總體經濟基本面的預測表現，台灣與其他六國匯率模型之實證分析

The Prediction Performance for Macroeconomic Market Fundamentals An Empirical Analysis of Exchange Rate Models for Taiwan and Other Six Countries

實務上，匯率是否確實為隨機漫步走勢，一直頗具爭議。除了難以由特定模型預測外，學界也習見匯率在實證上與總體經濟基本面反倒脫勾的看法。有鑑於此，本文擬套用財務領域常見的資產訂價模型，期能找出匯率與總體經濟基本面的潛在關聯。實證結果顯示，匯率對於市場基本面變數大致享有高度的預測成效，足以進一步強化兩者仍互具影響力的可信度。然而，比較特殊的是，在台灣的案例中，經濟基本面與匯率卻相對缺乏顯著的因果關係。我們初步推測，相較本文研究的其餘部分國家，台灣貨幣當局可能在政策執行層面，存在較高程度的非市場干預力量，導致台幣匯率在解釋基本面變數上，得不到較明確的結果。

In empirical study, whether nominal exchange rates behave like random walk or not, it remains controversial nowadays. Excepting the difficulty in model forecasting, some researchers argued that the nominal exchange rates are disaffiliated with macroeconomic fundamentals empirically. Accordingly, on the basis on asset-pricing theorem used in financial field, we want to figure out the potential relationship between nominal exchange rates and macroeconomic fundamentals. The empirical results show that the nominal exchange rates performed well in forecasting macroeconomic fundamentals, which enhanced the credibility of interactive affection among these variables. However, especially in the case of Taiwan, we found less evidence on the significance of Granger causality between the variables mentioned above. Basically, comparing to some of other 6 countries studied in this work, we supposed that Taiwan monetary authorities adapted more non-market intervention strength than over countries, thus we can't obtain the significant result in Taiwan.