Towards a method for guiding monetary policy in times of great uncertainty

Chart 1 – The optimal monetary policy path differs depending on the scenario chosen

Monetary policy decision-making by central banks is generally underpinned by a baseline macroeconomic projection scenario, corresponding to the most likely expected trajectory for real GDP and inflation. This approach has recently been criticised by Bernanke (2024) in his report to the Bank of England. According to the winner of the 2022 Nobel Prize in economics, relying on a single scenario to guide monetary policy is both insufficient and inappropriate in times of uncertainty. Bernanke recommends reducing the importance given to the baseline forecast in monetary policy discussions and communications, and giving greater visibility to alternative scenarios. The 2025 strategic review of the European Central Bank (ECB) also highlighted the usefulness of alternative scenarios to round out the baseline projection scenario in times of great uncertainty, in order to assess alternative monetary policy paths.

That said, one question remains regarding scenarios: when several economic scenarios are considered, each implying a different monetary policy path, how can the diversity of these scenarios be adequately taken into account to guide action?

Rather than treating all scenarios equally or focusing solely on the baseline scenario, this blog post proposes a (statistical) approach that consists in assigning each scenario a specific weighting, reflecting its economic plausibility. These weightings can then be used to construct a single monetary policy path that incorporates both the scenarios considered and their likelihood. 

This approach builds on the ideas presented by the Governor of the Banque de France, François Villeroy de Galhau, at the CEPR annual symposium in December 2025, where a risk management framework based on a weighted synthesis of scenarios had already been discussed.

The optimal monetary policy path differs depending on the scenario chosen 

Take, for example, the Eurosystem's macroeconomic projections for June 2025. There was considerable uncertainty at the time surrounding US trade policy, and the Eurosystem had to develop its macroeconomic projections without knowing the outcome of the trade negotiations between the European Union and the United States. Two alternative scenarios to the baseline projection scenario were therefore factored in: a more moderate and a more severe scenario, based on changes in technical assumptions concerning potential tariffs and the level of trade uncertainty. Under each of these scenarios, the projections were not limited to a single trajectory but represented a possible distribution of trajectories (known as a “predictive distribution”), reflecting both the uncertainty surrounding future economic developments and the inherent margins of error in the projection tools.

The macroeconomic projections associated with each scenario were also based on the monetary policy paths anticipated by the markets, which was used as a technical assumption in the forecasting exercise. However, this path does not necessarily correspond to the path that would be optimal in terms of the ECB's macroeconomic stability objectives. An optimal monetary policy simulation was therefore performed for each scenario to identify the most appropriate interest rate path for reducing the gaps between actual and target inflation and between actual and potential output (Barnichon and Mesters, 2023; Lhuissier, 2025). This optimal path provided a benchmark for guiding monetary policy.

The results of these simulations corroborated the decision to lower the interest rate to 2% in June 2025, as anticipated by the markets, regardless of the scenario considered (Chart 1). However, the medium-term monetary policy path differed depending on the scenario chosen. The optimal policy simulation based on the baseline or moderate scenario indicated a slightly more accommodative stance than the path expected by the markets, with an additional 25-basis-point cut before the end of 2025. The optimal policy simulation based on the severe scenario suggested a much more accommodative stance, with two additional cuts in the second half of 2025.

Weighting of scenarios is determined by their plausibility

Given that each scenario implies a different optimal monetary policy trajectory, the question of their plausibility would appear to be essential. An adaptation of the Bayesian predictive synthesis method proposed by Tallman and West (2023) and Adrian, Giannone, Luciani and West (2025) is used to assess the likelihood of each scenario. The principle consists in combining the different scenarios to obtain a synthetic distribution (i.e. a weighted average of the predictive distributions associated with each scenario), and then adjusting the weighting of each scenario so that this aggregate distribution is as close as possible to an independent baseline forecast, in this case the ECB Survey of Professional Forecasters, dated 25 July 2025. The scenarios that contribute most to reproducing forecasters' expectations are thus assigned a higher weighting. Estimated weightings therefore do not reflect the probability of a scenario occurring, but rather its plausibility according to forecasters.

The analysis also includes an additional “artificial” scenario, designed to offset any incompleteness in the set of scenarios defined by the Eurosystem. This scenario, called “Other”, does not correspond to a specific economic hypothesis, but is a residual category grouping together all macroeconomic trends not covered by the Eurosystem scenarios and likely to reflect forecasters’ perspectives.

The weightings, estimated based on real GDP forecasts for 2026, show that the baseline and artificial scenarios each receive a weighting of around 30%, followed by the moderate scenario, with a weighting of close to 27%. The severe scenario is assigned a more modest weighting of around 12% (Chart 2). The relatively high weighting of the artificial scenario indicates that the range of scenarios used by the Eurosystem is incomplete and does not fully reflect the expectations of professional forecasters.

Chart 2 – Weighting of scenarios reflects their plausibility according to forecasters

Combining the scenarios results in a clear monetary policy stance

Using the estimated weightings, it was now possible to obtain a synthetic optimal interest rate path, factoring in both the baseline and alternative scenarios (Chart 3). This synthesis points to a cut in the interest rate in June 2025 to 2%, as well as a further reduction of 25 basis points before the end of 2025, compared with the path anticipated by the markets. (In reality, the second interest rate cut was not implemented as the economy was more resilient than expected in the second half of the year.)

This optimal synthesis is very similar to the optimal policy obtained under the baseline and moderate scenarios, for two reasons. First, these scenarios have the highest weightings and imply very similar optimal policy paths. Second, the severe scenario, which implies a very different optimal path, is assigned a limited weighting, so that its impact on overall monetary policy decision-making is small.

The optimal monetary policy synthesis therefore allows uncertainty to be explicitly factored into monetary policy decision-making, integrating both the scenarios considered and their likelihood, rather than relying on a single and uncertain future scenario.

Chart 3 – Optimal monetary policy trajectory incorporating the three possible scenarios