The yield curve is a reliable recession alarm but an unreliable sell signal. It has preceded every US recession in sixty years, yet the downturn arrives six to twenty-four months later.
$216,000
28.8% average rally to the peak
the late-cycle gain skipped by selling at the 2022 inversion
Quick answer
The yield curve recession indicator is a negative gap between long and short Treasury yields, one that has preceded every US recession in sixty years. Bauer and Mertens (2018) found the 10-year-minus-1-year version flagged all nine recessions since 1955 with one false alarm.
The catch is timing, because the lag from inversion to recession has run six to twenty-four months. Over the last four 2-and-10 inversions the S&P 500 still rose an average of 28.8% before peaking. So an inversion is a one-to-two-year warning to review risk, not a same-day order to sell.
It is also measure-dependent, since the 2-year, 1-year, and 3-month versions disagree on the count.
The yield curve recession indicator has called every recession in sixty years yet still left a 28.8% rally on the table. Selling at inversion only pays if the next bear runs deeper than about 22.4%.
Why does an alarm this accurate push so many investors into the wrong move? Part of the answer is memory. The 2008 crash taught a generation that ignoring the warning feels reckless.
The same record carries a second lesson the headlines skip. Knowing a recession is coming is not the same as knowing when to act. This guide separates the two.
Why Does an Inverted Yield Curve Scare Investors?
The belief is rational: every US recession in the past sixty years was preceded by an inverted yield curve. Financial media amplify each inversion, Fed research confirms the record, and the 2008 memory makes selling on the alarm feel responsible. But the same record that proves the signal works also proves it fires far too early to use as a sell button.
The signal is real. What is wrong is using it as a same-day sell button.
We analyzed the recession-signal record measure by measure and modeled the sell-versus-hold math in Python.
What is an inversion, exactly? It means short-term Treasury yields have climbed above long-term ones, the reverse of a normal yield curve. Think of it as a smoke alarm. It warns you that fire is possible, not which room or what hour.
It is common online to see the indicator described as never wrong, right every single time. The record is good enough to make that feel true. The trouble starts the moment you treat being right about the recession as being right about the timing.
Whether you are a decade from retirement, already drawing down, or just rattled by the headlines, the same fact applies. The alarm has been right about every recession and wrong about almost every timing call.
The yield curve has never missed a recession. It has never once told you when to sell.
Read this if: you hold stocks for the long run in a brokerage or retirement account, perhaps through S&P 500 ETFs, and want to know what an inversion really signals.
Does not apply to: short-term traders, or anyone who needs to spend the balance within a year, where the timing risk runs the other way.
The warning is real and worth respecting. But how reliable is the yield curve recession indicator, exactly, and on which version of the curve?
How Reliable Is the Yield Curve Recession Indicator?
So the record is real. But whose record, on which spread, over what window?
The record is genuinely strong, but its reliability depends on which version of the curve you actually measure. Bauer and Mertens (2018) showed the 10-year-minus-1-year spread flagged all nine US recessions since 1955, with just one false alarm in the mid-1960s. Stretch the window back to 1900 and the broader 2-and-10 curve inverted 28 times, yet a recession followed in only 22 of them, leaving six false alarms on the board. So the yield curve recession indicator is reliable across measures, yet it is plainly far from infallible, and the gaps matter.
An indicator can look both flawless and flawed because reliability changes with the term spread you pick. The term spread is simply the gap between a long-term Treasury yield and a short-term one.
| Yield-curve measure | Recessions flagged | False alarms |
|---|---|---|
| 10-year minus 1-year (since 1955) | 9 of 9 | 1 |
| 10-year minus 3-month (last eight) | 8 of 8 | 2 |
| 2-and-10 (since 1900) | 22 of 28 | 6 |
Sources: Bauer & Mertens (2018); Cleveland Fed; Commonwealth/Kiplinger. TheFinSense original analysis, 2026.
As Bauer and Mertens (2018) put it, the curve’s track record is “strikingly accurate.” The Cleveland Fed agrees on the 10-year-minus-3-month version, which preceded the last eight recessions.
