臺灣對外貿易與實質匯率之關係研究

An Economic Analysis on Relationship between Exchange Rate and Taiwan Foreign Trade

陳婉華(Wan-Hua Chen)

指導教授 ： 何瓊芳

中原大學/商學院/國際貿易研究所/碩士(2008年)

https://doi.org/10.6840/cycu200800099

摘要

本研究針對臺灣對外貿易作實證模型，主要探討實質匯率的變動對國際收支餘額的影響，目的是了解臺灣的貿易狀況下，馬歇爾-勒納理論（Marshall-Lerner condition）模型是否成立，以及實質匯率的變動對國際收支餘額的影響是否存在J曲線效果。進一步以匯率的變動來預測國際收支餘額的變動幅度，用以縮減貿易的循環周期，進而提昇臺灣的貿易競爭力。

　　本研究先對各個變數做單根檢定，因為所選取的皆為經濟方面的時間序列變數，而時間序列變數必須呈現恆定狀態，因為在定態的時間序列，才符合漸進分配理論，其各項相關的檢定才能呈現其正確性。也由於經濟方面的時間序列變數，通常都具有非定態的性質，因此為了使其能真實的估計實證模型，就必須檢視是否存在「共整合現象」，即非定態的時間序列變數的線性組合變成定態的形式，來檢視變數是否真正存在因果關係，進而避免造成假性迴歸的情況產生。之後再以OLS（ordinary least squares）來做多元迴歸估計長期均衡迴歸模型。

　　實證結果顯示，在減少政府財政支出的同時也能使總支出水準降低，這也同時影響了國際收支餘額，使得國際收支餘額得到改善。且也因政府財政支出為落後一期，所以對國際收支餘額具有遞延的效果。

　　以貨幣分析法而言，實證結果無法符合貨幣分析法的理論，在實證下實質匯率（Mt）的係數為正數，無法達到政府在使用緊縮信用策略來調節市場供需時，所預期能調整國際收支餘額逆差現象的結果。而且貨幣供給量也沒原始設計模型預期的存有時間落後進行可影響國際收支餘額的現象。

　　實質匯率的部分，馬歇爾-勒納的理論模型在台灣的短期是不成立的。而實質匯率的落後期也因實證中的實質匯率係數有先呈負值而後轉換成正值的型態，顯現出實質匯率的變動情形對國際收支餘額的影響是存在J曲線效果的。而當新台幣貶值到第二季的時候，國際收支餘額才會出現改善的預期效果，且這種因貶值所產生出來的改善不論是短期亦是長期而言，效果均是相同的，都會使得國際收支餘額在相當長的時間內呈惡化現象，也因這些實證結果來做判斷的話，以台灣目前的經濟狀態下，新台幣的貶值是不會導致國際收支餘額趨於改善的，必須由整體經濟之基本面去著手才是。

關鍵字

實質匯率

並列摘要

This research takes Taiwan as the observed objective model. The study firstly discusses weather the real exchange rate of NT dollar may change the balance of international payment (BOP) with long-lasting influence. It then examines if Marshall-Lerner condition works or not in Taiwan. Further, the J-curve effect is also checked as well. Moreover, the exchange rate change is used to forecast the BOP and to analyze if it shorten the trade cycle period, so as to enhance Taiwan trade competitive ability.

A single root test on each variable was performed to make sure that the time series variable must fulfill the constant condition, because only the stationary state time series can be accurately predicted by the theory. The economic time series variable usually has the non-stationary nature. Therefore, in order to obtain the real estimated results, one must correct the non- stationary state time series and modify the variable into the stationary state so as to check the causality relation to avoid creating spurious regression. Ordinary least squares model was applied to analyze the long-term balanced regression.

According to the empirical results, a depreciation of NT dollar can reduce the government expenditure and aggregate consumption level. This simultaneously can also improve the BOP.

In terms of monetary analysis, due to all estimated coefficients of the real exchange rate (Mt) are positive, it is unable to achieve the governmental goal to reduce expenditure only by using the credit strategy to adjust the market.

As for the real exchange rate, Marshall-Lerner condition is not associated with Taiwan's short-term trade balance. However, there is a J-effect only for a very short time period, while not in the long run. Therefore, a depreciation of NT dollars does not affect Taiwan’s BOP. The improvement of BOP should be eventually made through enhancing the basic economic situation.

並列關鍵字

Exchange Rate

參考文獻

方文碩、張倉耀、葉志權（2005）。匯率貶值及其風險與出口。國立臺北大學經濟學系經濟研究期刊。第四十一卷第一期。頁105-139。

劉宗欣、張銘仁（2000）。進口物價的匯率轉嫁與不對稱性：台灣的實證研究。經濟論文。十二月。第二十八卷第四期。頁369-395。

Bahmani-Oskooee, M. (1985), Devaluation and the J-curve : some evidence from LDCs. Review of Economics and Statistics, 67(3), pp. 500-504.

Bahmani-Oskooee, M. (1992), What are the long-run determinants of the U.S. trade balance?. Journal of Post Keynesian Economics, 15(1), pp. 85-97.

Bahmani-Oskooee, M. (1998), Cointergation approach to estimate the long-run trade elasticities in LDCs. International Economic Journal, 12, pp. 89-96.