Structural scenario analysis with SVARs

Macroeconomists constructing conditional forecasts often face a choice between taking stand on the details of fully-specified structural model or relying correlations from VARs and remaining silent about underlying causal mechanisms. This paper develops tools for economically meaningful scenarios with VARs, proposes metric to assess compare their plausibility. We provide unified treatment forecasting scenario analysis, relating them entropic tilting. A careful uncertainty makes our methods suitable density risk assessment. Two applications illustrate methods: assessing interest-rate forward guidance stress-testing bank profitability.