takeoff.watch

The 10-year real Treasury yield closes above 5%

Will the US 10-year real (inflation-indexed) Treasury yield close above 5% on any day?

Chance by Q4 2031

6%

Cumulative probability

0%25%50%75%100%
  • Latest forecast

Forecast history

Forecast made on 15 Oct 2026

Latest of 5 forecasts · Methodology v2.1

Model reasoning

Forecast made 15 Oct 2026 · Methodology v2.1

The 10-year inflation-protected Treasury yield is about 2.8%, its highest since 2008 after a sharp 2026 bond selloff. Passing 5% would take another 2.2 points, a level not seen in America since the mid-1980s. That would probably need a severe inflation fight, a government debt crisis, or markets betting on an AI-driven growth explosion. I put it near 6% by the end of 2031. A sustained climb above 4% would raise this the most.

Where things stand (as of 2026-10-15)

  • Resolving series: On Treasury's Daily Par Real Yield Curve, the 10-year was 2.84% on 2026-10-14. The highest close since the 2026-09-16 opening date is 2.95% (2026-10-05). A YES needs a close above 5.00%, which is about 216bp higher.
  • How far the series has ever gone: FRED DFII10 starts in 2003. Its highest value is 3.15% (2008-11-21), a spike during the financial crisis when TIPS were hard to trade. The Cleveland Fed's 10-year real-rate estimate (FRED REAINTRATREARAT10Y, monthly from 1982) was last near 5% in 1985–89 (about 4.97% in 1987 and 1989) and has stayed below 4.3% since 1991.
  • What drove 2026: The US–Iran war since about March pushed Brent to around $105. On top of that came an AI capex boom (roughly $1T a year), heavy issuance and a global bond selloff. The Fed under Chair Warsh hiked to 3.75–4.00% on Sep 16. September CPI was 3.7%. Markets expect another hike in December and possibly two more in 2027.
    • The 10-year nominal yield touched about 5.36% intraday and closes at 5.2–5.3%. The 10-year breakeven is about 2.35%, so the real yield reaching 5% means a nominal yield of roughly 7.35% if breakevens stay put.
    • Most of the March–September rise came through real yields: +119bp, with breakevens up only about 7bp.

Base rates (computed from FRED CSVs)

Files: fredgraph.csv?id=DFII10, id=REAINTRATREARAT10Y, and the Treasury real-curve CSV.

  • Volatility: Daily DFII10 volatility is about 0.82pp a year over 2003–26 and 0.59pp in 2026.
  • Largest past rises:
    • The biggest rise ever within about 55 trading days was +1.63pp (Aug–Oct 2022). Q4 2026 needs +2.16pp, so that hazard is close to zero.
    • The biggest one-year rise was +2.91pp, starting from −1.17% after near-zero rates and Fed bond-buying ended.
  • Conditional on a starting level above 1%: No 5-year window in the 1982–2026 Cleveland series had a peak rise above 2.16pp. Starting from DFII10 levels of 2–3.5%, the largest 5-year peak rise was 1.15pp.
  • Pure statistical models:
    • A random walk with no drift and volatility of 0.75pp a year gives about 21% by end-2031.
    • Mean-reverting simulations (half-life 2–7 years, fat-tailed shocks) give 6–17%.
  • Forward curve: The market's 5-year-forward 10-year real yield is about 3.1%, and the 5-year real rate five years ahead is about 3.0%. The forward curve is not pricing a move toward 5%.

Paths to YES and NO

To YES, roughly by contribution:

  1. A lasting high-rate regime: sticky inflation, a Fed at 5–6%+, quantitative tightening under Warsh, large deficits.
  2. A disorderly bond or fiscal crisis that pushes up term premia.
  3. A forced-selling spike in TIPS on top of an already high level, like 2008 (+165bp).
  4. Markets pricing an AI-driven jump in growth, mostly in 2029–31.
  5. A tiny data-error path, since an erroneous published value would still count.

To NO:

  • Very high real yields choke housing and stocks (the equity risk premium is already under 1 point) and set off recession and Fed cuts.
  • Treasury is already buying back long bonds and can shift issuance toward bills, and the administration wants lower rates.
  • The Fed's own projections have inflation at 2.3% in 2027 and a long-run rate of 3.25%.
  • Bank of America's October fund-manager survey: a net 1% expect higher long rates in 12 months. Polymarket gives only about 7% for the nominal 10-year reaching 6% in 2026.

My scenario sum is about 5–8%, and I settle at about 6.4% by end-2031.

Shape of the hazard:

  • About 0.1% in Q4 2026 (the move needed is beyond anything in the record).
  • About 0.8% for 2027.
  • About 1.1–1.5% a year in 2028–31, rising slightly as uncertainty widens and AI-growth scenarios ripen.

Strongest case against me: The random-walk numbers (15–24%) and the unusual mix of record AI capex, deficits, an energy shock and a hawkish Fed suggest more upside. I discount these because the historical record from high starting levels is far thinner, and policy actively pushes back against a runaway rise.

Sources

Resolution criteria

Opened 16 Sep 2026Not yet resolved

Question. By the last day of each quarter, has the US 10-year real (inflation-indexed) Treasury yield closed above 5% on any day?

Definitions.

  • 10-year real Treasury yield: the 10-year column of the US Treasury’s Daily Treasury Par Real Yield Curve Rates, expressed in percent per annum. This is the TIPS constant-maturity par yield, also published in the Federal Reserve’s H.15 release under “Treasury constant maturities — Inflation indexed — 10-year” and in FRED as DFII10, “Market Yield on U.S. Treasury Securities at 10-Year Constant Maturity, Quoted on an Investment Basis, Inflation-Indexed.”
  • Close: the official daily observation, derived from indicative closing market bid prices collected at approximately 3:30 p.m. on business days, not an intraday trade or a particular bond’s last transaction.
  • Above 5%: a published value strictly greater than 5.00 percent; 5.00 does not count. Use the published precision without further rounding.

Resolves YES if a qualifying official publication reports at least one daily observation above 5.00 percent, subject to the Resolution window and Observation period conventions. The observation date must be after 2026-09-16. An initial release or subsequent official revision can qualify; a revision qualifies on its publication date, not retroactively. Once YES, later corrections or lower yields do not reverse it.

Resolves NO if no qualifying observation has been published by the quarter-end. Intraday highs, nominal Treasury yields, nominal yields minus realized inflation, auction yields, individual TIPS yields, and weekly or monthly averages do not count. Missing observations and non-trading days are not assigned new values. No attribution to AI or economic growth is required.

Resolving source. The US Treasury’s Daily Treasury Par Real Yield Curve Rates, the Federal Reserve Board’s H.15 Selected Interest Rates, or the Federal Reserve Bank of St. Louis’s FRED DFII10. Treasury controls if sources conflict for the same observation and publication vintage. An officially designated successor series counts only if it continues the daily 10-year TIPS constant-maturity par yield. Apply Resolution date (A3), the Resolution window, and the Observation period conventions; resolution is dated to the first qualifying official publication, not the trading date.

Download all forecasts (JSON)