Fundamental analysis · Advanced

Yield Curves and Inversions

An inverted curve has preceded most US recessions with lead times from six to twenty-three months. That makes it a regime signal, never an entry.

Last reviewed 15 September 2026 by the MarketPro research desk · Editorial policy

In short

A yield curve inverts when short-dated yields sit above long-dated ones, which is the bond market pricing in future rate cuts and, historically, a slowdown ahead. In the United States an inversion has preceded most postwar recessions, but on a sample of roughly eight observations and with lead times ranging from about six months to about twenty-three. It is useful as a regime indicator and unusable as a timing tool.

What to remember

An inversion prices in cuts, not a date. Lead times in the US record run from about six to twenty-three months.
The 2s10s and 3m10y spreads invert months apart. In 2022 they were roughly six months apart, so the start date depends entirely on which one is quoted.
A long yield is expected short rates plus a term premium. Bond buying compressed the second, so recent inversions are not measuring what pre-2008 ones measured.
The causal part is bank margins. Borrowing short and lending long gets less profitable when the curve inverts, so credit supply tightens.
Funding a ten-year at the overnight rate at minus 108bp costs 1.08% a year, which a 13 basis point fall in yield just covers.

Step 10 of 10 on Fundamentals & Macro

PreviousConsumer Confidence and Sentiment Indices NextEnd of the path

A yield curve plots the yields of one issuer's bonds against their maturities. It normally slopes upwards, because lending for ten years carries more uncertainty than lending for two. It inverts when short-dated yields sit above long-dated ones, which is the bond market saying it expects the policy rate to be lower in future than it is today. In the United States that shape has preceded most postwar recessions, with lead times varying by well over a year.

Minus 108 basis points, and the trade it implies

In early July 2023 the US two-year note yielded about 4.94% and the ten-year about 3.86%. The spread was roughly minus 108 basis points, the deepest inversion since 1981. With the overnight policy rate at 5.00% to 5.25% and one further hike still expected later that month, a ten-year yield of 3.86% says the market expected the average policy rate across the following decade to sit well below where it stood that week. That means cuts, and a lot of them.

Now price the position that view implies. Buy the ten-year, fund it overnight, and the carry is negative by roughly the spread, about 1.08% a year. At a duration of 8.5, a fall in the ten-year yield of 13 basis points produces 8.5 multiplied by 0.13, or about 1.1% in price, which covers exactly one year of that carry. Everything past that is gain and everything short of it is a slow bleed. That arithmetic is why an inverted curve is an expensive place to be early, and being early is the normal outcome.

Expectations plus term premium, and why a recent inversion is not the old measurement

A long yield decomposes into two parts: the average short rate the market expects over the life of the bond, and a term premium for bearing the risk of being wrong about it. An inversion can come from either. Expected cuts pull the first part down. A compressed or negative term premium pulls the second down, and produces a chart that looks identical while saying something quite different.

This matters more than it sounds. Term premium estimates from the standard New York Fed model were negative for much of the period between 2015 and 2021, a state with no real precedent before the era of large-scale bond buying, and central bank purchases were removing duration from the market precisely in order to compress it. Every inversion before 2008 happened with a positive term premium. Reading the recent ones as the same measurement as the old ones is the most common error in this subject.

2s10s, 3m10y, and a start date that moves by six months

There is no single yield curve. The spread in the headlines is usually the two-year against the ten-year. The spread in the academic work on recession probability is the three-month bill against the ten-year, which is the pair behind the New York Fed's published probability model. They do not invert together.

11 more sections in MarketPro

In 2022 the two-year against ten-year spread inverted briefly in early April and then decisively from early July. The three-month against ten-year spread did not invert until late October. So the answer to "when did the curve invert" moves by around six months depending on which pair is quoted, and every published lead time inherits that choice. When someone says an inversion preceded a slowdown by fourteen months, the first question is which spread they measured.

The re-steepening is the part most people miss. In several past US cycles the curve returned to a positive slope shortly before the downturn was eventually dated, because the central bank had started cutting the short end. Waiting for an inversion to end before turning defensive has therefore often meant turning defensive last.

Eight observations, and lead times from six to twenty-three months

The sample is tiny. There are roughly eight inversion-and-recession pairs in the US record since the mid-1950s. Eight observations is an anecdote with a chart attached, not a distribution.
The lead time is unusable for timing. Published estimates run from about six months to about twenty-three. A signal with a timing error longer than a year cannot carry a stop loss.
It says nothing about the path in between. Equity indices have risen substantially between an inversion and the eventual peak in more than one cycle. Being right about the destination and short for the journey is the classic way to lose money on a correct macro call.
It is a US statistic. The record elsewhere is weaker, and where a central bank has explicitly controlled the curve, as the Bank of Japan did under yield curve control, the shape is partly an administered price rather than a market opinion.

