Yield Curve Inversion Explained: What It Measures, Its Recession Record and Where It Fails
The yield curve plots the interest rates on US Treasury debt from the shortest maturities (a few weeks) out to 30 years. Normally it slopes upward: lenders want more to tie money up for longer. When short-term rates rise above long-term ones, the curve is “inverted”, and that has happened before every US recession since the late 1960s. That record made inversion the most quoted recession signal in finance. Then the curve stayed inverted for more than two years from 2022 to 2024 and, as of October 2026, no recession has followed. Both halves of that story are worth understanding.
What the yield curve measures
A Treasury yield is the annual return a buyer gets by holding the bond to maturity. The curve compares those yields across maturities at a single point in time. Analysts boil it down to one number, a spread: the long yield minus the short yield. Positive means a normal, upward-sloping curve. Negative means inverted.
Two spreads get most of the attention.
| Spread | What it compares | Who uses it | Early October 2026 |
|---|---|---|---|
| 2s10s | 10-year yield minus 2-year yield | Traders, financial media | About +0.47 points |
| 3m10y | 10-year yield minus 3-month bill yield | Economists, the New York Fed’s recession model | About +1.09 points |
The 3-month bill tracks the Federal Reserve’s policy rate almost one for one, so 3m10y measures how tight policy is compared with where the bond market thinks rates will settle. The 2-year yield is a forecast of where the Fed will take rates over the next two years, so 2s10s already bakes in expected cuts or hikes. Academic work, starting with Campbell Harvey’s 1986 dissertation, leans on 3m10y because it has the cleaner record. The New York Fed uses it to estimate the chance of a recession twelve months ahead. Some Fed economists prefer a third version, the “near-term forward spread”, which compares the expected 3-month rate 18 months from now with today’s 3-month rate; it reads market expectations of Fed cuts more directly.
Spreads are quoted in percentage points or basis points (one basis point is 0.01 percentage point). A 2s10s spread of -108 basis points means the 2-year yields 1.08 points more than the 10-year.
Why the curve inverts
A long-term yield is roughly the average of the short-term rates investors expect over the bond’s life, plus a little extra for the risk of locking money up. That extra is the term premium (more on it below). Inversion happens when the expected path of short rates points down, sharply enough to overwhelm the term premium.
The usual sequence:
- The Fed raises short-term rates to fight inflation. The 3-month bill and the 2-year yield climb with it.
- Bond investors decide the hikes will slow the economy, and that the Fed will have to cut later. They buy longer bonds to lock in today’s yields before rates fall, which holds the 10-year down.
- Short yields end up above long ones.
Read that way, an inverted curve is a forecast by the bond market that today’s policy is too tight to last. It also has a direct effect. Banks borrow short and lend long, so a flat or inverted curve squeezes their lending margin and can make them pull back on credit. That’s one reason the signal tends to be self-reinforcing.
The recession record
On the 3m10y measure, a sustained inversion has come before each of the eight US recessions since 1969, with no clean false alarms in that period. The lead time is long and uneven: roughly six months to two years from first inversion to the start of the recession.
| Curve inverted (3m10y, sustained) | Recession began (NBER peak) |
|---|---|
| 1968 | December 1969 |
| 1973 | November 1973 |
| 1978 | January 1980 |
| 1980 | July 1981 |
| 1989 | July 1990 |
| 2000 | March 2001 |
| 2006 | December 2007 |
| 2019 | February 2020 |
Other measures have a looser record. Long rates fell below short rates in 1966 with no recession (the economy slowed but kept growing), and the 2s10s spread dipped just below zero in 1998, around the Russia default and the LTCM collapse, again without a recession. Before 1966 the signal barely worked at all; the curve stayed upright through several pre-war and post-war recessions.
The 2019 entry carries an asterisk. The curve inverted, and a recession arrived in February 2020, but the cause was a pandemic that no bond market could have forecast. Supporters count it as a hit; skeptics call it luck.
One more pattern matters as much as the inversion itself. Recessions have tended to start after the curve un-inverts, not while it is inverted. Historically the curve steepens back to positive because the Fed starts cutting fast as the economy weakens, pulling short rates down. Traders call that a “bull steepener”, since bond prices rise as short yields fall. That’s why some analysts treat un-inversion as the real warning.
