Abstract:
The main aim of this paper is to provide forecast intervals for inflation and unemployment rate in Romania, bringing methodological novelties in the construction and evaluation of the prediction intervals. Considering the period 2004-2017 as forecast horizon, only few intervals included the registered values on the variables, but in the last stage when all the prior information has been used, the forecast intervals were very short. The proposed Bayesian technique for assessing prediction intervals was better than traditional approaches based on statistic tests.