The false signals are worth naming as well. The curve inverted in 1966 with no recession following, and briefly in 1998 during the emerging market crisis. The inversion that began in 2022 then ran longer than any previously recorded, which by itself should temper any claim that the relationship is mechanical rather than a tendency with a small sample behind it.

The one channel that is genuinely causal

The curve is not only a forecast. Part of it acts on the economy directly, which is the best argument for watching it at all. Banks borrow short and lend long, so their net interest margin compresses when short rates exceed long ones. Lending standards tighten, credit growth slows and marginal borrowers get refused. That is a real mechanism connecting the shape to activity, and it operates with the same long and variable lag as everything else in credit.

There is a second use if you trade currencies. An inverted curve in a high-yielding currency is the market stating that the yield advantage is temporary, which is exactly the risk a carry position is exposed to and is not paid for. The rate expectations that price a pair are visible in the curve before they ever appear in a policy statement, and a carry trade entered against a deeply inverted curve is a bet that the bond market has the path wrong.

Using the shape as a dial, not a trigger

Check the three-month against ten-year spread monthly rather than daily. Record the level and, more usefully, the direction of change: a deepening inversion and a re-steepening one are different regimes, and the second has historically been the later warning of the two. Then pair it with data describing the present rather than expectations, weekly jobless claims and the new orders component of the manufacturing surveys being the usual choices.

What you do with the reading is a position sizing decision, not an entry. A regime you judge to be late in the cycle argues for smaller positions, tighter limits on correlated exposure and shorter holding periods, all of which is covered in the risk rules, and the cost of getting the size wrong is easiest to see on the drawdown calculator. A macro view that changes the size of what you trade is useful. A macro view that becomes a leveraged short with a two-year horizon is the thing this lesson exists to argue against.

Finish this lesson in MarketPro

The rest of Yield Curves and Inversions, a quiz that checks it landed, and your place kept in the curriculum.

11 more sections of this lesson
All 320 lessons, with no cap and no paywall
Free on any account, including a free one

In your browserContinue on webor continue on web

Same account as the app · No install · Free to start

Or scan with your phone camera to get the MarketPro app.

Download on theApp Store Get it onGoogle Play

Free account, no card needed.

Not investment advice. Past performance is not indicative of future results.

Run the numbers while this is fresh

Work it through on your own position size. Nothing leaves your browser and there is no sign-up: Position size calculator.

Written and reviewed by the MarketPro research desk

MarketPro is a trading-signal and trading-education app operated by Harajuku Holdings LTD (Cyprus). Our signal desk publishes and tracks every trade idea in-app, and every page here is checked against the app's live behaviour before publishing. Last reviewed 15 September 2026. How we produce signals and content · About MarketPro

FAQ

Yield Curves and Inversions: questions

What does an inverted yield curve actually mean?

It means investors accept a lower yield to lend for ten years than for three months or two years, which only makes sense if they expect the policy rate to be cut substantially at some point over that longer horizon. Since central banks cut when growth or inflation is falling, the shape is read as the bond market pricing in a slowdown. It carries no information about when, and the historical lead times vary by more than a year.

Which yield curve spread should I watch?

The three-month bill against the ten-year note is the pair used in the recession probability literature and in the New York Fed's published model, so it is the better one for that purpose. The two-year against the ten-year is what appears in headlines and inverts earlier. In 2022 the two spreads inverted roughly six months apart, which is why quoted lead times differ so much between sources.

How long after an inversion does a recession usually start?

In the US record the gap has ranged from about six months to about twenty-three, measured across roughly eight episodes since the mid-1950s. That range is wider than most trading horizons, and there are false signals in the record as well, including 1966 and a brief inversion in 1998. Treating the inversion as a regime input alongside current activity data is defensible; treating it as an entry trigger is not.

Trade · Learn · Earn

The whole academy is free in MarketPro

All 320 lessons, 12 guided paths, quizzes and a daily streak. Free on every account, in the app or your browser, with vetted free signals alongside it.

In your browserContinue on webor continue on web Download on theApp Store Get it onGoogle Play

Same account as the app · No install · Free to start

Not investment advice. Past performance is not indicative of future results.