근사 글로벌 경제 구조하에서의 달러화 환율 불균형 오차 조정과정 분석

This paper discusses on the analysis for the dynamics of disequilibrium errors in dollar exchange rates in an approximated global model. Our approach is based on a transformed error correction model which is derived from an observationally equivalent transformation in Kim(2008) and Kim and Park(2008). In particular, we multiply a nonsingular matrix to the left hand side of VAR (vector autoregressive) model to exploit a disequilibrium error dynamics. So the transformed model is composed of the equations representing autoregressive (AR) error dynamics and the partial error correcting equations. The AR nature of the dynamics enables us to discuss on the stability of disequilibrium error, while conventional error correction model is not possible to conduct it explicitly. The disequilibrium error is defined from the concept of Engle- Granger cointegration of which theoretical back grounds are based upon the monetary approaches for the determination of exchange rates. The variables included in the global model are dollar real exchange rates, short and long term interest rates, relative growth rates of industrial productions comparing with United States(US) and stock prices for the United Kingdom (UK), Euro area and Japan. The categories of models are classified following Lucas(1982), Bilson(1978), Frankel(1979) and model of Frankel with a stock price(say extended model). From standard Johansen and Shin(1994)’s tests, we can not reject the existence of a cointegration for the exchange rate of each individual country. So we estimate the long run cointegration relation by OLS(Ordinary Least Square) regression for these different models. The data frequency is monthly from January 1999 to August 2007. The CUSUM tests did not find any evidences of structural breaks during the suggested period. We found moneys are critical factors to determine the long run cointegration relation while the signs of coefficient estimators were consistent with the most theoretical expectations of monetary approaches. The disequilibrium error may be estimated from the above OLS regression and it may be exploited to construct an estimated AR error dynamics and error correcting equations of trade surplus changes and relative growth rates between a country and the US economy. According to the estimation of these short run dynamics, we found the disequilibrium errors of exchange rates significantly affect to the real sectors and vice versa. In particular, the changes of trade surpluses in UK and Euro area are affected by the disequilibrium errors of own country and especially the change of trade surpluses in the US is largely affected by the disequilibrium error of Euro area. Further, the relative growth rate of Euro area is heavily affected by the disequilibrium errors of Yen․Dollar and Pound․Dollar. The disequilibrium error of Pound․Dollar showed instability while the stability of disequilibrium error of Euro․Dollar is remarkable. These empirical results suggest that the policy makers should carefully monitor the disequilibrium errors as well as the long run equilibrium level of exchange rates for the stability of real sectors in economy.