient capital markets by Fama (1970). Much phenomenon of empirical results are not consistent with the claim of efficient capital markets, for example, the January effect, the turn-of-the month effect, the day-of-the-week-effect and the holiday effect. The issue of this paper is the day-of-the-week-effect; we get whitening beta by using the GM model and then compare those data with the actual beta to know if the whitening process can make the betas steady. The author uses the Bartlett test, the SUR, one-way ANOVA, Scheffe test, and the correlation test. The empirical result of this thesis shows that the whitening data from GM(1,1) can do that.
|
|---|
