Testing market structure assumptions for DSGE modelling in Croatia using the SVAR model with long-run restrictions

Abstract

Assumptions on market structure are crucial in formulating dynamic stochastic general equilibrium (DSGE) models. The inclusion of the price stickiness assumption in DSGE models has questioned the money neutrality, which is a characteristic of DSGE models with perfect competition, and has thus opened the space for monetary policy analysis. One of the criteria used to determine which DSGE models are better suited to the characteristics of an observed empirical economy is the impact of technology shocks in the structural vector autoregression (SVAR) model with long-run restrictions. In DSGE models, assuming perfect competition with no price rigidity, i.e. real business cycle (RBC) models, changes in productivity that are driven by productivity shocks increase employment. On the other hand, new Keynesian (NK) models that assume imperfect competition and price rigidities, suggest that technological shocks decrease employment, since companies cannot adjust to excess production by reducing prices. In order to assess the impact of productivity shocks, the SVAR model with long-run restrictions is estimated using data on labour productivity, working hours and consumption in Croatia. The estimated impact of labour productivity shocks on working hours is statistically significant and negative, whereas the estimated impact on consumption is statistically significant and positive. The conducted analysis points to the fact that the DSGE model with imperfect competition and price rigidities describes the empirical characteristics of the Croatian economy better than the model with perfect competition.