本研究利用原油價格來探討十個國家的名目匯率預測能力之分析,建構出線性與非線性的油價預測模型,相對於隨機漫步(random walk)在樣本外之預測能力,研究期間為2004年1月到2017年4月的日資料和月資料,使用滾動樣本迴歸估計法(rolling regression)進行樣本外預測,並以相對RMSE和DM檢定衡量預測能力之優劣,實證研究結果發現,匯率的預測能力須依資料的頻率、預測期的比例、模型的設定和國家的選取而定,當使用日資料並且樣本內滾動視窗占總樣本的比例為1/10估計時,澳幣、盧布和南非幣油價預測模型都顯著優於隨機漫步模型;澳幣、英鎊、盧布和南非幣顯著優於匯率AR(1)模型。;This paper examines whether oil prices have significant predictive ability in forecasting nominal exchange rates out-of-sample, and we compare the linear and non-linear oil price models forecast with those of the random walk, which, to date, is the toughest benchmark to beat. We implement the Diebold and Mariano’s (1995) test of equal predictive ability by comparing the Root Mean Squared Error(RMSE) recorded in the daily and monthly data of ten countries between 2004 and 2017.
Under rolling regression analysis, empirical results indicate that predictability depends on forecast horizon, frequency of the data, model specification, and forecast evaluation method. Our empirical results suggest that oil prices can predict the nominal exchange rate at a daily frequency and small in-sample estimation window sizes. However, the predictive ability is not evident at monthly frequencies.