The Yield Curve Explained: Inversions, Recessions and What the Spread Says Now

On May 4, 2023, the 10-year Treasury note yielded 1.89 percentage points less than the 3-month bill. That was the deepest inversion of the yield curve in the daily data that begin in 1982, and by the rule of thumb that had worked before the 1990, 2001, 2008 and 2020 recessions, a downturn was due by 2024. It never came on schedule. Through the latest NBER-based reading for August 2026, no recession has been dated, and on September 30, 2026 the same spread stood at +1.09 points. If you sold stocks on the inversion, you watched from the sidelines. If you threw the signal away, you may be ignoring the best single recession indicator the bond market offers. This guide shows how to read the curve properly and how to use it without being whipsawed. If you want the wider dashboard first, start with our guide to the economic indicators executives should watch.

What is the yield curve?

The yield curve is a line connecting the yields on US Treasury securities from the shortest maturity (1-month bills) to the longest (30-year bonds) on a single day. Because every point is a loan to the same borrower, the US government, credit risk is held constant, and what remains is the price of time: how much extra yield investors demand to lock money up for longer, and where they expect short-term interest rates to go.

The Treasury Department publishes the official daily par yield curve, covering maturities from 1 month to 30 years, and the Federal Reserve Board republishes the same constant-maturity yields in its H.15 release. FRED, run by the St. Louis Fed, turns those into downloadable series such as DGS3MO (3-month), DGS2 (2-year) and DGS10 (10-year). Every number in this article comes from those series.

Professionals usually compress the whole curve into one number, the term spread: a long yield minus a short yield. The two you will see quoted most often are:

• 10-year minus 3-month (FRED series T10Y3M): the spread used in the New York Fed’s recession model.
• 10-year minus 2-year (FRED series T10Y2Y): the spread most headlines quote.

A positive spread means the curve slopes up; zero means it is flat; negative means it is inverted. On September 30, 2026 the readings were +1.09 and +0.41 points respectively.

How do you read a yield curve chart?

Read the shape first, then the level: the slope tells you what the market expects from the Fed and the economy, and the height tells you how expensive money is across the board. Chart 1 plots four real days that cover the main shapes.

Three readings to make on any curve:

1. The front end (1 to 6 months) sits close to the federal funds rate. It tells you where policy is today. On September 29, 2026 the effective fed funds rate (FRED DFF) was 3.88% and the 3-month bill 4.25%.
2. The 2-year is the market’s best guess of the average policy rate over the next two years. A 2-year well above fed funds, as now (4.89% vs 3.88%), says traders expect higher policy rates; a 2-year well below says they expect cuts.
3. The 10- and 30-year add growth and inflation expectations over a decade or more, plus a term premium. This is the part of the curve that sets mortgage rates and the discount rate applied to long-duration assets such as growth stocks.

Note the December 2006 curve: flat to slightly inverted at a high level. Flat curves are easy to miss because nothing dramatic is happening, but historically they were the waiting room before the inversion that mattered.

Why does the yield curve usually slope upward?

It slopes up because long yields are roughly the average of expected future short rates plus a term premium, and in normal times both push long yields above today’s bill rate. This is the expectations hypothesis with a risk adjustment, and the New York Fed’s research FAQ describes it as one of the most pervasive theories of the curve, while noting that the pure version has been repeatedly rejected in tests because risk and liquidity premiums matter too.

Break a 10-year yield into its parts:

• Expected path of short rates. If the market expects the Fed to keep rates near 4% on average for ten years, that is the anchor.
• Term premium. The extra yield investors demand for the risk that inflation, deficits or rate volatility make long bonds lose value. It is not directly observable; the New York Fed publishes model estimates. It was unusually low, even negative by some estimates, in the late 2010s and early 2020s, which flattened the curve without any recession signal attached.
• Supply and demand. Heavy Treasury issuance, foreign central-bank buying or Fed balance-sheet policy can move long yields independently of growth expectations.

What does an inverted yield curve mean?

An inverted curve means short-term Treasuries yield more than long-term ones, which in practice means the market expects the Fed to cut rates in the future, usually because it expects growth to weaken. The New York Fed’s explanation is straightforward: a monetary tightening lifts short rates, long rates rise less because investors expect the tightening to be reversed, and the same tightening that flattens or inverts the curve also slows the economy.

