CFCM Seminar: Ivan Petrella (University of Warwick)

Structural scenarioa analysis with SVARs

Abstract: Macroeconomists seeking to construct conditional forecasts often face a choice between taking a stand on the details of a fully-specified structural model or to rely on empirical correlations from vector autoregression (VAR) models and remain silent about the underlying causal mechanisms. This paper proposes to construct "structural scenarios" using identified structural VARs that can be given an economic interpretation.  We provide a unified and transparent treatment of conditional forecasting and structural scenario analysis and relate our approach to entropic forecast tilting. We provide for a careful treatment of uncertainty, making the methods suitable for density forecasting and risk assessment. We also propose an approach to assess and compare the plausibility of alternative scenarios. We illustrate our methods with two examples: comparing alternative monetary policy options and stress testing the reaction of bank profitability to an economic recession.