Fan chart or Monte Carlo simulations for assessing the uncertainty of inflation forecasts in Romania?

Abstract

The objective of this paper is to make a comparison between two methodologies used to assess the forecasts of uncertainty: a numerical method based on Monte Carlo simulations and a graphical representation represented by the fan charts. For the inflation rate predictions made for Romania over the period Q4:2012– Q4:2013, a fan chart based on BVAR models with non-informative priors presents a lower degree of uncertainty compared with a fan chart using VAR models. The numerical procedure is based on forecasts that use auto-regressive models and the Monte Carlo method. In this case, the probabilities that the inflation forecasts are greater than the National Bank of Romania’s (NBR’s) target and the previous value increased from a quarter to the next. Therefore, this method of assessing the forecast uncertainty is a better tool than the fan charts. Moreover, the simple NBR methodology that did not take into account the probability distribution of the forecasts should be replaced by the fan charts. The forecast’s uncertainty assessment is necessary for the establishment of monetary policy.