On the application and use of DSGE models

To put our claim in context, we describe the 'core/non-core' approach used in the Bank of England's Quarterly Model, and compare it with two others: the 'measurement error' approach of Ireland and the 'shocks-in-parameters' approach utilised by Smets and Wouters and others. Using a mock forecast scenario for illustration, we argue that each of these approaches would present model users with difficulties in communicating with policymakers. We conjecture that it is this problem, not theory or fitting the data, that currently hinders more widespread influence of DSGE-type models on policy-making.