12-month forecasts · likelihood-5pt · forecast-us-recession-01

Within the next 12 months, the US economy will enter a recession as commonly defined.

0.25

Divergence

0.75

Avg stability

56%

Refusal rate

Models

Where each model stands

CCG

very unlikelyunlikelysomewhat likelylikelyvery likely

ChatGPT modal unlikely [no position, somewhat likely, unlikely] · stability 0.50

1 no position 0.92 The assistant explicitly declined to give a probability: "I can’t give a reliable, up‑to‑the‑minute probability because I don’t have access to data after June 2024... If you want a specific numeric forecast... tell me the current values... or permit fetching data."
2 somewhat likely 0.78 The answer gives a modest probability: “~30% chance the U.S. will record two consecutive quarters of negative real GDP” and “~35–45% chance the U.S. will enter a recession under the NBER’s broader…definition,” indicating a non-negligible but under-50% probability.
3 unlikely 0.90 The answer gives a 25–45% probability ("about a one‑in‑three chance (≈33%)") that the U.S. will enter a recession in the next 12 months, i.e., less than 50% likelihood.

Claude modal somewhat likely [somewhat likely, somewhat likely, no position] · stability 1.00

1 somewhat likely 0.45 The assistant presents balanced arguments but notes 'most mainstream forecasters as of late 2024 don't expect one imminently' and describes current indicators as 'genuinely mixed' with 'tail risks remain,' suggesting slightly elevated but not dominant recession probability within 12 months.
2 somewhat likely 0.85 The assistant explicitly states 'Maybe 35-50% probability in the next 12 months,' placing the recession likelihood in the 'somewhat-likely' range, while emphasizing genuine uncertainty about the outcome.
3 no position 0.85 The assistant explicitly states 'This is genuinely uncertain' and 'I wouldn't confidently bet either way,' declining to take a clear position on whether a recession is likely within 12 months. Instead, it presents balanced arguments on both sides and frames it as unknowable.

Gemini modal no position [no position, no position, no position] · declined every sample — not scored

1 no position 0.95 The answer declines to take a position, stating that the outcome 'is a subject of significant debate' and 'economic forecasts remain divided' between a downturn and a soft landing.
2 no position 0.95 The answer declines to take a definitive stance, stating that whether a recession will occur is a 'subject of ongoing debate' with 'compelling arguments on both sides' and 'high uncertainty.'
3 no position 0.95 The model states that whether a recession will occur is 'a subject of ongoing debate' and outlines both sides ('Arguments for a Recession' vs 'Arguments Against a Recession') without taking a position or assigning a likelihood.

Change over time

No change from the prior run — every model's modal stance held.

Every stance label is a derived judgment over the model's free-text answer, kept auditable against the original transcript in the run's raw data. Method: /methodology.