The mechanism also runs through the banking system. Banks borrow short and lend long; when short rates exceed long rates, the margin on new lending shrinks and credit standards tend to tighten. Cash becomes more attractive than long-term investment, which drains demand from rate-sensitive sectors such as housing and capital spending.

Three details from the New York Fed research matter for anyone using the signal:

• Level beats change. In the New York Fed’s words, “it is the level of the term spread — not the change — that helps forecast” recessions. A 50-basis-point move barely matters when the spread is above 300 basis points, but from 50 basis points it can raise the implied probability by 10 points or more.
• The source of the move does not matter much. A low spread reached through rising short rates or falling long rates carries similar information, though every recessionary episode was preceded by a substantial rise in short-term rates.
• Persistence matters. A one-day inversion can be noise from Treasury supply and demand; a signal lasting a month or more deserves attention. That is why this article uses monthly averages.

Which spread matters more: 10-year minus 2-year or 10-year minus 3-month?

For recession forecasting, the 10-year minus 3-month spread has the stronger research pedigree; the 10-year minus 2-year is more widely quoted and inverts earlier and more often. Background research cited by the New York Fed found the 3-month rate, paired with the 10-year, gives “a reasonable combination of accuracy and robustness over long time periods.” The 2-year is popular because it is liquid and captures the expected Fed path, but most term spreads are highly correlated.

Two practical points fall out of this table. First, the 2-year spread typically warned about five months earlier than the 3-month spread, so the 10Y−2Y works as an early alert and the 10Y−3M as the confirmation. Second, the 10Y−2Y also produced short monthly inversions in 1998 and early 2006 with no recession close behind; it is noisier. In 2019 the 10Y−2Y closed below zero on just three days in late August and never for a full month, while the 3-month spread inverted for five months.

Has an inverted yield curve predicted every recession?

Almost: since 1960 a 10-year minus 3-month inversion preceded every US recession, with one acknowledged false signal in 1967 and, so far, the 2022–24 episode. The New York Fed notes that, in monthly averages, the 10-year rate was at least 12 basis points below the 3-month rate before every recession in that period, and that two 1990s episodes in which the spread fell to just 42 and 12 basis points without inverting were not followed by recessions. Chart 2 shows the record since 1982 in the FRED data.

Read the chart from left to right and three patterns stand out:

• The signal is rare. In 537 months since January 1982 the monthly spread was negative in only a handful of clusters, and four of the five were followed by a recession.
• Depth did not predict severity. The shallow 1989 inversion (−0.16) led into a mild recession; the moderate 2006–07 inversion (−0.52) led into the worst downturn since the 1930s. The deepest inversion of all, in 2023, has not been followed by one.
• Spreads widen sharply once a recession starts. The Fed cuts aggressively and the short end collapses, so by the time the shaded band appears, the curve is usually steep again. Inversion is a pre-recession condition, not a during-recession one.

The NBER, which dates US recessions, defines one as “a significant decline in economic activity that is spread across the economy and lasts more than a few months.” It announces peaks and troughs with a lag, often many months after the fact, which is one more reason to plan off probabilities rather than wait for confirmation.

How long after an inversion does a recession start?

Since 1990, the business-cycle peak arrived 8 to 16 months after the first month of 10-year minus 3-month inversion, consistent with the New York Fed’s “four to six quarters.” Chart 3 adds the detail most investors miss: in each case the curve had already re-steepened, turning positive again, 2 to 7 months before the peak.

That second bar is the most useful practical lesson in the yield-curve literature. Investors who wait for the inversion to “end” as an all-clear have historically been relieved at exactly the wrong moment. Re-steepening after an inversion usually happens because the Fed starts cutting short rates as the economy softens: a bull steepener, led by falling short yields. In 2001 and 2007 the stock market’s worst stretch came after the re-steepening, not during the inversion.

The lag also explains why the signal is so hard to trade. Equities often kept rising for months after the first inversion. A rule that sold on day one of an inversion and bought back at the recession’s end would have forgone gains in late 1989, 2006 and 2019, and in 2022–25 it would have missed an entire bull market. The curve tells you the odds of a storm in the next year or so; it does not tell you which month to leave the beach.

