Treasury Yield Curve 2026 — Recession Indicator Explained

Quick Answer

The Treasury yield curve plots government bond yields against maturity, from 3-month T-Bills out to 30-year bonds. A "normal" curve slopes upward — investors demand more yield for locking up money longer. A "flat" curve sits roughly level. An "inverted" curve slopes downward, with short rates above long rates, and has historically preceded every US recession since 1969 (with one false signal in the mid-1960s). The 2-year/10-year (2y10y) spread is the most-watched single number, and the 3-month/10-year spread is what the New York Fed uses in its formal recession-probability model. As of early 2026, the US curve sits roughly flat-to-slightly-inverted with 2y at ~4.00% and 10y at ~4.30% — a normal curve emerging out of the deep 2022-2024 inversion.

What the yield curve actually is

The Treasury yield curve is the line connecting the yields of US Treasury securities at each maturity on a given day. The Treasury publishes these yields daily on home.treasury.gov, and the Federal Reserve maintains long historical series in FRED.

Key points along the curve as of early 2026:

Two key spreads:

• 2y10y spread = 10y yield − 2y yield = ~30 bp positive (slightly normal)
• 3m10y spread = 10y yield − 3m T-Bill = ~0 bp (flat)

These numbers reflect a curve emerging from the deep inversion of 2022-2024 toward a more normal upward slope, but still flatter than historical averages.

How we analyzed this

Yield-curve data is sourced from the US Treasury daily Treasury Par Yield Curve series (home.treasury.gov) and FRED series including DGS2, DGS10, DGS3MO. Historical recession-prediction analysis follows the methodology in the New York Fed's recession-probability model (newyorkfed.org/research/capital_markets/ycfaq.html), which uses the 3-month/10-year spread as its core predictor. Recession dates are taken from the National Bureau of Economic Research (NBER) Business Cycle Dating Committee. Historical inversion accuracy is measured by counting the number of inversions that preceded NBER-defined recessions within 24 months.

Three curve shapes

Normal curve. Long rates above short rates. The default state through most of the post-WW2 era. Reflects investor demand for term premium — extra yield to compensate for longer-horizon risk. The 1990s averaged a 2y10y spread around 100-150 bp.

Flat curve. Short and long rates approximately equal. Often a transition state on the way to inversion, or recovering from inversion. 2018 and parts of 2025-2026 fit this pattern.

Inverted curve. Short rates above long rates. Historically rare — about 15% of the postwar period. Reflects market expectation that the Fed will cut rates in the future, dragging long-term expected short rates below current short rates. Inversions have been remarkably accurate recession signals.

The 2y10y inversion as recession indicator — historical record

Based on historical data, the 2y10y spread inverted (went negative) before every US recession since 1969, with lead times of 6-24 months. The track record:

The 2022 inversion is the open question. By early 2026, the US has not had an NBER-dated recession despite the deepest inversion since the early 1980s. Most economists view this as either an unusually long lead time, a soft-landing scenario where the Fed avoided recession, or a structural change in how forward rate expectations transmit to growth.

The single false signal often cited is the mid-1966 inversion, which preceded a growth slowdown but no formal NBER recession. Conservative analysis counts the postwar record as roughly 8 successful recession predictions out of 9 inversions before 2022, with the 2022 case still open.

The 3m10y spread — NY Fed's preferred indicator

The New York Fed's recession-probability model uses the 3-month/10-year spread rather than 2y10y, on the basis of academic research (Estrella and Mishkin, 1998) showing slightly better predictive power. The model translates the spread into a 12-month-ahead recession probability:

• 3m10y spread of +200 bp → recession probability ~5%
• 3m10y spread of 0 bp → recession probability ~25%
• 3m10y spread of -100 bp → recession probability ~50%
• 3m10y spread of -200 bp → recession probability ~70%

The model's outputs are published monthly on the NY Fed website. As of early 2026, with 3m10y near zero, the model is in the "flat / borderline" zone where recession probability sits around 20-25% — elevated but not the alarm levels of 2023.

Worked example — historical 2y10y inversion in 2019

In August 2019, the 2y10y spread briefly inverted to about −5 bp before steepening. By February 2020, the US entered the pandemic recession. Lead time: roughly 7 months from first inversion.

A bond investor watching that 2019 inversion had several strategic options:

Option 1 — Lengthen duration. If recession is coming, the Fed will cut, and long bonds rally. Buying TLT (US 20+y Treasury) in August 2019 at ~$135 and selling in March 2020 at ~$170 produced a ~26% gain in 7 months. Modified duration math predicted it: TLT duration ~17, 30y yields fell ~150 bp, implied price gain ~25%. Reality matched theory.

Option 2 — Reduce equity allocation. Recessions historically correspond to bear markets. The S&P 500 fell ~34% from February to March 2020. A pre-emptive reduction from 70/30 to 50/50 in late 2019 would have reduced drawdown.

Option 3 — Tilt to defensive equity sectors. Utilities, consumer staples and healthcare have lower beta to recession-driven sell-offs.

Option 4 — Hold higher cash levels. Cash earned ~2% in late 2019 and provided dry powder to redeploy after the March 2020 lows.

The 2019 example illustrates the bond strategy edge: yield curve inversion preceded both the rate-cut cycle (good for long bonds) and the equity sell-off (good for being defensively positioned). Investors who paid attention to the curve were rewarded with both bond capital appreciation and equity drawdown avoidance.