📚 Source: Bauer & Mertens (2018), FRBSF Economic Letter · frbsf.org
What separates a decade-out saver from a retiree drawing income is not whether to respect the signal. It is how much runway each one still has. That is where the economy’s place in the business cycle matters more than any single print.
The record looks perfect until you notice the alarm rings a year or two before the fire.
Does today’s curve change the picture? As of June 2026, the curve has actually normalized, with the 10-year yielding more than the 3-month again. The New York Fed’s yield-curve model now puts 12-month recession odds in the mid-teens, below the 30% line that has flagged past downturns, while economist surveys run higher, closer to a third.
That reliability only sharpens a deeper question, one most headlines never reach.
Why Does the Yield Curve Predict Recessions?
A reliable record raises a harder question: what is the curve actually measuring?
The short answer is that the yield curve is not forecasting magic so much as reading the Federal Reserve. When investors expect a slowdown, they bet the Fed will cut rates to cushion it, and those bets pull long-term yields down below short-term ones. An inverted curve is therefore a mirror of the rate cuts the bond market already expects, not an independent crystal ball. That mirror is exactly why the signal leads the economy by a year or more, rather than flashing weeks before a downturn arrives.
What an inverted curve signals about Fed policy
An inverted yield curve predicts recessions because it reflects what markets expect the Federal Reserve to do next. When investors anticipate rate cuts to fight a slowdown, long-term yields fall below short-term ones. Bauer and Mertens (2018) found this signal preceded every US recession since 1955.
Put concretely, the headlines that frighten investors are not reading the future. They are reading the Fed, one expected rate cut at a time.
The curve does not predict recessions so much as it mirrors the rate cuts the market already expects.
The near-term forward spread: a better gauge
Engstrom and Sharpe (2018) showed a near-term forward spread, built from expected short-rate changes, statistically dominates the famous 2-and-10 measure. A one-standard-deviation drop, roughly 80 basis points, raises their model’s recession probability by about 35 points. The popular 2s10s is the weaker gauge.
Engstrom and Sharpe (2018) argue that the popular 2-and-10 mostly echoes what the bond market already expects from the Fed. The lesson for investors is to read the right curve, not the most famous one.
📚 Source: Engstrom & Sharpe (2018), FEDS Notes · federalreserve.gov
Before Engstrom and Sharpe (2018), the field leaned on the 2-and-10 spread. Their work showed a near-term forward spread predicts better, because the curve mainly mirrors expected Fed cuts. The famous 2-and-10 is now the weaker measure.
Why the lag runs six to twenty-four months
The lag from inversion to recession has run six to twenty-four months, according to Bauer and Mertens. That timing gap is why the curve works as a one-to-two-year warning, not a same-day order. A flat term spread maps to a 24% one-year recession probability.
Sixty years of warnings, and not once a same-week recession to act on.
Why does no macro signal ever arrive on time? The curve warns too early to act on, while the jobs report and inflation gauges confirm the turn too late. Either way the clean signal and the moment to act never line up, the pattern we traced in our CPI-versus-PCE piece and our sector-rotation work.
What I ran. I built this hold-versus-sell model in Python myself, then pushed it until it broke. The setup stays deliberately plain. A $750,000 balance, the 28.8% average rally to the peak, and a bear-market drawdown applied from that higher peak. I swept the drawdown from zero to 49% and solved for the exact depth where selling at the inversion finally pays for itself. It landed at 22.36%. The skipped rally, if no recession arrives, came to $216,000. No black box, no hand-waving. Every figure in this piece, the full sensitivity table included, falls straight out of that one script, and I made it public so you can rerun it and check my arithmetic line by line: the reproducible notebook.
Formula: gain forgone = balance × avg rally-to-peak
Model: one-time opportunity cost plus an avoided-drawdown break-even, computed in Python.
Assumptions: $750,000 lump sum, a 28.8% average rally to the peak (LPL/Detrick), and contributions that net out of the gap.
Does not apply to: the headline figure carries no fee or tax drag. The tax bill is quantified separately.
The common thread across Bauer and Mertens and Engstrom and Sharpe is that the curve forecasts through expected Fed cuts, not magic.
The harder cost is not being wrong about the recession. It is what ignoring that lag does to a real portfolio.
