Skip to content

News Predictions

ECB

ECB reports a quantitative toolkit to evaluate scenarios for September 2026 projections

The ECB published a box in the Economic Bulletin introducing a framework that constructs joint predictive distributions for growth and inflation and scores scenario paths by empirical plausibility. Applied to the Middle East conflict scenarios in the September 2026 staff projections, the box reports high plausibility for the milder scenario, intermediate scores for the adverse scenario, and a considerably lower plausibility for the severe scenario.

PipDig Newsroom3 min read
ECB reports a quantitative toolkit to evaluate scenarios for September 2026 projections

What the ECB published

The ECB published a box in the Economic Bulletin that introduces a quantitative framework to evaluate alternative macroeconomic scenarios and to assess the plausibility of their outcomes, according to the box. The box appears as part of the ECB Economic Bulletin, Issue 7/2026, and is prepared by Niccolò Battistini, Mohammed Chahad and Luís Fonseca. The text says the framework is intended to judge scenario outcomes for real GDP growth and HICP inflation relative to the Eurosystem/ECB staff baseline projections.

The figures in this story
Cumulative absolute deviation — real GDP growth0.2 to 1.2 percentage pointsRange across the Middle East conflict scenarios relative to the baseline
Cumulative absolute deviation — HICP inflation1.0 to 4.3 percentage pointsRange across the Middle East conflict scenarios relative to the baseline

Each figure as the source states it.

How the toolkit measures plausibility

The box sets out a two-step approach. First it constructs joint predictive distributions for the variables of interest that capture temporal and cross-variable correlations. These joint densities are centred so their mode corresponds to the baseline projections and are computed over 20-year rolling windows starting from the early 1980s, the box says. Two complementary methods are used to derive the joint density. The first is a model-based approach that obtains univariate conditional distributions from Bayesian quantile regressions and combines them via the chain rule of probability. The second is a copula-based approach that builds joint densities by combining univariate marginals with historical patterns. The univariate marginals are taken from two sources: historical data and a survey-based distribution derived from the ECB Survey of Professional Forecasters (SPF), incorporating forecasters' forward-looking beliefs, according to the box.

How scenarios are scored and applied to the Middle East conflict

The box computes an empirical plausibility index for each scenario by comparing the likelihood of the scenario path for growth and inflation with the likelihood of the baseline path across historical subsamples. For each subsample the relative likelihood is calculated from the joint density; the final index is the highest relative-likelihood value across subsamples and ranges from 0 to 100, the box explains. Applied to the Middle East conflict scenarios in the September 2026 ECB staff macroeconomic projections, the box reports that scenarios differ in assumptions about energy commodity prices, financial uncertainty and second‑round effects on inflation. Cumulative absolute deviations from the baseline across the scenarios span 0.2 to 1.2 percentage points for real GDP growth and 1.0 to 4.3 percentage points for HICP inflation, the box says. The box reports the empirical plausibility index results: the milder scenario shows a high plausibility with an index ranging between 82 and 96 across approaches. The adverse scenario has a range between 56 and 85. The severe scenario is assessed as considerably less plausible, though the box says it remains important to analyse it to cover a wide range of conflict outcomes. The box also notes the plausibility assessment is robust across the different frameworks used to derive the joint densities.

Comparison, limits and concluding points

The box illustrates that empirical plausibility is not the same as severity. It compares the Middle East adverse scenario with an ad hoc flat‑path scenario that has a similar cumulative deviation from the baseline and finds the adverse scenario far more plausible: the adverse scenario scores 56 while the flat‑path scenario scores almost 0 in the comparison reported in the box. The box emphasises that accounting for temporal and cross‑variable correlations matters. Omitting either or both dependency dimensions can yield partial and potentially misleading plausibility assessments, the box says. The authors conclude that the framework provides a measure of empirical plausibility for scenario outcomes but does not assess the economic plausibility of narrative events or their probability of materialisation; those remain matters for policy judgement.

Source: Narratives and risks: a quantitative toolkit to evaluate scenarios applied to the Middle East conflict

More from the desk

All market news →