A fresh college degree came with a quiet edge in the job market. New graduates were likely to do a better job than the average worker, and this advantage persisted for as long as anyone tracked it. Not anymore. They now face higher levels of unemployment than the overall workforce, and the gap is the largest on record.
What makes it weird is the timing. The reversal didn’t start with ChatGPT, and it didn’t start with the pandemic. It started in early 2019, before it was even on anyone’s radar.
The chart tracks a single number, the unemployment rate of recent graduates minus the rate for all workers. Below the zero line grades overtake the average worker, and above it they lag behind.
This comparison is worth noting. “All workers” is the entire American labor force, and most of them are older and more experienced than a new graduate, so a new graduate starts out at a natural disadvantage. Degrees canceled out that loss for decades. Now it is not so.
Degree was a buffer for decades
A recent graduate almost always has a better chance of getting employed than the average employee. The Sahara was real, and it was biggest when the economy was at its worst.
The rise peaked in the depths of the Great Recession. In the mid-2010s, graduate unemployment was around 7%, while the overall workforce was near 10%, the largest gain ever. The recession hit construction and manufacturing first, sectors that rely heavily on workers without degrees, so diplomas were most valuable when jobs were disappearing.
Before AI and before COVID, the edge disappeared in 2019
The gap crossed zero in February 2019 and the 12-month average has remained positive every month since then. That timing defies both easy explanations. This flip was several years before the generative-AI boom and a year before Covid.
It was a slow structural drift, not a sudden shock. The Cleveland Fed still traces the erosion. Young graduates’ job search advantage has been declining since around 2000, and their lead over high-school workers leveled off around 2019. The pandemic also did not cause this, and the spike in 2020 is clear proof of this. When unemployment rose that year, both lines rose together, so the gap remained roughly constant through 2020 and 2021. The fine was already there. The lockdown pushed it down to a huge number.
This is now the largest gap on record
Unemployment for all recently graduated workers at the beginning of 2026 was 5.6% versus 4.2%, the largest difference on record. Its prevalence has increased almost every year since the beginning of 2019.
What makes the record even stranger is the background. This is not a recession story. Overall unemployment stands at 4.2%, yet new graduates are struggling. Each previous increase in their unemployment coincided with the Great Recession. This is theirs only.
Unemployment is only half the picture. Of those new graduates who do have jobs, about 41% are underemployed, working roles that never required a degree in the first place.
Remote work, or AI?
So what’s broken? The honest answer is that economists are still debating it. In June 2026, the New York Fed made the case that remote work, not AI, is the main culprit, and attributed the nearly 64% increase in youth-graduate unemployment to it. The Fed argues that employers are wary of hiring inexperienced people into remote roles, where the on-the-job advice that turns a new graduate into a productive worker is harder to provide. The timing is opportune, as the ascent began before AI took hold.
Stanford researchers look at AI’s fingerprints anyway. Their study found that early-career workers aged 22 to 25 in the most AI-exposed jobs saw a nearly 16% decline in employment by the end of 2022, a decline that persisted even after stripping out remote-friendly roles. Both can be true. Either way, entry-level notches are being taken out, and the technology is at its sharpest. The unemployment rate for recent computer science graduates is the highest of any major, as the number of CS degrees added to the shrinking pile of vacancies has more than doubled.
On-ramp broken, no degree
This is an entry-level problem, not proof that a degree has stopped paying off. Old degree holders are doing well. According to the Bureau of Labor Statistics, there was only 2.8% unemployment among American workers age 25 and older with a bachelor’s degree or more in April 2026, significantly lower than the rate for high-school graduates. The damage is almost entirely concentrated in youth. Since 2019, recent graduates have borne the brunt of the increase, while unemployment for older degree holders has barely increased, according to the St. Louis Fed, and the New York Fed still estimates lifetime returns near 12.5%.
New graduates are not far behind their college dropout peers. Unemployment among young workers without a degree is 7.2%, well above the 5.6% among graduates. One degree is still better than no degree. Now what it doesn’t do is beat the average.
None of this has been decided. The Economic Policy Institute argues that the picture is more mixed, with the college wage premium flat for years and new graduates still performing no worse than young workers without degrees. Degrees still open doors. It no longer gets you through faster than everyone else.
How was this chart created
This chart was created by an AI agent and graded against the Tufte Test, a data visualization quality standard created by GoodEye Labs at TrueSight.
Data Source: The labor market for recent college graduates is constructed from the Current Population Survey from the Federal Reserve Bank of New York, the U.S. Census Bureau, and the Bureau of Labor Statistics. “Recent graduates” are non-students aged 22 to 27 with at least a bachelor’s degree, “young workers” are aged 22 to 27 without a bachelor’s degree, and “all workers” are aged 16 to 65. The cleaned dataset is available here.
<a href