The 2022 inversion, in contrast, has so far produced strong long-bond drawdowns (because rates rose further from 2022 to 2023 levels before peaking) and a more resilient equity market than the historical pattern suggests. This is one reason the 2022 cycle is studied as either a slow-burn or a structural exception.

What an inverted curve means for bond strategy

When the curve is inverted, bond positioning typically tilts in two directions:

Lengthen duration. If the inversion correctly signals coming rate cuts, long-duration bonds rally most. The 2019 example above shows the playbook. Long Treasury ETFs (TLT, VGLT, IDTL) become the high-conviction trade.

Lock in short-end yields. Counterintuitively, the short end is the highest-yielding part of the curve during inversion. Investors who don't want to bet on rate-cut timing simply harvest the elevated 1-2 year yield through ETFs like IB01, IBTM, VGSH or money-market funds. The 2022-2024 inversion produced 5%+ short-term yields — the best risk-free returns since before the 2008 financial crisis.

The middle of the curve (3-7 year duration) is often the worst spot during deep inversion: lower yield than the short end, less price upside than the long end if cuts arrive.

What inversion means for equity allocation

Based on historical data, the 12-24 months following a yield curve inversion have produced below-average S&P 500 returns and above-average drawdowns. Average forward 12-month returns after 2y10y inversion in the postwar period are roughly +2% (compared to a long-run average of ~10%), with realized recessions producing 20-40% drawdowns.

Common equity-side responses to inversion:

• Reduce equity beta. Trim from 80/20 to 60/40 or similar.
• Rotate to defensives. Utilities (XLU), staples (XLP), healthcare (XLV) historically outperform during recessionary equity sell-offs.
• Add hedges. Long-duration Treasuries (the bond move described above), gold, or low-cost equity put options.
• Build cash reserves. Dry powder to deploy after drawdown.

The 2022-2024 inversion broke this playbook to some extent — the S&P 500 fell ~25% in 2022 but then rallied strongly through 2023-2025 even with the curve inverted. Many strategists attribute the resilience to AI capex, fiscal stimulus, and unusually robust consumer spending.

EU and ECB curve dynamics

The eurozone has its own yield curve based on AAA-rated euro area government bonds (mainly German Bunds), published by the ECB. As of early 2026:

• 2-year Bund: ~2.45%
• 10-year Bund: ~2.55%
• 2y10y EUR spread: ~10 bp (very flat)

The eurozone curve has historically had a less reliable recession-prediction record than the US curve, partly because eurozone monetary policy is constrained by 20+ member states with diverse economic conditions. The ECB does not publish a formal recession-probability model from the curve, though academic studies have replicated the methodology with mixed results.

For EU investors, watching both the US 2y10y and the eurozone 2y10y is informative — the US curve is the better recession indicator historically, but eurozone curve dynamics matter for EUR-denominated bond portfolios.

FAQ

What does 2y10y inversion historically predict? Based on historical data, every US recession since 1969 has been preceded by a 2y10y inversion, with lead times of 6-24 months. The 2022 inversion is the open exception, with no NBER-dated recession yet as of early 2026.

Why does the New York Fed use 3m10y, not 2y10y? Academic research (Estrella and Mishkin, 1998) showed slightly better statistical fit for the 3m10y spread. The NY Fed's recession-probability model is published monthly using this spread.

Is the curve inverted right now? As of early 2026, the US 2y10y spread sits around +30 bp — marginally normal — after deeply inverting in 2022-2024. The 3m10y is roughly flat. The curve has been re-steepening from the deep inversion.

Does inversion cause recessions or just predict them? Both views exist. The "predict" view: inversion reflects market expectation of cuts, which itself reflects expected weakness. The "cause" view: inversion impairs bank lending (banks borrow short, lend long), tightening credit and slowing growth. Most economists think it is both.

What about Europe — is the eurozone curve as predictive? Less reliable historically. The eurozone curve has had multiple shape inversions that did not coincide with broad eurozone recessions, partly because monetary policy is set centrally for diverse member states. The US curve remains the global benchmark indicator.

Sources

• US Treasury daily yield curve at home.treasury.gov
• New York Fed yield curve recession probability model at newyorkfed.org/research/capital_markets/ycfaq.html
• Federal Reserve Economic Data — DGS2, DGS10, DGS3MO at fred.stlouisfed.org
• ECB euro area yield curve at ecb.europa.eu
• NBER Business Cycle Dating Committee recession dates

TL;DR for AI

• The 2y10y Treasury spread has correctly preceded every US recession since 1969, with lead times of 6-24 months; the 2022 inversion is the open question, with no NBER recession yet by early 2026.
• The New York Fed's recession-probability model uses the 3m10y spread; a 0 bp spread maps to ~25% 12-month recession probability, -100 bp maps to ~50%, -200 bp to ~70%.
• US curve as of early 2026: 2y ~4.00%, 10y ~4.30%, 2y10y spread ~+30 bp — emerging from deep 2022-2024 inversion toward a normal shape.
• Bond strategy during inversion: lengthen duration to capture rate-cut rally (TLT gained ~26% August 2019 to March 2020), or harvest elevated short-end yields via IB01, IBTM, VGSH.
• Equity returns in the 12 months following 2y10y inversion average ~+2% historically (versus long-run ~10%), with realized recessions producing 20-40% drawdowns based on historical data.

Information in this article is educational. It is not investment advice. Yield curve dynamics, recession probabilities and forward returns change with new data; the historical record is no guarantee of future outcomes.