本研究旨在運用Diebold and Yilmaz（2009, 2010）波動外溢效果（volatility spillovers effect）衡量方法，計算各國（美國、加拿大、英國、德國、法國、日本、新加坡、香港與韓國）在全球金融危機期間，股票市場相互影響之外溢指標強度與擴散方向，樣本期間取自2005年1月至2012年12月之日資料。本研究波動外溢指標乃利用向量自我迴歸（vector autoregressive, VAR）模型架構下的一般化預測誤差變異分解（generalized forecast error variance decomposition）來計算市場彼此之間外溢效果之動態變化，除了修正傳統VAR模型會受到變數排列順序的影響外，並可具體闡述當市場受到外在衝擊時對其他市場影響之總外溢效果（total spillovers）與市場彼此之間影響的方向性外溢效果（directional spillovers），不僅能提供金融預警系統功能，亦有助於監控與追蹤危機發展進度。

This paper carries out the methodology suggested by Diebold and Yilmaz (2009, 2010) to measures daily volatility spillovers effect between stock markets during global financial crisis for nine countries, including American, Canada, England, Germany, France, Japan, Singapore, Hong Kong and Korea, from the period January 2005 to December 2012. Volatility spillovers measure based on generalized vector autoregressive framework in which forecast-error variance decompositions are invariant to variable ordering, the generalized VAR approach enables us to understand how much of shocks to total volatility spillovers and the direction of volatility spillovers across major asset. It provides early warning systems for emergent crisis, and to track the progress of extant crisis.