The Building Of Forecasts Intervals

In this paper I built forecasts intervals for the inflation rate in Romania, using the quarterly predicted values provided by the National Bank of Romania for 2007-2010. First, I used the historical errors method, which is the most used method, especially by the central banks. 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. I introduced as a measure of economic state the indicator d-relativevariance of the phenomenon at a specific time in relation with the variance on the entire time horizon Then, I calculated the relative volatility in order to know the change that must be brought to the root mean squared error in order to take into account the state of economy. Finally, I proposed a new way of building forecasts intervals, when the date series follows an autoregressive process of order 1. In this case the length of forecasts interval is smaller and I got a slightly higher relative variance. I consider really necessary the building of forecasts intervals, in order to have a measure of predictions uncertainty, which is quantified by the National Bank of Romania using the prediction intervals based on a simple methodology. I calculated the forecasts intervals using MAE (mean absolute error), the indicator chose by National Bank of Romania and the MSE (mean squared error) indicator.