The bond market’s most famous recession signal is not flashing right now. As of October 30, 2026, the U.S. Treasury yield curve slopes upward: investors earn more for lending to the government for ten years than for two. That is the normal shape — worth understanding because the abnormal shape, an inverted curve, has preceded every American recession since 1955.
The timing makes the question live. On October 28, the Federal Reserve held its policy rate at 3.75%–4.00% and signalled a December hike is on the table, while the 10-year Treasury yield touched 5% earlier in October — its first time above that mark in more than two decades — before easing to 4.94% at month-end. Short and long rates are moving for different reasons, and the gap between them is where the recession signal lives.
What the yield curve actually is
A yield curve is a snapshot of what the U.S. government pays to borrow for different lengths of time. Each point is the yield on a Treasury security of a given maturity: three months, two years, ten years, thirty years. Plotted together, they form a line that usually slopes upward, because lenders normally demand higher compensation for tying up money longer. That extra compensation has two parts: expectations about future short-term rates, and a “term premium” for bearing the risk that inflation or policy shifts erode a long bond’s value.
The curve moves every day, and for different reasons at different points. The short end is anchored by the Federal Reserve: the 3-month bill yield tracks the policy rate closely, and the 2-year yield reflects where markets think the policy rate is heading. The long end is set by the market’s view of growth, inflation and the supply of government debt — plus a “term premium” for bearing the risk that inflation or policy shifts erode a long bond’s value.
That split is the key to reading the curve. When the short end rises while the long end stays put, the market is saying the Fed will tighten now but growth will cool later. When the short end rises above the long end, the curve inverts — the market is pricing a future in which the Fed will have to cut rates, usually because the economy is heading into trouble.
What “inverted” means — and the two spreads that matter
Inversion is simple to define and easy to misread. A curve is inverted when a shorter-maturity yield exceeds a longer-maturity yield. But “the curve” is not one number: the slope can be measured between any two points, and two particular spreads dominate the discussion because they have different track records.
The 2s10s spread — the 10-year yield minus the 2-year yield — is the version quoted on television and in headlines. It inverted before every U.S. recession since 1955. Its latest long episode ran from July 2022 until late summer 2024, the longest inversion in the modern record.
The 10-year minus 3-month spread is the version the Federal Reserve Bank of New York uses in its recession-probability model, because the 3-month bill tracks the policy rate more directly than the 2-year note does. Its record was even stronger: it inverted eight times over more than sixty years, and a recession followed every time — until the ninth episode, in late 2022, when it bottomed near negative 200 basis points and produced no recession, breaking the 8-for-8 streak.
The two spreads can disagree — in October 2026 the 2s10s has been steepening since its inversion ended, while the 10-year minus 3-month gap has been flattening. Neither is negative, but they are moving in opposite directions: the curve is a set of readings, not a single dial.
The yield curve today
The table below uses the U.S. Treasury’s own daily par yield curve rates for October 30, 2026, the last trading day of the month; the two spreads are our own calculations from those published yields.
| Maturity | Yield, October 30, 2026 |
|---|---|
| 3 months | 4.16% |
| 6 months | 4.25% |
| 1 year | 4.36% |
| 2 years | 4.56% |
| 5 years | 4.71% |
| 10 years | 4.94% |
| 30 years | 5.31% |
| 2s10s spread (10Y − 2Y) | +0.38 pp (+38 bp) |
| 10Y − 3M spread | +0.78 pp (+78 bp) |
Source: U.S. Department of the Treasury, Daily Treasury Par Yield Curve Rates, October 30, 2026; spreads calculated by Macrometer.
Two things stand out. First, both spreads are positive: the curve is not inverted. Second, the long end is still elevated: the 10-year touched 5% earlier in the month, a two-decade high, before easing to 4.94%, while the 30-year remains above 5% at 5.31% — a move driven more by real yields and the term premium than by inflation expectations. After the October 28 FOMC decision to hold rates, the curve steepened further, with the 2s10s spread widening to 38 basis points by October 30 — Reuters had put it at about 34 basis points immediately after the decision.
Why inversion has predicted recessions
The curve’s forecasting power works through three channels that reinforce each other.
The bank lending channel. Banks borrow short and lend long. When short rates exceed long rates, the margin on new loans turns negative, credit standards tighten, and lending to households and businesses slows. A credit squeeze follows the inversion with a lag, and spending slows with it.
