The Reduction of Uncertainty in Making Decisions by Evaluating the Macroeconomic Forecasts Performance in Romania

The evaluation of macroeocnomic forecasts performance does not include only the calculating of some statistical measures, rather controversial in literature, like root mean squares error or absolute mean error. In theory and economic practice, three directions have been traced regarding the evaluation of forecasts performance: the analyse of accuracy, bias and efficiency. Using the forecasted values on medium run of inflation rate and unemplyment rate through theperiod from 2004-2010 in Romania, we get a better degree of accuracy and a lower efficiency for forecasts made by National Commission of Forecasting comparing to those based of Dobrescu model used by Institute of Economic Forecasting. Following the international tendency, the forecasts are, in all cases, biased because of difficulties in precise anticipation of shocks which affect the economy. Forecasts performance is indestructible related by their uncertainty, RMSE, the measure of evaluating the accuracy being used in building forecast intervals based on historical errors. For forecasted values of inflation rate published by National Bank of Romania we propose a new way of building forecast interval in order to take into account the economic shocks.