Do College Degrees Still Pay Off?

1. The Crack in the Foundation

For most of the twentieth century, the deal was simple and broadly honored: go to college, get a degree, get a job. The credential served as a socially legible signal that you were trainable, disciplined, and minimally educated. Employers trusted it. Families sacrificed for it. Society organized around it.

That deal is not broken. But it has been significantly renegotiated — and millions of students entering college today are doing so without anyone showing them the new terms.

In November 2025, researchers at the Federal Reserve Bank of Cleveland published what may be the starkest single data point of this shift: young college graduates between ages 22 and 27 are now spending more time unemployed than job hunters with only a high school diploma. Their monthly “job-finding rate” — the share who either land work or exit the search — was 37.1%, compared to 41.5% for their peers who stopped at high school. The gap between degree-holder and non-degree-holder employment prospects, they wrote, is at its smallest point since records were reliably tracked.

That quote, from the chief economist of the Burning Glass Institute — a nonprofit labor analytics firm, not an axe-grinding pundit — ran in a CNBC report in November 2025. He wasn’t being provocative. He was reading data.

The Burning Glass Institute’s own July 2025 analysis found that one year after graduation, 52% of the class of 2023 were working in jobs that didn’t require a degree at all. Their description: a “flipped pyramid.” Demand for experienced, senior-level workers remains relatively robust. The rungs at the bottom of the ladder — the entry-level positions that used to turn graduates into competent professionals — are disappearing.

This isn’t simply a bad economy for everyone. The broader labor market is soft, yes — but the specific deterioration at the entry level of degree-requiring white-collar work is happening faster and more structurally than the general picture suggests. Young graduates aren’t just caught in a cyclical downturn. They’re caught in a structural one, and these tend not to reverse when the economy picks up.

2. AI Is Already Eating Entry-Level Jobs

The technology story here isn’t speculative. It’s not “AI might disrupt white-collar work.” It’s already doing it, and entry-level college graduates are the first casualties.

SignalFire, a data-driven venture capital firm, found that between 2019 and 2024, there was a 50% decline in new roles for people with less than one year of experience at top tech companies. In 2024 alone, hiring of recent graduates by big-tech companies fell 25% compared to 2023, according to SignalFire’s research published by TechCrunch. In sales, design, HR, engineering, and legal departments — fields where graduates once landed their first jobs — the traditional footholds are shrinking.

Companies have been explicit about why. In 2025, executives at several large employers publicly stated they were using AI to replace entry-level functions. The rationale is straightforward: a junior analyst who spent four years learning to research, summarize, and format reports can now be partially or wholly replaced by AI tools that do those tasks faster and without benefits. That doesn’t eliminate the need for analytical judgment, but it does eliminate the headcount needed to perform the labor-intensive parts of the job.

The Class of 2025 submitted more job applications per person than the Class of 2024 did — 10 versus 6, on average, according to the National Association of Colleges and Employers (NACE). They received fewer offers: a mean of 0.78 vs. 0.83 for the previous class. More effort, worse outcome. That trajectory doesn’t usually reverse without structural change.

About half of employers — 51%, the highest share since 2020–21 — rated the job market for this year’s college seniors as “poor” or “fair.” That’s not from students complaining online. That’s from the employers themselves.

It’s worth noting what AI automation does not do: it doesn’t eliminate all knowledge-work jobs. The roles that remain and grow are precisely the ones that require judgment, client relationships, creative direction, and the ability to supervise and prompt AI systems. These roles skew toward experienced workers, which compounds the problem for new graduates who can’t get experience because there are no entry-level roles to gain it in.

3. The Wage Premium: Still Real, Often Misleading

The strongest argument for college is also the most misused one. Yes, bachelor’s degree holders earn significantly more, on average, than those with only high school diplomas. The Federal Reserve Bank of Minneapolis’s September 2025 analysis put average lifetime earnings at $2.8 million for degree-holders versus $1.6 million for high-school-only workers — a gap of $1.2 million. The New York Fed, in a widely cited April 2025 report, found a median 12.5% annual return on investment from a college degree, comparing favorably to long-run stock market returns.

These figures are real. They’re not fabricated by higher-education lobbyists. The problem is they’re aggregate averages that obscure extraordinary internal variation — variation that matters enormously when you’re the person making a $42,000 borrowing decision at age 18.

Put differently: the college earnings premium is real in aggregate, but at least one in four college graduates earns a return on their education investment of less than 2.6%. That’s below inflation. For those students, the degree didn’t just fail to pay off — it actively destroyed wealth, particularly when combined with significant borrowing.

