Bayesian Estimation of a DSGE Model with Inventories

This paper introduces inventories in an otherwise standard dynamic stochastic general equilibrium model. Based on U.S. data we estimate the parameters of our model using Bayesian methods. We find that inventories enter the Phillips curve as an additional and significant driving variable and make the inflation process less backward-looking. Moreover, our analysis reveals that the impulse responses and the contribution of structural shocks to variations in the observable variables change significantly when allowing for inventories. In contrast to previous studies, we can attest only a subordinated role to technology shocks in explaining the forecast error variance of real variables.