Does Term Spread Predict the Swings in Exchange Rate?

本文參考Chen (2011) 之研究方法, 以12 個已開發國家的月資料為樣本, 對匯率升貶值之趨勢進行預測。本研究以期間利差作為關鍵解釋變數設立模型。由於期間利差對市場資訊反應極為迅速, 我們預期加入期間利差可以增加模型的解釋能力。為了比較期間利差對於模型解釋能力的影響,我們進一步考慮不放入期間利差的模型。此外我們亦關心期間利差相對於利率所增加的解釋能力,因而進一步建立一個以利率取代期間利差的模型。除了自設模型外, 本文亦使用貨幣模型、購買力平價模型以及考慮利率平滑的泰勒法則模型作為比較。儘管模型對於短期以及長期的匯率趨勢預測, 無法於所有國家皆具最佳的解釋力, 但實證結果顯示, 加入期間利差的模型其整體表現較佳。本文的研究結果可對於未來匯率趨勢的預測提供可靠的參考意見。This paper follows the emprical models from Chen (2011), using monthly data from 12 developed country to predict the swings in exchange rate (major trends in depreciation or appreiation). In our research, term spread is treated as the key variable to construct the empirical model. Because its sensitivity to the market information is quite strong, it would expected to increase the explanatory power of the exchange rate swing. We also consider a model without term spread to compare the marginal effect, and a model with interest rate instead of term spread. In addtion to our model, we also apply monetary model, purchase power parity model and Taylor rule model with interest rate smoothing for robustness. Although our model do not being very suceccful at all horizon on all country, however, our empirical result shows the term spread model performs better on long-horizon forecasting. Out result provides a reliable prediction for future exchange rate trends