The Minneapolis Fed’s September 2025 analysis is careful to emphasize that this is an average, not an individual prediction. Major, institution, graduation timing, and labor market conditions at graduation all influence outcomes. A computer science degree from a state school completed in four years with minimal debt carries a fundamentally different risk profile than an art history degree from a private college completed in five and a half years with $80,000 borrowed.

Deming’s research is compelling and shouldn’t be dismissed. But it’s important to note what it captures: a cohort of workers who graduated before AI automation, credential inflation, and the current entry-level market disruption took hold. Whether that trajectory holds for workers entering the market in 2025 and 2026 is genuinely uncertain. The historical data is backward-looking by definition.

4. Credential Inflation and the Employer Reversal

One of the stranger dynamics of the past few years has been the pendulum swing on degree requirements among employers. When labor was scarce (roughly 2021–2023), companies dropped credential requirements to widen their talent pool. When labor loosened, many put them back — and in some cases, added new ones.

The Cengage Group’s 2025 Graduate Employability Report found that 71% of employers now require a two- or four-year degree for entry-level roles, up sharply from 55% in 2024. When employers have more applicants than jobs, they use credentials as a filtering mechanism — not necessarily because the degree predicts job performance, but because it’s a low-cost screen for discipline and persistence.

There’s a darker layer to this, too. A 2025 survey of over 150 companies by recruiting intelligence firm Veris Insights found that more than a quarter — 26%, up from 17% in 2022 — were recruiting exclusively from a brief selection of schools. The credential isn’t just “do you have a degree.” For an increasing number of employers, the relevant question is “did you attend one of our target schools?”

That said, the employer landscape is genuinely bifurcated. Companies like IBM, Google, Apple, Tesla, and Bank of America have made high-profile moves to eliminate degree requirements for many roles — and in IBM’s case, now fill roughly half of open roles without requiring one. A 2025 survey found 81% of U.S. companies embracing skills-based hiring in at least some capacity, compared to 73% in 2023. But the critical qualifier is “in at least some capacity.” The positions where degree requirements have been dropped tend to be specific, and many employers retain them broadly while publicly signaling openness to alternatives.

Meanwhile, a 2025 survey by Intelligent.com found that 53% of hiring managers said their company had eliminated bachelor’s degree requirements for some roles — and 76% said they favored experience over education in at least some situations. These numbers coexist comfortably with the Cengage finding that degree requirements are going up. The truth is that the labor market for college graduates is not moving in one direction. It’s fragmenting.

5. The Debt Trap: When the Numbers Stop Working

The financial case for or against college cannot be evaluated without confronting the debt reality. Total U.S. student loan debt crossed $1.83 trillion as of 2025, according to multiple financial tracking sources. Average federal student loan debt per borrower stands at approximately $39,375, with total debt including private loans reaching roughly $42,673 for many graduates.

These aren’t abstract numbers. They translate into monthly payment obligations that structurally reshape the first decade of adult financial life: delayed home ownership, constrained emergency funds, deferred retirement contributions, and reduced ability to absorb career setbacks.

In July 2025, the One Big Beautiful Bill Act was signed into law, ending the Grad PLUS loan program for borrowers beginning programs on or after July 1, 2026. While the stated rationale was controlling ballooning graduate debt, the practical effect for many students will be restricted access to graduate programs — the very programs that have become increasingly necessary to differentiate in a credential-inflated market.

The ROI math on college debt is genuinely favorable for the median student who graduates in four years with average debt in a market-rate major. It becomes significantly less favorable for anyone who takes five or six years, borrows heavily for a low-wage field, or attends a high-cost private institution for a program without strong placement rates. And for the bottom quartile of graduates — at least 25% of everyone who gets a bachelor’s — the Federal Reserve’s own data suggests the return doesn’t justify the investment.

6. What’s Actually Working Instead

The conversation about whether to go to college has, for too long, been framed as binary: go to a four-year university, or don’t go at all and make do. That framing is now badly outdated. A growing ecosystem of alternatives has matured to the point where they represent genuine career paths, not fallback options.

Coding and Tech Bootcamps

The bootcamp model, originally built for software development, has expanded into data science, cybersecurity, UX design, digital marketing, and product management. Programs typically run 12–24 weeks, cost $10,000–$20,000 (a fraction of a semester at many universities), and many offer income-share agreements that delay payment until the graduate is employed. Starting salaries for bootcamp graduates in tech roles average $50,000–$75,000, with placement rates that — at the better programs — are genuinely competitive with traditional CS degrees.

