The Implementation of Scenarios using DSGE Models

The new generation of dynamic stochastic general equilibrium (DSGE) models seems particularly suited for conducting scenario analysis. These models formalise the behaviour of economic agents on the basis of explicit micro-foundations. As a result, they appear less prone to the Lucas critique than traditional macroeconometric models. DSGE models provide researchers with powerful tools, which allow for the design of a broad range of scenarios and can tackle a large range of issues, while at the same time offering an appealing structural interpretation of the scenario specification and simulation results. This paper provides illustrations of some of the modelling issues that often arise when implementing scenarios using DSGE models in the context of projection exercises or policy analysis. These issues reflect the sensitivity of DSGE model-based analysis to scenario assumptions, which in more traditional models are apparently less critical, such as, for example, scenario event anticipation and duration, as well as treatment of monetary and fiscal policy rules.