Evaluation of the Performance of Combined Methods in Real-Time Forecasting of Inflation in Iran

Inflation rate is one of the most crucial macroeconomic variables and accurate real time forecast of it is of great importance to policy making institutions, especially central banks. One of the methods that have been proposed to increase the accuracy of forecasts is making use of information content of an extensive set of diverse variables using forecast combination methods. In this paper, some of the forecast combination techniques are implemented to forecast Iran’s inflation rate and then evaluated in real time. Our results show that firstly, simple combination methods have better performance compared to the combination methods that estimate the optimal weights. Secondly, ignoring the correlation between simple forecasts-when estimating the weights-will increase the accuracy of forecasts. Thirdly, the combination methods with optimal weights beat the methods with shrinking weights. Finally, the performance of the methods with shrinking weights is improved by increasing the shrinkage factor.