Forecasting Iran's Economy Inflation with DSGE-VAR Model (Theory and Technique)

Inflation is one of the key macroeconomic variables that its precise forecasting is the goal of policy makers and in particular the central bank.VAR model has a long history as a tool for forecasting and policy analysis; but the problem with this method is that it uses a little theoretical information about the relationships between variables. In addition, in VAR models, many parameters need to be estimated, some of them may be meaningless. Following this, the idea of hybrid models was introduced. One of the hybrid models is the DSGE-VAR model. This model combines DSGE, which is a structural model and more reliant on theory, with a VAR that provides better fitting data. In this study at the beginning, the theoretical structure and results of the estimation of the DSGE-VAR model for Iran's Economic Data are presented. Further, the forecasting of this method are compared with other models such as unrestricted VAR and Minnesota VAR. All three models are estimated recursively from the period 1991:1 to 2012:4 and then used to forecast inflation for one to eight-quarters-ahead over an out-of sample from 2011:1 to 2015:4. Comparison of the accuracy of forecasting of the above methods using the RMSE index shows a better performance of the DSGE-VAR approach compared to other models.