外匯市場效率性檢定—以台灣為例

Abstract

本研究欲探討自1991年重新開放之後至今的台灣外匯市場是否具有效率性，採用新台幣兌美元的即期匯率與10天期、30天期、60天期、90天期，以及180天期的遠期匯率，匯率資料期間涵蓋了1991年12月11日至2004年12月31日間之日資料。 在研究方法方面，除了傳統的單根檢定之外，本研究亦加入panel單根檢定法進行檢測，以提高傳統單根檢定統計檢定力，並且避免產生不當結論，實證結果顯示傳統單變量單根檢定的結果與多變量的panel單根檢定之間存在差異。 單根檢定結果顯示，除了10天期的遠期溢酬為定態I（0）序列，其餘30天期、60天期、90天期與180天期的遠期溢酬皆為I(1)序列；Johansen共整合檢定結果顯示，30天期、60天期、90天期與180天期的遠期溢酬之間存在共整合關係；此外不同天期的遠期溢酬間，大多存在回饋的因果關係；衝擊反應檢測結果顯示，各天期的遠期溢酬並無法立即反應完畢。在遠期溢酬反應貨幣風險溢酬隨機結構的前提下，本研究的實證結果，不支持樣本貨幣風險溢酬為定態，從而不支持台灣外匯市場具效率性。

This paper tried to figure out whether the foreign exchange market in Taiwan is efficient or not after the market reopened in December, 1991. Thus, we examine panel of forward premium series including 10, 30, 60, 90 and 180 days forward contract maturities. The contradictory findings in the literature may reflect the well-known limited power of conventional unit root tests against stationary alternatives in small samples. Therefore, in addition to the conventional univariate regression used to test the weak-form efficiency hypothesis in the foreign exchange market, we applied the panel unit root test to forward exchange premiums by utilizing cross-sectional information from their term structure. There are unit root tests, Johansen cointegration tests, vector error correction models, Granger causality tests and impulse responses used to find out the existence of the long run and short run stability relationships in these variables. Our finding implicate that forward exchange premiums are non-stationary and given that the forward premium reflects the stochastic structure of the currency risk premium, our finding shows non-stationary behavior for the risk premia of NT dollar (relative to the US dollar), and thus cannot support the weak-form foreign exchange market efficiency under risk aversion.