Advanced4 min read

Yield Curve Inversion as a Recession Signal

Short yields above long ones has preceded every modern US recession. The record is genuinely striking and the lag is long enough to make it close to useless for timing.

MadStockAlerts Research · Updated August 28, 2026

What to take away

  • An inverted yield curve means the market expects policy rates to be lower in future.
  • It has preceded US recessions with an unusually consistent record.
  • The lag has historically been many months and has varied widely.
  • Equities have frequently continued rising well after an inversion appeared.
  • The un-inversion has sometimes been the more proximate signal.

MAD Academy Training Video · 0:45

The Signal With No Timetable

An inverted curve has preceded recessions with a good record and terrible timing, which makes it a warning rather than a trigger.

This lesson is part of a Stock Alerts + Tools plan.

See the library

What inversion actually says

Long yields reflect the expected average of short rates over the life of the bond. If ten-year yields sit below two-year yields, the market is saying it expects short rates to be materially lower in the years ahead than they are now.

Rates are usually cut because the economy weakened. Inversion is therefore the market pricing a future in which the Fed has had to ease, which is an indirect statement about growth. It is a market expectation, not a mechanism that causes anything.

The two shapes, and what each one is saying
The two shapes, and what each one is saying3%4%5%6%Ten-year money cheaper thanthree-month money3 mth1 yr2 yr5 yr10 yr30 yrYield

Scroll the chart sideways to see all of it.

  • Normal
  • Inverted
An upward curve says money costs more the longer it is lent, which is ordinary. An inverted one says the market expects the policy rate to be lower in years than it is now, and the usual reason to expect that is a weaker economy.

The record and its limits

  • It has preceded every US recession since the 1960s, which is a small sample of events.
  • The lag from inversion to recession has ranged from several months to around two years.
  • It has produced at least one widely cited false positive.
  • Equity indices have frequently made new highs between the inversion and the eventual downturn.

A small number of observations spread across sixty years of very different monetary regimes is a thin basis for confidence, however consistent the pattern looks. Eight or nine data points would not support a strong conclusion in any other field.

There is also a mechanism

Banks fund short and lend long, so an inverted curve compresses lending margins and can make credit less available at the margin.

This is a real channel by which inversion could contribute to a slowdown rather than merely predicting one, and it is one of the reasons the signal is taken seriously rather than dismissed as coincidence. A signal with a plausible mechanism is a better signal than one without, even when the sample is small.

Un-inversion

Curves have historically returned to a normal upward slope shortly before or during recessions, as short rates are cut faster than long ones.

The steepening that follows an inversion has sometimes been the more proximate signal, which is a further reason not to treat the inversion date as the event. The shape is a continuous variable and reducing it to a binary inverted-or-not discards most of what it is saying.

Which spread, and why they disagree

The yield curve is not one number, and the spread chosen changes both when it inverts and how good its record looks. Two are conventionally reported and they have inverted months apart from each other.

SpreadWhat it isCharacter
10-year minus 2-yearThe most quotedInverts earliest, driven by policy expectations
10-year minus 3-monthPreferred in much of the researchInverts later, and has the stronger statistical record
5-year minus 3-yearRarely quotedSometimes inverts first, in the belly of the curve
30-year minus 10-yearThe long endSays more about term premium than about the cycle

The disagreement is not noise. The two-year is dominated by where the policy rate is expected to go over that horizon; the three-month is close to where it is now. A curve where the ten-year is below the two-year and above the three-month describes an expectation of cuts that has not yet begun.

Base rates, and the honest version of the record

The claim usually made is that inversion has preceded every US recession since the 1960s. That is accurate and it is a smaller claim than it sounds, for reasons that are worth stating plainly.

  • The sample is small. Roughly eight to nine recessions in six decades is not a sample from which a reliable indicator can be established, whatever the hit rate.
  • The lag has ranged from about six months to about two years, which is wide enough to make the signal difficult to act on even if it is real.
  • There has been at least one inversion without a recession following, depending on which spread is used, so the record is not perfect.
  • The indicator is now watched by everybody, and a widely watched indicator can change the behaviour it was measuring.

There is also a structural argument that the relationship has weakened: a decade of large-scale asset purchases compressed the term premium, which is one of the components an inversion is meant to reveal. That does not make the signal meaningless, and it does mean the historical base rate was drawn from a different regime.

The mechanism is the part that does not depend on the sample. A curve where short money costs more than long money makes lending less profitable, since banks fund short and lend long, and lending standards do in fact tighten in those periods. That channel is observable directly, in the senior loan officer survey, rather than inferred from eight data points.

Educational content only. MadStockAlerts provides market commentary, research, and educational content. It is not personalized investment advice, and nothing here is a recommendation to buy or sell any security. Trading and investing involve substantial risk, including loss of capital. See the Risk Disclosure and Customer Agreement.