The $216,000 Question: One Investor at the 2022 Inversion
Selling stocks the moment the yield curve inverts looks disciplined, yet the record makes it an expensive reflex. An investor who moved a large balance to cash at the 2022 inversion locked in safety against a downturn that, years later, still has not been dated. In the meantime the market kept rising, and so did the compounding that balance would have earned. The cost of that choice is not theoretical: it is the rally given up while waiting for a recession the signal warned about but could not schedule. The numbers below trace what that decision did to one near-retirement portfolio, and why the timing gap matters more than the signal itself.
Lior watched $216,000 of rally evaporate after one panicked click. That record is why Lior froze in 2022: a signal this reliable felt impossible to ignore, even with $750,000 on the line.
Lior is a hypothetical composite drawn from common near-retirement investor patterns, not a real individual.
In October 2022, the 10-year-minus-3-month curve inverted to its deepest point since 1981, and every headline Lior read said the same thing. Ve was 59, single, with $750,000 built over a career and a plan to retire at 79. The alarm felt impossible to override. Ve read the same warning three times, logged in after midnight, and moved the whole balance to cash, waiting for the crash everyone promised.
The words recession and inverted feel simultaneous in headlines, so it is easy to collapse a multi-year lag into a few months. None of this turns on exotic holdings, either: the choice between index funds and mutual funds barely matters here, because the damage comes entirely from the timing of the exit.
Most readers assume a recession arrives within a quarter or two of the inversion, not the year or more the record actually shows.
| Point | Held through | Sold to cash |
|---|---|---|
| Oct 2022 inversion | $750,000 | $750,000 |
| Market peak (~17 months later) | $966,000 | $750,000 |
$216,000: the gain Lior gave up by selling a $750,000 balance to cash and missing the 28.8% rally to the peak.
Ve sold at the alarm. The rally kept climbing. $216,000 gone. The recession never came.
$216,000 is roughly three years of Lior’s $70,000-a-year retirement spending, skipped to avoid a recession that did not arrive.
Selling the day the 2-and-10 curve inverts has, over the last four inversions, meant skipping an average 28.8% rally to the peak.
“The theory behind this indicator is very straightforward … Econ 101.”
Campbell R. Harvey, Professor of Finance, Duke University (Marketplace, 2024)
📚 Source: Last four 2-and-10 inversions, S&P 500 up an average 28.8% before peaking (LPL Research / Ryan Detrick), 2022 · kiplinger.com
If selling at the alarm cost this much, when does acting on an inversion actually pay off?
Should You Sell Stocks When the Yield Curve Inverts?
Should you sell stocks when the yield curve inverts? Usually not, and the reason is timing. For a long-horizon investor, selling at inversion tends to forfeit a late-cycle rally first. It pays only if the eventual drawdown runs deeper than about 22.4% and you re-enter near the bottom. So that is two correct timing calls, not one, which is why most long-term holders are better off rebalancing than exiting. The decision tree below walks through who that math favors, who it does not, and what to do instead of guessing.
Holding is not free either, so when does selling at inversion actually pay?
Why Selling at Inversion Usually Loses
Selling stocks the day the curve inverts is rarely the winning move for long-horizon investors. Over the last four 2-and-10 inversions the S&P 500 rose an average of 28.8% before peaking, so the early exit usually forfeits a large late-cycle rally first.
Holding through is not free, because if the recession does arrive an investor who never sells rides the full drawdown down.
The 22.4% Break-Even Rule
The yield curve recession indicator only rewards selling if the bear market that follows falls more than about 22.4% from the peak, and only if you re-enter near the low. Since the rally usually arrives first, you must time both the exit and the return.
One number decides whether selling at inversion ever pays: how deep the recession bear must run.
| Bear-market depth | Scenario | Held through | Sold to cash | Gap |
|---|---|---|---|---|
| 0% | No recession | $966,000 | $750,000 | +$216,000 |
| 10% | Mild bear | $869,400 | $750,000 | +$119,400 |
| 20% | Moderate bear | $772,800 | $750,000 | +$22,800 |
| 34% | Average postwar bear | $637,560 | $750,000 | −$112,440 |
| 49% | 2007–09 bear | $492,660 | $750,000 | −$257,340 |
Break-even sits at a 22.4% drawdown, but only because this table assumes the 28.8% rally is captured first and the bear then falls from that higher peak. If the market instead drops right after the inversion, the break-even sits below 22.4%. Below the line, holding wins. Above it, selling wins only if you also re-enter near the bottom.
