Synthesizing exogenous scenarios: A distributional approach to macroeconomic risk
Tobias Adrian, Domenico Giannone, Matteo Luciani, and Mike West analyze how can externally specified macroeconomic scenarios be combined and evaluated against a reference predictive distribution to produce a coherent assessment of macroeconomic risks.
The authors take the baseline and alternative scenarios as given inputs. Each externally constructed scenario is represented as a conditional predictive density, while a separate statistical, survey-based, or judgmental density summarizes the broader distribution of possible outcomes. Scenario Synthesis then chooses weights on the scenarios so that their mixture resembles this reference distribution as closely as possible.
The method therefore evaluates, weights, and combines scenarios; it does not endogenously generate the scenarios or identify the structural shocks behind them. The weights are usually measures of each scenario’s relative contribution to matching the reference distribution, not literal occurrence probabilities.
Findings
- Applying the framework to Federal Reserve scenarios shows that externally constructed scenarios vary in how successfully they span the risks found in predictive distributions.
- In 2007, the scenario set did not adequately capture the severe downside risks identified by the New York Fed’s reference distribution.
- In 2018, the scenarios captured the reference risks considerably better, although their inferred importance depended on the chosen reference distribution.
- Considering GDP growth and inflation jointly changed which scenarios were most informative.
- Residual disagreement between the synthesized and reference distributions can reveal missing, insufficiently severe, or redundant scenarios and guide future scenario design.

Risks and Uncertainty in Monetary Policy
Authors: Tobias Adrian, Domenico Giannone, Matteo Luciani, Mike West
From: International Monetary Fund, Johns Hopkins University, Federal Reserve Board, Duke University
Generating endogenous scenarios: A structural approach to macroeconomic risk
Maximilian Schröder analyzes how structural shocks and prevailing economic conditions jointly determine future macroeconomic scenarios, their probabilities, and associated tail risks.
Empirically, Schröder asks whether a geopolitical shock comparable to that preceding Russia’s invasion of Ukraine would recreate the same inflationary and stagflationary risks under the economic conditions prevailing at the end of 2025.
The author develops a regime-dependent mixture VAR in which scenarios are generated endogenously rather than specified in advance. Economic conditions determine the probability of entering different regimes, while identified structural shocks affect both outcomes within each regime and the likelihood of moving between regimes.
The framework uses heteroskedasticity, sign restrictions, and narrative restrictions to identify structural shocks. Simulations then trace how each shock changes expected outcomes, the full predictive distribution, and the probabilities of policy-relevant scenarios such as recession and stagflation.
Findings
- The same shock can generate very different macroeconomic risks depending on the economy’s starting conditions. A geopolitical shock calibrated to January 2022 produces a more contained risk profile and a smaller rise in stagflation risk, when applied to conditions prevailing in December 2025.
- The results show that scenarios cannot be understood solely from the size or type of shock. Their paths and probabilities emerge jointly from the shock, the initial economic environment, and endogenous transitions between macroeconomic regimes.
Multivariate uncertainty and distributional transmission
Author: Maximilian Schröder
From: BI Norwegian Business School
How do the two papers compare:
