在產業垂直分工的架構下，各國股市之間的關聯性勢必存在。從彼此股市之間的報酬率相關性，可了解任兩國股市之間相互影響的程度與長期間此種影響的變化。許多研究直接針對股市的報酬率研究，固然得到具共整合的結論，可是卻難以觀察彼此之間關聯性的高低。因此，本研究主在探討半導體產業間的資訊傳遞效果，透過VAR-EGARCH模型之應用，其研究發現，在不同期間，資訊傳遞效果和預測績效的因果關係都會有所變化。亦即既便國際資本市場存在著高度的共整合現象，但仍會受到各股市之特質性因素，如管制、解除管制、系統風險、市場規模與投資人理性等因素所影響，甚至於各國產業在國際上的地位及重要性的消長與轉變都將影響著此種傳遞效果，建議可再從產業分工的角度更廣泛而深入地討論。

Given a note to the structure of vertical disintegration carrying out by industries, a relation between stock markets among nations is always perceived. It can tell, viewing the relativity between return rate of stock markets, the degree the inter-influences between stock markets and the changes in those influences in the long run. Accordingly, a number of return rate studies directly derived a conclusion of having co-integration, but failed to observe how profound their connection is. Therefore, this study aims to examine the information spill-over effect in semi-conduct industry through the application of VAR-EGARCH model. Our results show that information spill-over effect and cause-effect relation in performance prediction would change as time changes. It means, even an obvious phenomenon of high co-integration exists in the international capital markets, information spill-over effect remains variation when factors such as regulation, deregulation, systematic risk, market scale and investor's rationale have changed, suggesting a call for a comprehensive investigation should be drawn.