Jan Babecký, Michal Franta, Jakub Ryšánek
In this paper we explore the potential of the DSGE-VAR modelling approach for examining the effects of fiscal policy. The combination of the VAR and DSGE frameworks leads theoretically to more accurate estimates of impulse responses and consequently of fiscal multipliers. Moreover, the framework allows for discussion about the differences of the effects of fiscal shocks in DSGE and VAR models and to some extent discussion about misspecification in fiscal DSGE models. The DSGE-VAR model is estimated on Czech data covering the period from 1996 to 2011 at quarterly frequency. The government consumption multiplier attains a value close to 0.4 at the horizon of four years. The public investment multiplier is about 0.4 higher, which confirms findings in the literature. On the other hand, the DSGE model alone implies a similar government consumption multiplier but a much lower public investment multiplier, suggesting misspecification of the fiscal DSGE model.
JEL codes: C11, E62, F41, H30
Keywords: DSGE-VAR model, fiscal multipliers, fiscal shocks, identification, model misspecification
Issued: November 2016