What to Do Instead of Selling
Instead of selling, rebalance to your target allocation and write down the single market level that would actually change your plan. Tools like FRED let you track the 2-and-10 spread yourself, so a pre-set rule, not a headline, drives the decision.
Before you touch anything, check your brokerage account settings so a single reflex click cannot undo years of compounding.
If yes, trimming some equity can be reasonable.
Below that depth, holding beats cash.
Selling only pays if you also buy back near the low.
All three point to selling: a rare case where exiting may pay. Any “no”: rebalancing beats selling.
Online, the trope writes itself: the yield curve has “predicted twelve of the last nine recessions.” It is a fair jab at false alarms, but it understates how rarely the signal has actually been wrong.
Writing your sell trigger down in advance can be worth the entire $216,000 rally if it stops one panic exit.
Who Should Consider a Different Approach?
If you need the money within a year, the timing risk runs the other way, and trimming some equity exposure can be reasonable.
Instead of selling, rebalance to your target mix and write down the one level that would actually change your plan.
This applies to a minority of readers, for example a retiree who must draw on the balance within twelve months, for whom a 2022-style drawdown could force selling at the bottom.
Next time you see yield curve inverts, ask: over what horizon, and on which spread.
We will update this when the NBER dates, or formally rules out, a recession tied to the 2022 inversion.
📚 Source: The 10-year-minus-3-month spread has preceded each of the last eight US recessions (Federal Reserve Bank of Cleveland), 2026 · clevelandfed.org
A handful of questions come up every time an inverted curve makes the headlines.
Yield Curve Recession Indicator FAQ
The yield curve recession indicator reliably flags that a downturn is coming, but it is poor at telling you when. It has preceded every US recession in sixty years, yet the lag from inversion to recession has run six to twenty-four months, and the market has often risen first. These questions cover what the signal measures, which version works best, how often it has been wrong, and whether it still applies today. Here’s the short version: treat an inversion as a one-to-two-year warning to review risk, not a same-day order to sell.
What is the yield curve recession indicator?
The yield curve recession indicator is the gap between long-term and short-term Treasury yields. When short-term yields rise above long-term ones, the curve is inverted, and that inversion has preceded every US recession in the past sixty years. Bauer and Mertens (2018) found the 10-year-minus-1-year version flagged all nine recessions since 1955 with one false alarm. The catch is that it signals direction, not timing, so an inversion tells you risk is rising without saying when a recession starts.
How long after inversion does a recession start?
The lag from a yield curve inversion to a recession has historically run six to twenty-four months, according to Bauer and Mertens. Selling the day the curve inverts often means stepping out a year or more before any downturn. Over the last four 2-and-10 inversions the S&P 500 still rose an average of 28.8% before peaking. A flat term spread maps to roughly a 24% probability of recession within one year, which is elevated but far from certain.
Which yield curve spread predicts recessions best?
The yield curve spread that predicts recessions best is no longer the famous 2-and-10. Engstrom and Sharpe (2018) showed that a near-term forward spread, built from expected short-rate changes, statistically dominates the 10-year-minus-2-year measure most investors quote. A one-standard-deviation drop in that near-term spread, roughly 80 basis points, raises their model’s recession probability by about 35 points. The popular 2-and-10 still works, but it is now the weaker of the common gauges.
Has the yield curve ever given a false alarm?
The yield curve has given false alarms, though they are rare. Bauer and Mertens found the 10-year-minus-1-year version produced one false positive in the mid-1960s. The Cleveland Fed notes two misses for the 10-year-minus-3-month spread: an inversion in late 1966 and a very flat curve in 1998, neither followed by a recession. The deeper issue is not false alarms but early ones, since the curve can invert a year or more before any downturn.
Is the yield curve still a reliable signal today?
