Abstract
This study is based on the evaluation of the long-run performance of the monetary model approach of exchange rate determination for the newly entered EU members and Turkey. First, we tested the cointegration relationship between exchange rates and monetary variables. Then, the forecast estimates of the monetary approach were used for comparing the performances with the random walk model.
Notes
1Variables in countries other than Latvia are found to be nonstationary I(1). Also, in Lithuania, monetary differential is found stationary I(0). Thus, we tested the cointegration relationship only for variables exchange rate and output differential. Unit root test results are available upon request.
2Wald tests for DOLS and ARDL show that in all countries, except Lithuania (DOLS), the test results are statistically significant.
3Therefore, we excluded Czech Republic, Cyprus and Malta where appeared to be stationary.
4There is no autocorrelation or heteroscedasticity and the errors are distributed normally.