A Comparison of the Performances of the Direct and Iterated Methods in Real Time Forecasting of Inflation in Iran

Inflation rate is one of the key macroeconomic variables that policymaking institutions and central banks in particular, need to forecast accurately for several periods ahead in order to make proper policies. Direct and iterated methods are two common techniques which are suggested in the literature for multi-period forecasting. In this paper, using a wide range of quarterly economic variables we compare the performance of these two techniques in real time forecasting of inflation in Iran. The results show that as the forecast horizon increases, iterated method outperforms direct method. For the information criteria which select shorter lags (e.g. Schwarz criterion), direct method and iterated method performs better in short forecast horizons (1 and 2 periods ahead) and long forecast horizons (3 and 4 periods ahead), respectively, while for the information criteria which select longer lags (e.g. Akaike criterion), iterated method generally performs better, irrespective of the forecast horizon.