THE EVALUATION OF QUARTERLY FORECAST INTERVALS FOR INFLATION RATE IN ROMANIA

Abstract

The forecast uncertainty was one of the causes of the recent economic crisis and its evaluation became more necessary nowadays. The aim of this paper is to build and assess different types of forecast intervals for quarterly inflation rate in Romania. The Bootstrap Bias-corrected-accelerated (BCA) forecast intervals outperformed the intervals based on historical errors, four out of six values of inflation rate being placed in the first type of intervals during Q3:2013-Q4:2014. The likelihood ratio tests and the chi-square test indicated that there are significant differences between the ex-ante probability of 0.95 and the real probabilities for both types of forecast intervals. As a methodological novelty, Monte Carlo and bootstrap simulations were used for assessing the uncertainty in inflation rate forecasts in Romania.