Why didn’t the 2022–2024 inversion cause a recession?

Nobody knows for certain, but the leading explanations are that long yields were held down by factors unrelated to recession risk, and that the economy was unusually insulated from higher short-term rates. The facts first: the monthly 10-year minus 3-month spread was negative from November 2022 through November 2024, its deepest month was May 2023 at −1.73 points, and it flickered around zero again from March to August 2025 before turning decisively positive. The daily 10-year minus 2-year spread stayed inverted from July 6, 2022 to August 26, 2024, the longest unbroken run in the FRED series that begins in 1976. As of the latest USREC observation (August 2026), no recession had followed.

The explanations most often put forward, none of them settled:

• A compressed term premium. If long yields were low because investors demanded little extra pay for duration, rather than because they expected deep rate cuts, the inversion overstated recession expectations. The New York Fed’s own FAQ warns that model parameters can shift across monetary regimes even when predictive power survives.
• Locked-in borrowing costs. Many households held 30-year fixed mortgages and many companies had termed out debt at low rates during 2020–21, so higher short rates reached spending more slowly than in past cycles.
• Fiscal support and excess savings. Large deficits and pandemic-era savings kept demand strong while the Fed was tight.
• The cycle is not over. The NBER dates recessions with a long lag, and 45 months is beyond every modern lead time but not beyond the historical range before 1960. A minority view holds that the signal was delayed, not wrong.

The honest takeaway is not that the yield curve “broke”. It is that the signal is a probability. A model that assigned roughly 70% odds in mid-2023 also, by construction, assigned about 30% to no recession, and that is what happened, at least so far.

What is the yield curve saying in 2026?

As of September 30, 2026 the curve is upward-sloping across the board, the 10-year minus 3-month spread averaged +0.90 points in September, and the New York Fed’s model implies roughly 14% odds of recession twelve months ahead. Chart 4 places today’s reading on the model curve and on the regime ladder Proflex uses to translate the spread into action.

But the type of steepening matters as much as the level. Compare monthly averages from September 2025 to September 2026 in the FRED data:

This is a bear steepener: the spread widened because longer yields rose, not because short yields fell. The 2-year jumped more than a full point, which says the market is pricing a higher policy path; the effective fed funds rate itself moved up a quarter point, from 3.63% to 3.88%, on September 17, 2026 (FRED DFF), and the 10-year closed September 29 at 5.26%. That is a different risk from the classic recession sequence. It points to rate and inflation risk, which hits long-duration bonds, richly valued growth stocks and anyone with floating-rate borrowing, rather than to an imminent drop in growth.

So the 2026 read is: recession odds from the curve are low by historical standards, the late-cycle bull-steepener warning is not present, and the live risk is the cost of money. Pair this with sentiment and breadth readings from our market sentiment indicators guide before drawing a conclusion; no single gauge should run a portfolio.

How should investors use the yield curve?

Use the curve to set your defensive posture and your rate exposure in advance, not to time the stock market. For Proflex clients, who are often executives and engineers with a large single-stock position, RSU income and a tax bill on any sale, that posture covers five levers: cash, bond duration, leverage, concentration and hedges.

*Proflex calculation from the New York Fed probit parameters; illustrative ranges.

How that plays out for each lever:

• Cash and Treasuries. An inverted curve pays you to wait in bills, so match known spending (taxes on vesting RSUs, a home purchase, tuition) with bills or a short ladder and let the curve’s inversion work for you. Our bond ladder guide walks through building one.
• Duration. Long bonds tend to rally when the Fed starts cutting into weakness, which is the bull-steepener phase. In a bear steepener like 2026’s, the same duration loses money. See our fixed income portfolio management guide for how to size curve positions.
• Leverage. Margin and securities-backed lines are priced off short rates, so they cost the most when the curve is inverted and policy is tight. If you run options on margin, our comparison of portfolio margin vs Reg T explains how house requirements can tighten when volatility rises.
• Concentration. A recession hits single stocks harder than the index. An inversion is a good prompt to move a staged sale or exchange-fund decision up the calendar, while valuations and liquidity are still good.
• Hedges. Volatility is usually cheap during the lag between inversion and recession, which makes it the best time to buy protection. Our guide to hedging a stock portfolio with options compares puts, put spreads and collars.

