3-Month / 10-Year Curve Inverts
The Fed's own preferred recession model input
The 10Y-3M spread is the basis for the New York Fed's recession probability model. When this measure inverts, the NY Fed model typically shows recession probability above 30%, a level that has preceded every recession since 1968.
| Date | 1M return | 1Y return | 5Y return |
|---|---|---|---|
| 1982-02-01 | -5.8% | +22.7% | +128.9% |
| 1989-03-27 | +5.6% | +16.1% | +62.1% |
| 1998-09-10 | +0.4% | +37.9% | +5.0% |
| 2000-04-07 | -6.9% | -25.0% | -24.6% |
| 2006-01-17 | -0.2% | +11.2% | +0.9% |
| 2007-07-20 | -5.8% | -17.9% | -10.3% |
| 2019-03-22 | +4.8% | -20.1% | +86.3% |
| 2022-10-18 | +6.4% | +15.0% | — |
| 2025-02-26 | -4.4% | +15.5% | — |
What history says
Editorial commentary written by ALAN analysts. Figures cited below are analyst-authored context — they are not derived from the chart above and may reflect different windows or sources.
The NY Fed's model translates the 10Y-3M spread into a recession probability. Readings above 30% have preceded every recession. This is publicly available and updated monthly.
The NY Fed model estimates recession probability 12 months ahead. A high reading today says 'recession within 12 months is likely,' not 'recession is happening now.'
The 10Y-3M remained inverted for over 24 months in 2022-24, the longest inversion on record. This has tested the model's patience — either a recession is delayed, or the relationship has changed.
A model built around a 12-month lead hands you 12 months of preparation: consider rebuilding cash reserves, upgrading bond credit quality, and writing down — in advance — what you will and won't do if a recession arrives. None of that requires selling equities, which have historically kept rising well after this signal first appears.