早有文獻指出系統風險並非固定,而是會隨時間變動而改變,本篇以Bollerslev(1990)所提出雙變量非均齊條件變異數模型(Bivariate GARCH Model)計算國內八大類股及其主要個股隨時間變化之系統風險。樣本期間自2007年1月2日至2008年12月31日,共有492筆日資料。為了證實系統風險的非固定性,本文分別以2007年7月27日、2008年5月19日及2008年9月15日為結構轉變點,並進行chow test探討八大類股及其個股之股票報酬率及系統風險是否有明顯轉變。結果顯示八大類股及個股其股票報酬率在2007年7月27日較無變化,而在2008年9月15日則有顯著轉變。其中八大類股及其個股之系統風險在這三個時間點大多皆有明顯轉變。
There exist evidences that show systematic risk is unstable and time varying. This paper applies Bivariate GARCH Model to estimate time-varying systematic risk (time-varying beta) using daily data in eight industries and sixteen individual stocks of Taiwan from Jan. 1, 2007 to Dec. 31, 2008. In addition, returns are compared with systematic risks in the case of different events. Results show that returns changes are totally different and most of the systematic risk switched from the event occurred. Finally, this paper examines some time specific structural changes of the returns and the systematic risk or not. The presupposed is that the fell of Taiwan stock market(Jul. 27, 2007), presidential inauguration(May. 19, 2008) and bankruptcy of Lehman Brothers(Sep 16, 2008) .The result demonstrates that returns have different outcome and most of the systematic risk almost changed.