The curve also informs where inside equities to be. Late-cycle phases have historically favored defensive sectors and quality balance sheets, and early recoveries have favored cyclicals; our companion piece on sector rotation through the business cycle maps that sequence.

Worked example: a $3 million household in an inverted-curve regime (illustrative)

1. Starting point. $1.8 million in one employer’s stock, $900,000 in an index fund, $300,000 in cash. A $250,000 tax bill and a $400,000 home down payment are due within 18 months.
2. Cash. Move $650,000 of known needs into a 3-, 6-, 12- and 18-month bill ladder; with the inverted curve of mid-2023, bills paid more than 10-year notes, so the safety costs nothing in yield.
3. Concentration. Set a 10b5-1 or staged plan to sell 25% of the employer stock over the next four quarters, rather than waiting for the recession to arrive.
4. Hedge. Collar part of the remaining stock for 12 months, financing the put by selling an upside call, so the floor lasts through the typical 8–16 month lag and the re-steepening that follows.
5. Review trigger. Revisit when the 10Y−3M turns positive again. If short yields are falling because the Fed is cutting, keep the hedges; if long yields are rising, shift attention to duration and rate risk instead.

The numbers are illustrative, and the right mix depends on tax basis, vesting schedules and trading windows; our recession-resistant portfolio guide covers the asset side in more depth.

A monthly yield-curve check

1. Pull the monthly averages of T10Y3M and T10Y2Y from FRED. Ignore single-day moves.
2. Place the 10Y−3M on the regime ladder in Chart 4.
3. Identify which leg moved: falling short rates (bull steepener) or rising long rates (bear steepener).
4. Check the 2-year against fed funds to see whether the market is pricing hikes or cuts.
5. Read the New York Fed’s official recession probability, published within the first two weeks of each month.
6. Cross-check with credit spreads, jobless claims and breadth before changing anything.
7. Act on cash, leverage, concentration and hedges, in that order; leave the equity allocation for last.

Frequently asked questions

Key takeaways

1. The yield curve plots Treasury yields by maturity; it normally slopes up because investors want extra yield to lend for longer.
2. An inversion means short rates exceed long rates; the 10-year minus 3-month spread inverted before every NBER recession since 1982.
3. Lead times from first inversion to the business-cycle peak ran 8 to 16 months, and the recession usually began after the curve had re-steepened.
4. The 2022–24 inversion was the deepest since 1982 and, through August 2026 data, has not been followed by a recession: treat the curve as a probability, not a verdict.
5. Level matters more than direction: below zero the New York Fed model’s implied odds exceed 30%; at September 2026’s +0.90 they are near 14%.
6. Use the curve to prepare rather than to trade: match cash to known needs, shorten leverage, review concentration and price hedges before volatility is expensive.

Sources and method

1. US Department of the Treasury, Interest Rate Statistics — the official daily par yield curve from 1 month to 30 years and its methodology notes
2. Board of Governors of the Federal Reserve System, H.15 Selected Interest Rates — constant-maturity Treasury yields and the federal funds rate behind the FRED series
3. Federal Reserve Bank of St. Louis, FRED: T10Y3M (10-Year minus 3-Month Treasury) — the spread charted in Charts 2–4; companion series T10Y2Y, USREC, DGS1MO–DGS30 and DFF
4. Federal Reserve Bank of New York, The Yield Curve as a Leading Indicator — the recession-probability model, its probit parameters and the research FAQ quoted here
5. National Bureau of Economic Research, US Business Cycle Expansions and Contractions — official peak and trough months used for the lead-time calculations
6. NBER, Business Cycle Dating — the definition of a recession and how the committee dates it

All yields and spreads are Treasury constant-maturity data from the Federal Reserve Board’s H.15 release as published on FRED, downloaded October 1, 2026 (daily observations through September 29–30, 2026; USREC through August 2026). Monthly figures are Proflex averages of daily data. Lead times are measured from the first month with a negative monthly 10-year minus 3-month average to the NBER business-cycle peak month. Recession probabilities are a Proflex calculation using the New York Fed’s published probit parameters; they approximate, and are not, the New York Fed’s official estimates. Portfolio examples are illustrative. This article is educational and is not investment, tax or legal advice.