This is not a one-year fluke dressed up as a trend. It has survived recessions, recoveries, and a pandemic that scrambled almost everything else.

Georgetown CEW reports that prime-age workers with a bachelor’s degree face an unemployment rate of 2.9 percent, compared with 6.2 percent for workers whose education stops at high school. Georgetown’s report, published in October 2025, also says those degree holders earn 70 percent more at the median than workers with only a high school diploma.

The long view is just as hard to shrug off. Brookings reported last year that college graduates have had the lowest unemployment rate of any education group for the last 20 years. That is a boring finding in the best possible way. It keeps showing up.

The monthly BLS series backs that up. Bureau of Labor Statistics data, extracted in June 2024, show unemployment for adults 25 and older with a bachelor’s degree or higher hovering around 2.0 percent to 2.3 percent through most of 2022 and 2023, then at 2.1 percent in early 2024. Even the pandemic spike was temporary. The same series shows the rate jumping to 8.4 percent in April 2020, then falling back to about 2.0 percent by 2022.

The bigger point is not that college degrees make people recession-proof. They don’t. It’s that, across very different labor markets, degree holders keep ending up near the bottom of the unemployment table. The labor market has had plenty of chances to stop rewarding credentials. It has not taken them.

The pandemic made a mess of plenty of labor-market assumptions, so it is worth being careful here. Some of the comparisons should be handled with a light touch, because NCES flags that 2020 through 2022 estimates should be compared with prior years cautiously because of disruptions to survey interviewing and response rates.

Still, the directional story is clear enough. NCES says that for 25- to 34-year-olds overall and at each level of educational attainment, unemployment in 2023 was not measurably different from 2019, the year before the pandemic. That does not mean the years in between were calm, only that the labor market settled back into a familiar shape once the shock passed.

For younger workers, the hierarchy persisted. NCES says unemployment rates were generally higher at lower levels of educational attainment for both males and females. And while men had a higher overall unemployment rate than women in 2023, the same source says there were no measurable gender gaps at any specific level of educational attainment.

That is a useful detail because it trims away some easy narratives. The degree premium is not just a story about one gender doing better than another, or one cycle being unusually favorable. It shows up across groups, and it keeps showing up in the same direction.

Unemployment is the first half of the case. Earnings are the second, and they matter more than the campus debate sometimes allows.

Brookings says college grads come out ahead by age 26 or 27. That matters because it undercuts one of the most common ways the anti-college argument gets built: by focusing on the first few post-graduation years, when graduates are still finding footing and non-graduates may already be earning. Short horizons make many good investments look bad.

The broader earnings trend has moved in college’s favor, not away from it. Brookings says the college-to-high school earnings ratio was 2.1 in 2000 and rose to 2.7 by 2023. That is the opposite of a fading premium. Even more, Brookings says the gap has risen, not fallen, since 2000.

The payoff is not confined to the middle of a career. Brookings says the typical college graduate will be ahead by well over a million dollars by retirement age. That is a large number, but the more important point is how early the lead starts. Once the graduate pulls ahead in the late 20s, compounding does most of the work.

Cost is the obvious objection, and it is not a silly one. Brookings argues that the real cost of college has been flat for years now. If that holds, it changes the frame: the problem is not a relentlessly rising sticker price so much as a premium that has stayed large while the real cost has stopped climbing.

That still does not settle the individual case. Tuition, debt, and lost wages are real, and they do not show up neatly in a lifetime earnings ratio. The gross return looks strong. The net return depends on where a student goes, what they study, and how they finance it. The spreadsheet is never the whole story, even if people keep trying to make it one.

This is where the anti-degree argument becomes more credible. Not because the average degree is worthless, but because averages can hide a lot of ugly variation.

Georgetown CEW is helpful here because it does not treat a bachelor’s degree as one uniform product. Its October 2025 report looks at outcomes across undergraduate fields and majors for prime-age workers and recent graduates, including median earnings, graduate-degree premiums, and unemployment rates by field. That is the right way to think about this question. A degree is not a monolith, and never has been.

Some fields do plainly better than others. The report’s whole point is that major choice changes the labor-market outcome, sometimes a lot. That is why a clean yes-or-no answer to “is a college degree worth it” can be misleading. The better question is which degree, from which institution, at what price, and with what labor-market odds attached.

The evidence on recent graduates also explains why the skepticism feels stronger than the data would suggest. Brookings says the typical graduate does not pass the non-graduate in cumulative earnings until the late 20s. That means someone staring at a first job that barely pays the rent can be forgiven for feeling unimpressed by the grand theory of human capital. The return may be real and still feel far away.

And there is one more missing piece that matters. The research here is about unemployment, not underemployment. A graduate working in a job that does not use their degree still counts as employed. A barista with a philosophy degree does not vanish from the statistics; they simply disappear into them.

That is the strongest version of the skeptic’s case. Not that college is useless, but that the spread between degree paths is wide enough to make some programs excellent bets and others noisy ones. On that point, the skeptics are not wrong. They are just overstating the conclusion.

Strip away the slogans and the picture is fairly plain. College graduates are still the hardest group to knock out of the labor market, and their lifetime earnings premium has widened since 2000.

Georgetown CEW says bachelor’s degree holders aged 25 and over earn 70 percent more at the median than workers with only a high school diploma, and face much lower unemployment rates. Brookings says the earnings ratio between college and high school graduates moved from 2.1 in 2000 to 2.7 in 2023. That is not evidence of a credential losing value. It is evidence of one still being bid up.

There is still plenty of room for disappointment. Some degrees pay well, some do not, and debt can turn a good bet into a bad one if the numbers are wrong from the start. But the broad claim that college degrees are “useless” does not survive contact with the data.

The more useful takeaway is less ideological and more practical. The evidence supports college as an investment on average, but it also punishes lazy choices. Field matters. Price matters. Timing matters. So does the job market a student is actually walking into, not the one in the brochure.

That leaves the cultural skepticism in an awkward spot. It is not irrational. Tuition is expensive, debt is visible, and young workers often get hit with the bill before the payoff arrives. But if the question is whether a bachelor’s degree is broadly worth it, the last 20 years of unemployment data and the earnings evidence point in the same direction.

Not every degree is a good deal. But the degree itself, on average, still looks like one of the more reliable ways to lower unemployment risk and raise lifetime earnings. The argument has not disappeared. It has just gotten more specific.