Table 3 reports our maximum likelihood estimates of the real exchange rate according tothe system of Eqs. (2)–(4). We first-differenced the data in Eq. 2, except inflation so as toinduce stationarity and eliminate spurious regression bias. Note that the variables are expressed in terms of growth rates, using log-differenced dataon RER, productivity, producer prices, and oil prices (all multiplied by 100). The results arereported with six lags. The choice of lag length is based on the Akiake criterion.12 As notedabove, we include two dummy variables in the mean equation to account for the effect ofthe Russian crisis in August 1998 and the switch to a floating exchange rate system in April1999.The results reported in Table 3 indicate that the productivity growth rate has astatistically significant impact on real exchange rate movements, but with a significant lag(recall that an increase in the real effective exchange rate indicates an appreciation of thetenge). Productivity growth lagged by five periods is significantly associated with anappreciation of the domestic currency in the current one. These results reflect theimportance of including lagged variables in estimations aimed at capturing the dynamicadjustment of real exchange rate to changes in such explanatory variables as productivity
đang được dịch, vui lòng đợi..