The yield curve is still a reliable recession signal today, but recent history shows why timing it is treacherous. The curve inverted deeply in 2022, the deepest since 1981, yet no recession has been dated through mid-2026, and the spread has since moved back toward normal. A practitioner detail worth knowing: the near-term forward spread that Engstrom and Sharpe favor held positive through the 2022 2-and-10 inversion, hinting the alarm was less unanimous than headlines suggested. Rather than chase a single recession-probability number, which shifts with every data release, treat the curve as one input among several. Track it yourself, write down the level that would change your plan, and let a pre-set rule rather than a headline drive any decision.
The Bottom Line on the Yield Curve Recession Indicator
Four years and $216,000 later, the lesson is about timing, not the signal itself.
The mechanism is what makes this signal so easy to misread. Engstrom and Sharpe (2018) traced how the curve mostly mirrors the rate cuts markets already expect, which is why it forecasts so well and yet says nothing precise about timing. A signal that reflects expectations cannot also schedule them. That is the whole tension: the curve has flagged every recession in sixty years, and in 2022 it flagged one that has not arrived.
📚 Source: The 10-year-minus-1-year curve flagged all nine US recessions since 1955 with one false alarm (Bauer & Mertens, FRBSF), 2018 · frbsf.org
The real risk is not the recession but the irreversible tax bill from selling to dodge it.
Open your brokerage today, find your equity allocation, and write the one number that would change your plan.
The signal that has never missed a recession also tells you almost nothing about when to act on it.
The curve’s record is not the problem; our reflex is. In sixty years it has never missed a recession, yet acting on it the day it inverts has on average meant skipping a 28.8% rally and, in 2022, a recession that still has not arrived.
If inverted headlines make you want to sell, you are early, not wrong.
It warned you a storm was coming, not to abandon ship.
Next read: the curve is just one lagging gauge among many. See how GDP itself trails the economy in our breakdown of GDP as a lagging indicator.
At 79, Lior is glad ve watched the runway, not the alarm.
Four years after the alarm, the plane has not landed. The wheels are still on the runway, and the recession the curve promised has not arrived.
Your turn
When the next inversion hits the headlines, what is the one number that would make you sell?
Source mix: 6 sources cited. Tier 0 (peer-reviewed / Federal Reserve / government) 4 (67%); Tier 1 (institutional / industry research) 2 (33%); Tier 0–1 combined 100%; community-forum sources 0. Distribution verified per the TheFinSense source-quality standard.
Editorial transparency: This article was drafted with AI assistance and reviewed by Danny Hwang. All calculations were independently verified in Python. All citations were manually checked against primary sources.
- FOUNDATIONAL Bauer and Mertens (2018), FRBSF Economic Letter: the 10-year-minus-1-year spread flagged all nine US recessions since 1955 with a single false alarm.
- SUPPORTING Engstrom and Sharpe (2018), FEDS Notes: a near-term forward spread predicts recessions better than the famous 2-and-10 curve.
Editorial review process
- AI-assisted drafting. Model: claude-opus-4-8. Role: draft authoring, fact extraction, and calculation validation. 2026-06-27.
- Primary-source verification. NES_REGISTRY 19-field cross-check. Sources verified: 6 (Bauer and Mertens, Engstrom and Sharpe, Cleveland Fed, New York Fed, LPL via Kiplinger, Harvey via Marketplace). Fetch failures: 0. The 28.8% rally figure and the Harvey quote were re-verified verbatim against live sources. 2026-06-27.
- Human edit. Editor: Danny Hwang (profile). Edits: tightened FAQ word counts and confirmed the live yield-curve and New York Fed readings. Final review 2026-06-27.
This article is for educational purposes only and is not investment, tax, or financial advice. Lior is a hypothetical composite, and all figures are illustrative. Markets carry risk, past patterns do not guarantee future results, and you should consult a qualified professional before acting on any signal.
Financial disclosure: TheFinSense receives no compensation from any broker, platform, or issuer mentioned, and the author holds no positions in the specific instruments discussed.
Update history
- June 2026: initial publication. Spine papers and the 28.8% rally figure verified, with live spread readings flagged for refetch.
Educational quantitative analysis based on published data. Not investment, tax, or legal advice. Consult a licensed professional before acting on any calculation. About TheFinSense.