The prediction of inflation in Romania in uncertainty conditions

Based on data of inflation forecasts provided quarterly by the National Bank of Romania, forecast intervals were built using the method of historical forecast errors. Forecast intervals were built considering that the forecast error series is normally distributed of zero mean and standard deviation equal to the RMSE (root mean squared error) corresponding to historical forecast errors. We introduced as a measure of economic state the indicator– relative variance of the phenomenon at a specific time in relation with the variance on the entire time horizon. For Romania, when inflation rates follows an AR (1), we have improved the technique of building forecast intervals taking into account the state of the economy in each period for which data were recorded. We consider really necessary the building of forecasts intervals, in order to have a measure of predictions uncertainty.