Publication details

Comparison of prediction power of BVAR and DSGE models



Year of publication 2014
Type Article in Proceedings
Conference Proceedings of 32nd International Conference Mathematical Methods in Economics
MU Faculty or unit

Faculty of Economics and Administration

Web Conference proceedings
Field Economy
Keywords steady state BVAR model; New Keynesian DSGE model; forecasting performance
Description This paper analyses macroeconomic development in the Czech Republic using structural steady state Bayesian VAR model, classical VAR model, and New Keynesian DSGE model for a small open economy. The steady state BVAR was developed by Villani (2009), DSGE model is borrowed from Justiniano and Preston (2004). Foreign sector in VAR models is implemented as block-exogenous. The models are estimated on data of Czech economy covering period 1996Q3 – 2013Q3. All models are compared on the basis of prediction performance measured by RMSE, impulse responses and variance decomposition. The results show that steady state BVAR has better prediction power for longer time horizons which is in accordance with its structure. On the other hand, DSGE model better predicts behaviour of variables for short periods, up to four quarters. According to RMSE computed for different horizons over the whole dataset, classical VAR model predictions were the least accurate in the majority of cases, DSGE model have the best predictive power for behaviour of foreign variables, and steady state BVAR overcomes the other models in prediction of interest rates in domestic economy.
Related projects:

You are running an old browser version. We recommend updating your browser to its latest version.

More info