MODELING AND FORECASTING THE EXCHANGE RATE IN ROMANIA

Econometric modeling of the exchange rate saw successive progresses, the forecasts based on the ‘70s models having a rather good accuracy, as recent researches showed. In order to explain the monthly evolution of RON/USA exchange rate during 2007-June 2011, I used three econometric models: a simultaneous equations model, an autoregressive model of order 1 and a model respecting the sense of Granger causality. From the statistical analysis of forecasting accuracy for one-month-ahead forecasts for July and August 2011 based on these models I found that the best predictions are those based on the model that is compatible with the sense of Granger causality. The higher errors are those of the forecasts based on the AR(1) model. The importance of knowing the best exchange rate forecasts is related to the improvement of decision-making and the building of the monetary policy.