本研究以臺、港、日三地外匯市場為研究對象,並應用VEC GJR-GARCH模型與Panel共整合模型分析即期匯率間的互動性、外匯市場效率性與遠期匯率在不偏性下即、遠期匯率的互動關係,樣本期間為2003年1月2日至2008年6月15日。 實證結果證明各外匯市場之即期匯率皆存在自我外溢效果、GARCH效果、ARCH效果;除此之外,納入日本遠匯升貼水做為資訊變數之實證結果發現日本遠匯升貼水對臺灣與日本即匯市場有顯著之解釋能力,且原本存在之波動外溢效果皆已消失,顯示日本遠匯升貼水確實為一資訊變數;在波動不對稱效果方面,實證結果發現臺灣與日本之不對稱估計係數為正向顯著,表示壞消息衝擊下,自身市場的波動會比好消息的波動幅度還要大。 在外匯市場效率性的實證結果方面,本研究發現三外匯市場皆通過遠期匯率為未來即期匯率的不偏估計值,使用Panel FMOLS所得出之結論亦不拒絕外匯市場效率性之假設。
This study investigates the interactions, volatility spillovers and efficiency for foreign exchange markets of Taiwan, Hong Kong and Japan by using VEC GJR-GARCH model and Panel cointegration approach. The sample period of this study is from January 2, 2003 to August 15, 2008. The empirical results prove that the various foreign exchange markets the spot rate of all have the own-mean spillover effects, GARCH effects and ARCH effects. In addition, the results also show that Japanese forward exchange premium (or discount) has significant effect on the spot exchange markets of Taiwan and Japan when incorporating the Japanese forward exchange premium (or discount) into the model, and volatility spillovers effects all disappeared. It confirms that Japanese forward exchange premium (or discount) is a mediating variable of information. In the asymmetrical volatility effect aspect, the empirical results discover that asymmetrical estimated coefficients are positive and significant in exchange markets of Taiwan and Japan. This expresses that under the bad news impact, own market's volatility will probably be bigger than the good news. Based on the results in testing foreign exchange market efficiency, it indicates discover that three foreign exchange markets all accept the forward rates as unbiased predictors of future spot rates. Finally, this study also does not reject the hypothesis for foreign exchange market efficiency by using Panel FMOLS