General Assembly, Flatiron School, and others have built substantial placement track records. The key variable, which holds across all credential types including university degrees, is the quality of the institution and the relevance of the program to actual employer demand.

Industry Certifications

Google, AWS, Microsoft, and CompTIA have all built certification pathways that are now taken seriously by employers, particularly in cloud infrastructure, IT support, cybersecurity, and data analytics. Google’s Career Certificate programs, available through Coursera, cost under $500 and have resulted in hires at companies including Walmart, Infosys, and Bank of America. They’re not a universal substitute for a degree — but in specific technical domains, they’re increasingly a sufficient one.

Apprenticeships

The apprenticeship model — which pairs paid employment with structured training — has expanded well beyond the traditional trades into healthcare, finance, and technology. The core advantage is that it eliminates both the debt and the opportunity cost of four years out of the workforce. Apprentices earn while they learn, create direct pathways to employment with sponsoring organizations, and often outperform their university-educated peers on job-specific skills because they’ve been doing the work, not studying it.

It’s worth being honest here: there are careers where a four-year degree is not optional. Medicine, law, engineering requiring licensure, academia, and many government positions maintain strict credential requirements that aren’t going away. Anyone planning on those tracks doesn’t have a genuine choice about undergraduate education. The question of alternatives is most relevant for the large middle ground of professional work — knowledge work, tech, business, design, marketing, and the creative industries — where the credential has historically served as a filter rather than a genuine training requirement.

7. Which Degrees Still Pay — and Which Don’t

If there’s one thing that both the skeptics and the defenders of higher education agree on, it’s this: not all degrees are equal, and the variance in outcomes has widened. The question isn’t just “is college worth it?” It’s “is this specific degree, from this specific institution, worth it for this specific person entering this specific labor market?”

The Georgetown Center on Education and the Workforce has documented extensively that 28% of associate degree holders and 16% of high school graduates earn more than the median bachelor’s degree holder. That’s not a fringe outcome. A licensed electrician in a major metro area with a decade of experience regularly earns more than an English literature graduate from a private university who went on to work in publishing or nonprofits. The trades have been systematically undervalued in cultural prestige relative to their financial reality.

8. Making the Call: A Framework for 2026

Given the complexity, here’s a practical decision framework — not a prescription, but a set of questions worth answering honestly before signing on for $40,000 or more in debt.

The honest answer, in 2026, is that college remains the right choice for many people and the wrong choice for a significant minority who would be better served by alternative paths. The problem isn’t college itself — it’s the cultural insistence that it’s the only legitimate route to a good career, combined with a system that prices itself as though it holds a monopoly it no longer fully holds.

Families with the ability to fund education without significant borrowing face a different calculus than families who can’t. A student who attends a public university, graduates in four years with modest debt, majors in a high-ROI field, and enters a career where the degree is genuinely relevant is still making a sound investment by any measure. A student who borrows $90,000 for a low-ROI major from a private institution because “it’s what you do” is playing a very different game.

10. Final Thoughts

The headline of this article is deliberately provocative, but the argument isn’t. College degrees are not becoming worthless across the board — but for a growing segment of students, a specific combination of factors is making them a poor investment: the wrong major, at the wrong school, with too much debt, in a labor market that’s structuring away entry-level positions. That combination is more common than the higher education industry would like to acknowledge.

The more honest framing is this: the automatic value of a college degree — the idea that any degree, from any school, in any field, justifies any price — is what’s becoming worthless. The discriminating value of a well-chosen degree, in the right field, at reasonable cost, from an institution with genuine career infrastructure, remains real.

The AI disruption is not a background hum. It’s already measurable in hiring data, employment rates, and the explicit statements of employers. The bottom rung of the white-collar ladder is eroding, and that rung is where almost every college graduate’s career begins. There’s no obvious mechanism for this to reverse; as AI capabilities improve, more of the tasks that justified entry-level hiring will be automated, and the runway problem will worsen.

What this calls for is neither panic nor dismissal. It calls for honest analysis — exactly the kind that most colleges, for understandable financial reasons, are reluctant to provide. High school seniors and their families deserve the full picture: the aggregate earnings data, the distribution underneath it, the debt reality, the alternatives, and the honest risk profile of specific programs at specific institutions. Armed with that, the decision can be made rationally rather than by default.

The fact that this conversation is finally happening — visibly, in mainstream publications, with real data — is, in a strange way, encouraging. The information environment around higher education has been systematically skewed toward the institutional interest for decades. It’s belatedly correcting, and students will be better equipped for it.