本文以動態追蹤資料迴歸分析方式探討1975至2010年間,以中國官方2005年宣布的管理浮動釘住的11個國家對美元貨幣為樣本,重新估計人民幣失衡情形。應變數以購買力平價說(PPP)為基礎的Penn World Tables資料考量的相對購買力平價之實質匯率,自變數納入文獻探討關於BEER/FEER內外部均衡的重要解釋變數,建構其動態追蹤資料模型。另外,我們比較用傳統計量方法所得出的近期人民幣失衡狀況與本文主要的動態追蹤資料迴歸分析法(Dynamic Panel GMM),2010年分別為顯著高估31.3%與低估52.2%。本文以考量總體經濟理論因果關係與持續性效果的模型,所用落後一期的實質匯率納入研究方法較傳統方法能解決變數之內生性問題所造成的統計偏誤,故較合理且有效。且兩種方法的正負符號結果並無差別,支持了本文使用動態追蹤資料迴歸分析法(GMM)對實質匯率的解釋有效,甚至在各變數顯著性上優於傳統回歸方法。研究結果綜合如下三點。 (一)關注實質匯率具有持續性的效果上,實質匯率落後一期估計係數在各模型皆為正且顯著(0.61~0.82)。(二)人口、政府政策與金融發展之變數,在扶養年齡結構與政府營運為正相關;唯貨幣供給為負相關。以資本管制與制度之資本帳公開性指數與國家清廉皆與實質匯率成正相關。經常帳盈餘與實質匯率呈負相關。(三)匯率失衡部分,本文用動態追蹤資料法估計失衡值,2010年低估52.2%。
This thesis estimates China’s equilibrium exchange rate and misalignment since China announced a shift in exchange rate policy from a simple peg to the US dollar to a basket peg about 11 countries by using the dynamic panel GMM estimators with the period of 1975-2010. We estimate the China’s equilibrium exchange rate based on using the purchasing power parity from Penn World Tables with the consideration of Balassa-Samuelson effect, and the others variables from the theory defining equilibrium exchange rate, including BEER/FEER. Compared panel data (fixed/random) model with dynamic panel GMM Estimators, we find Renminbi is significantly overvalued by 31.3% and undervalued by 52.2% .We concern real exchange rate with not only the persistent but also cause and effect with other Macroeconomic variables.We consider the results of exchange rate misalignment using dynamic panel GMM estimators solving endogeneity problems by lag term of real exchange rate are more reliable and significant. With long term estimation, sign dosen’t change across model specifications. We summarize our findings as follows: 1. The coefficients of lagged real exchange rate are positive (0.61~0.82). 2. Real exchange rate have positive relationship with government operating balance(0.0047~0.0099),and negative relationship with money supply(-0.03~-0.068);have a positive relationship with capital account openness(0.007~0.03); Current account with negitive relationship (-0.003~-0.004) 3. With dynamic panel GMM estimators ,we find Renminbi is significantly undervalued by 52.2% .