The expectations channel. The long end of the curve is, in effect, the market’s forecast of future short rates. When investors buy 10-year bonds at yields below the 2-year, they are betting the Fed will cut rates because growth is faltering. Inversion is the market writing down its own recession forecast.
The policy channel. Inversions almost always happen because the Fed has tightened aggressively to fight inflation, pushing short rates above the level the economy can sustain. The inversion marks the point where policy has become restrictive enough that something breaks — a credit event, a housing slowdown, a hiring freeze — and the downturn follows.
Timing matters as much as the signal: recessions have historically arrived six to eighteen months after the curve un-inverts — after short rates fall back below long rates, usually because the Fed has started cutting. The inversion is the warning; the un-inversion is the countdown.
The track record — and its first big miss
The table below compiles the 2s10s inversion episodes of the last four decades against the recessions that followed. Inversion dates follow Federal Reserve research; recession dates are the National Bureau of Economic Research’s official dates. Lags are our own calculation.
| Inversion began | Recession began (NBER) | Lag, inversion to recession |
|---|---|---|
| June 1989 | July 1990 | 13 months |
| February 2000 | March 2001 | 13 months |
| December 2005 | December 2007 | 24 months |
| August 2019 | February 2020 | 6 months |
| July 2022 | None as of October 2026 | 51 months and counting |
The pattern is striking — and the last row is the problem. Four of the five episodes were followed by a recession within two years. The fifth, the July 2022 inversion, was the longest in the modern record, running until late summer 2024, and no recession has followed. The New York Fed’s preferred 10-year minus 3-month measure tells the same story: eight inversions, eight recessions, and then a ninth in late 2022 that bottomed near negative 200 basis points and produced none.
That does not erase the indicator’s history, but it changes how it should be used. The 2022 episode coincided with a strong labour market, large pandemic-era household savings buffers, and an economy with far less leverage than in 2007. Inversion is a strong signal, not a law — and the latest evidence says its lead time can stretch well beyond the historical range, or fail outright, when the economy’s buffers are unusually large.
What inversion does not mean
Three cautions keep the signal honest.
First, inversion does not cause recessions. It reflects the market’s expectations and the stance of monetary policy; the recession is caused by the underlying tightening, credit stress or shock. Treating the curve as a trigger rather than a thermometer leads to the wrong trades.
Second, not every inversion has been followed by a recession. The 1966 inversion produced a credit crunch and a sharp growth slowdown but no official recession, and the curve flattened toward inversion in 1995 and 1998 without a downturn following. The record is “every recession was preceded by an inversion,” which is not the same as “every inversion is followed by a recession.”
Third, the signal is about the un-inversion as much as the inversion. A curve that inverts and stays inverted is a warning; a curve that inverts and then steepens sharply is often the market pricing the recession itself.
What to watch next
The curve will be decided by two things: what the Fed does in December, and whether the long end keeps demanding a higher term premium. The FOMC meets December 8–9, with a hike currently the more likely outcome after October’s hold; the next CPI and PCE releases will shape that pricing. An inversion-based recession call would need the New York Fed’s 10-year minus 3-month measure to turn negative and stay there for a quarter, or the 2s10s to re-invert. Neither is close.
The curve is one indicator, not a verdict: it currently says the market expects the Fed to stay restrictive, and that holding long-term government debt has become expensive. What would change the story is a flattening driven by the short end — a sign the market thinks the Fed has tightened too far.
Is the yield curve inverted right now?
No. As of October 30, 2026, both leading spreads — the 2s10s and the 10-year minus 3-month — are positive. The curve slopes upward, though the 10-year touched 5% earlier in October, a two-decade high, before easing to 4.94% at month-end.
Which spread should I watch?
Both. The 2s10s is the headline measure with the longest public track record; the 10-year minus 3-month is the one the New York Fed’s recession model uses, because the 3-month bill tracks the policy rate most directly.
Does an inverted curve cause a recession?
No. Inversion reflects expectations of future rate cuts and tight credit conditions; it is a symptom of restrictive policy and pessimistic growth expectations, not a cause of the downturn itself.
This article is for informational purposes only and is not financial advice. Yield data: U.S. Department of the Treasury. Spread calculations and the episode table are Macrometer’s own analysis.
