Breaks in DSGE models

In this paper, we investigate the consequences of breaks in mean for the estimates of model parameters and the resulting inferences about the structure of the economy and policy implications. We explore the behavior of DSGE models that undergo occasional, but permanent shocks to the parameters that determine their steady state. We present a method for approximating the state space representation of the model that can be used for simulation and estimation using recently developed techniques for analyzing non-linear structural breaks. We apply our method to a simple RBC model in which the trend growth rate is subject to infrequent stochastic breaks and conduct two Monte Carlo experiments. First, we verify that our estimation technique, based on the Kalman …lter, is able to correctly uncover the parameters of the model when the econometrician is aware that some parameters undergo infrequent stochastic shifts. Second, we show that ignoring the breaks in parameters can lead to over-estimation of the persistence of the other shocks driving the model. Our results suggest that, if actual time series data are driven by breaks in parameters and econometricians ignore this, then they may mistakenly conclude that the data are well explained by a DSGE model driven by very persistent shock processes. We explore this issue by estimating our RBC model on US and UK data. This views in this paper are those of the authors and do not necessarily re‡ect those of the Bank of England’s Monetary Policy Committee or the International Monetary Fund.