原油股票指數對匯率之預測

Abstract

本文檢視原油相關的股票價格指數在短期內對於原油出口國之匯率是否具有預測能力。在日資料頻率下，透過樣本內以及樣本外預測的方法檢視發現，原油股票指數模型可以用來預測未來 1 期的匯率，且其預測能力優於隨機漫步模型，但而對於更長期數的預測，或是在月資料頻率下的預測中，原油股票指數模型並不具有預測未來走勢的能力。這樣的結果顯示原油股票指數能快速反應原油市場上的資訊，而這些資訊能反映出短期匯率的走勢，使得原油股票指數可用以預測短期匯率，但在長期下這些資訊已經過時，而造成長期的匯率預測結果並不顯著。

In this paper we investigate the ability of oil sensitive stock in predicting daily exchange rate of oil-exporting countries. Using in-sample and out-of-sample tests, our model beat the random walk model in one-day-ahead pre- diction but not in longer horizons prediction. The result indicates that the oil sensitive stock has a quick response to the change in oil market, providing useful information to predict short-run exchange rate movement. However, information contained in oil sensitive stock is out of date in the long-run and hence the exchange rate prediction over long horizons are not statistically significant.