Estimating DSGE Models: Recent Advances and Future Challenges
We review the current state of the estimation of DSGE models. After introducing a general framework for dealing with DSGE models, the state-space representation, we discuss how to evaluate moments or the likelihood function implied by such a structure. We discuss, in varying degrees of detail, recent advances in the field, such as the tempered particle filter, approximated Bayesian computation, the Hamiltonian Monte Carlo, variational inference, and machine learning, methods that show much promise, but that have not been fully explored yet by the DSGE community. We conclude by outlining three future challenges for this line of research.
We thank David Childers and Peifan Wu for useful comments and Sara Casella for superb research assistance. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research.