The comparison that broke
The New York Fed tracks the labour market for recent college graduates, and in the second quarter of 2026 it recorded something it had not recorded before: graduates aged 22 to 27 were unemployed at a higher rate than the base working population. 5.6% against 4.2%. In data reaching back to 1990, that ordering had never inverted.
It is a real inversion and it deserved the coverage it got. What it does not support is the conclusion drawn from it - that a degree has stopped paying. To read it that way you have to accept that 22-to-27-year-olds and all adults are alike in every respect except education, and they are not. One group is at the start of working life, where unemployment is always highest because people are entering, switching and being let go first. The other is an average that includes everyone at the peak of their career.
Comparing them measures age at least as much as education. The inversion tells you that early-career hiring has deteriorated badly relative to the rest of the market. It cannot tell you what a degree is worth, because the comparison never held education constant.
| Population | Unemployment |
|---|---|
| Recent graduates, 22 to 27 | 5.6% |
| Young workers, 22 to 27, no degree | 7.8% |
| College-educated workers, 22 to 65 | 3.1% |
| All adults | 4.2% |
The comparison that holds
Hold the age bracket still and the picture changes. Young workers without a degree were unemployed at 7.8%. Graduates in the same bracket: 5.6%. The degree is worth 2.2 points of unemployment to someone the same age - and against college-educated workers across the full 22-to-65 range, at 3.1%, the gap is not education either. It is experience, which the graduates do not have yet and will.
So the honest version of the headline is narrower and duller: early-career hiring is bad. It is bad for everyone entering the market, and it is less bad for those entering with a degree. That is a smaller claim than "the degree is finished", and it is the one the numbers actually make.
The wider market gives the same reading. US payrolls fell by 23,000 in July 2026 with unemployment at 4.1%, and entry-level hiring specifically is down about 7% year on year. Nothing there is a story about credentials. It is a story about the door being narrower for whoever arrives at it first.
The number that is actually alarming
Set the unemployment rate down and pick up the other one. 42% of recent graduates are underemployed - working, but in jobs that did not require the degree they hold. That figure edged up in the second quarter, and unlike the inversion it is not an artefact of comparing unlike groups. It is one population, measured against its own qualification.
It also resolves a contradiction that has been confusing this story. A separate survey found 77.2% of recent graduates landed a role within three months, up sharply from 63.3% the year before, and that was reported as evidence the market is fine. Both things are true at once: they are finding work quickly, and close to half of that work does not use what they trained for. The placement figure answers whether they were hired. The underemployment figure answers as what.
That is the sentence this edition exists for. A degree still gets you hired. It has stopped reliably getting you the job you took the degree for - and no single unemployment rate can show you that, which is why the number in circulation is the wrong one.
| Figure | Value | What it actually answers |
|---|---|---|
| Placed within three months | 77.2% | Did they find work at all |
| Unemployment, 22 to 27 | 5.6% | Are they without work right now |
| Underemployment | 42% | Is the work worth the degree |
Where the door is still open
The freeze is not uniform, and the exceptions are specific enough to act on. Entry-level postings in health care rose by 13 percentage points against a declining market. Skilled trades, cybersecurity and business operations are also taking graduates while the sectors that spent a decade recruiting them are not.
The sector that has pulled back hardest is tech - the one whose entry-level pipeline was, for most of the last decade, the default destination for a technical degree. Whether that is displacement by automation or simply a hiring correction after over-expansion is contested, and this desk is not going to settle it with a posting count. What is not contested is the direction.
For anyone reading this while applying: the useful move is not to send more applications into the sectors that are closed. It is to notice which doors moved, and that the answer differs by sector far more than it differs by candidate.
We publish how our own matching is tested, including where it is weakest, at our bias audit. If a number here is wrong, that page is where the correction should start.