先前關於匯率波動對出口貿易量的影響之文獻多以線性計量模型進行分析，然而，綜觀目前既存的文獻，其關係並無一致的結論。因此，本研究以貿易小國—臺灣與貿易大國—美國為研究對象，探討 1990 年 1 月至 2003 年 12 月期間，匯率波動對各國出口貿易量的非線性之影響。首先，利用自我迴歸條件異質性變異模型 (autoregressive conditional heteroscedasticity, ARCH model) 與一般化自我迴歸條件異質性變異模型 (generalized autoregressive conditional heteroscedasticity, GARCH model) 分別衡量新台幣實質有效匯率指數和美元實質有效匯率指數的波動程度，最後再採用 Teräsvirta (1994) 所提出的平滑轉換門檻誤差修正模型 (smooth transition error correction model, STECM) 來探討出口函數由短期動態調整至長期均衡關係的非線性過程。實證結果顯示在臺灣的 LSTECM 模型中，匯率波動除了對出口貿易量存在非對稱性的影響之外，亦同時具有正向和負向的關係；而在美國的 LSTECM 模型中，匯率波動對出口貿易量則產生正向和負向的線性影響。此外，由實證結果亦可發現貿易小國—臺灣的出口貿易量相對於貿易大國—美國的出口貿易量受匯率波動的影響較為敏感。 There are numerous theoretical and empirical studies that analyze the effects of exchange rate volatility on trade volume of export by linear econometrics models. The theoretical and empirical results, however, are contradictory. This paper investigates the nonlinear effects of exchange rate volatility on the export volume of the small trading nation (Taiwan) and the big trading nation (U.S.) during the period from January of 1990 to December of 2003. In this research, we first use conditional variances of the real effective exchange rate series modeled as an autoregressive conditional heteroskedastic (ARCH) and a generalized autoregressive conditional heteroskedastic (GARCH) process to measure the exchange rate volatility. We then adopt the smooth transition error correction model (STECM) to examine that the short-run export function adjustment toward the long-run equilibrium is a nonlinear process. The empirical results indicate that the nonlinear effects of exchange rate volatility on the export volume are asymmetric and there is a positive and negative relationship between them in the logistic smooth transition error correction model (LSTECM) in the case of Taiwan. Furthermore, the exchange rate volatility is positively and negatively related to the export volume in the LSTECM model of U.S. The empirical results also show that the export volume of the small trading nation (Taiwan) is more sensitive than the big trading nation (U.S.) in terms of the impact of exchange rate volatility.