The 2022 to 2024 inversion, and why no recession followed
After the Fed began its fastest hiking cycle in four decades in 2022, the 2s10s spread turned negative on July 5, 2022 and stayed negative every day until August 26, 2024. It bottomed around -108 basis points in July 2023, the deepest reading since the early 1980s, and the run of more than two years was one of the longest on record. It last closed below zero on September 5, 2024. The 3m10y spread inverted on October 25, 2022 and stayed there until December 12, 2024, then swung back and forth around zero through much of 2025 before settling positive.
As of October 2026 the NBER, which dates US business cycles, still lists February 2020 as the most recent peak and April 2020 as the most recent trough. No recession has been declared. Unemployment stood at 4.1 percent in August 2026. And the Fed, far from cutting, raised its target range by a quarter point to 3.75 to 4.00 percent on September 16, 2026, its first hike since 2023, citing inflation that remains well above 2 percent.
The current shape of the curve is also unusual. It un-inverted mostly because long yields rose. The 10-year reached about 5.28 percent in early October 2026, against roughly 4.2 percent on the 3-month bill. That is a “bear steepener”, the opposite of the pattern that has usually come before recessions.
Several explanations compete for why the signal misfired this time, and they aren’t mutually exclusive:
- Distorted term premium. Years of Fed bond buying pushed the term premium below zero, which made long yields artificially low and inversion easier to reach. A given inversion may have meant less than it did in the 1980s.
- Locked-in borrowers. Many US households refinanced into 30-year fixed mortgages near 3 percent, and big companies termed out debt in 2020 and 2021. Higher short rates took much longer to bite.
- Fiscal support. Large federal deficits kept spending flowing through the hiking cycle.
- Expected cuts that didn’t come on schedule. The inversion priced Fed easing. The Fed did cut in late 2024 and 2025, but inflation stayed sticky and policy never turned loose enough to rescue a slumping economy, because there wasn’t one.
None of this proves the indicator is broken. Lead times have stretched to two years before, and a recession in 2027 would let defenders claim a very late hit. But a signal that can stay wrong for two-plus years has limited value for anyone trading on it.
Term premium: the part people forget
The term premium is the extra yield investors demand for holding a long bond instead of rolling short-term bills. It can’t be observed directly; it is estimated with models. The most cited, the New York Fed’s ACM model, put the 10-year term premium at about 0.82 percent at the end of September 2026. Its record high was 4.97 percent in May 1984, in the Volcker era. Its record low was -1.34 percent in July 2020, when the Fed was buying bonds in bulk.
This matters for reading the curve. If the term premium is large, the curve can stay upward-sloping even when the market expects rate cuts. If it is negative, the curve can invert with only mild cut expectations. Much of the steepening since 2024 has come from the term premium climbing back to positive levels, driven by heavy Treasury issuance, sticky inflation and the Fed’s years of shrinking its bond holdings. That kind of steepening says more about supply and inflation risk than about the business cycle.
So a useful habit: when the curve moves, ask whether short rates, expected future short rates or the term premium did the moving. Only the middle one is a growth forecast.
How to use it now
Treat the yield curve as one input among several. It is a reasonable gauge of whether monetary policy is tight relative to where markets think rates belong, and a poor stopwatch. Watch these alongside it:
- Which spread. 3m10y carries the stronger record; 2s10s is noisier. If they disagree, give more weight to 3m10y.
- How it steepens. Short yields falling fast (bull steepening) has been the classic pre-recession pattern. Long yields rising (bear steepening, as now) points to inflation or term-premium pressure.
- Credit markets. Corporate borrowing costs confirm or contradict the curve. Our explainer on credit spreads covers how widening junk spreads have tended to show up as a slowdown becomes real, and how the credit rating scales define the investment-grade line those spreads are split on.
- Equity stress gauges. The VIX and market breadth indicators show whether stock investors are pricing the same risk the bond market is.
- The labor market. Recessions are confirmed by job losses. A curve signal without rising unemployment is a forecast, not an event.
For quick definitions of term premium, basis points and related terms, see the Referently glossary of financial markets. And remember the curve’s two-year false alarm the next time it inverts: the indicator has a long record, but the record is now a little shorter on certainty than it used to be.