我國遠期外匯市場重新開放後之效率性檢定

Efficiency of Taiwan's Forward Exchange Market: Evidence Since Market Reopening in 1991

何中達(Chung-Da Ho) ； 沈中華(Chung-Hua Shen)

《中國財務學刊》 3卷2期 (1996/01) Pp. 63-85

https://doi.org/10.6545/JoFS.1996.3(2).4

摘要

我國遠期外匯市場自1987年7月以後，由於中央銀行不再接受遠期外匯的拋補，且對銀行外匯部位的認定改為現金制而瀕於關閉，直到1991年11月才告重新開放。目前學者對我國外匯市場效率性之檢定，其研究期間均為重新開放之前，對遠匯市場重新開放後之研究則付之闕如，本文之目的即在探討重新開放後之遠匯市場是否較1987年之前更具有效率性。如市場效率性之假說成立，則遠期匯率為未來即期匯率的不偏預測值；且預測誤差（未來即期與遠期匯率之差）不具有序列相關性。在檢定上，由於即期與遠期匯率均為非定態（non-stationary）之時間序列，我們考慮其共整合（cointegration）之關係，並採用Johansen的最大概似法以檢定遠期匯率之不偏性。在檢定上，我們進一步以預測誤差之正交性來探討市場效率性。基於樣本期間僅一年餘，我們使用重疊之週資料，並採用一般化動差法（GMM）以修正資料期間重疊可能造成之偏誤。

並列摘要

Due to some regulations imposed by the central bank, the forward foreign exchange market in Taiwan has been inactive between July 1987 and November 1991. While the efficiency of Taiwan's forward market had been investigated by several researchers, none of the studies examines period after the market reopened in 1991. The purpose of this paper is to examine whether the efficiency of Taiwan's forward market has been improved since its reopening. If the market is efficient, then [1] forward rate is a unbiased estimator of future spot rate; and [2] forecast errors (the difference between future spot rate and forward rate) are serially uncorrelated. Since both spot and forward rates are non-stationary series, we consider their cointegration relationship and use Johansen's maximum likelihood method to examine hypothesis [I]. In testing hypothesis [2], we test the orthogonality of the forecast errors. Since the sample size is small, we use overlapping data and adopt GMM to eliminate possible bias.