金融危機之國際投資市場訊息傳遞效果研究

隨著全球資金流動的效率提高，市場資訊透明度增加，單一國家金融危機的發生對於國際金融市場間產生之連動效果成為學者關切的議題。近來一些研究指出，危機之所以很快傳遞到其他國家，除了存在外溢效果之外，同時存在的蔓延效果亦是主要的原因之一。本研究為了解台灣與美國股價連動性，及個股股價與其發行美國存託憑證關聯性，以Enron破產事件為劃分時點，採用向量自我迴歸(VAR)模型，探討樣本公司的美國存託憑證價格、普通股價格與台灣加權股價指數、那斯達克指數在重要金融危機後股市間的訊息傳遞效果。研究結果發現，以Enron破產事件為例，在重要金融危機前後，上述價格序列間均有長期穩定之均衡共整合關係；而在誤差修正檢定中，那斯達克指數對於台灣加權股價指數具有引導效果。此外，台美股市間呈現雙向回饋關係，投資人不僅考慮美國各大股價指數變動做為分析台股決策依據，亦考慮台股波動對於美股之影響。此外，雖然台積電與聯電都佔台灣加權股價指數很大的比例，但波動上也無法擺脫台灣加權股價指數影響，此結論可作為支持資產定價模型理論之實證。

關鍵字

金融危機；國際金融；國際投資；向量自我迴歸

並列摘要

With the increasing efficiency of capital flows around the globe and the enhancement of market information transparency, the impact of single-country financial crisis to international financial markets has become researched by many scholars. Some studies indicated that in addition to the spill-over effect, the contagion effect is also a major cause attributing to the quick spread of a crisis to other countries. In order to gain an understanding of the correlation of equity prices in Taiwan and the U.S. and the correlation between stocks and their corresponding ADRs, this paper adopts the VAR model by referring to the bankruptcy of Enron as the watershed to examine the signalling effects of ADR prices and ordinary share prices of sampled companies after major financial crises between TAIEX and NASDAQ Index. The results indicate that in the case of Enron's collapse, the abovementioned price series all exhibit long-term equilibrium cointegration relationships before and after the major financial crisis. According to error correction test, NASDAQ Index has leading effect on TAIEX. Meanwhile, there is a two-way feedback relationship between the equity markets in Taiwan and the U.S. Investors not only consider the equity indexes in the U.S. as the basis for analysis of investments in Taiwan's equities, but also the influence of the fluctuation of Taiwan's equities on the U.S. stocks. Although both TSMC and UMC account for a large portion of TAIEX, they are not immune to the fluctuations of TAIEX. This conclusion can serve as an empirical finding to support CAPM.