Why Employers Say They Want Skills, and Still Hire for Degrees

The question 0:01

Employers say they hire for skills. So why, when the degree requirement comes off the job ad, does almost nothing change about who gets hired?

This is WAC Signal, from the World Assessment Council. Evidence and ideas for the changing world of learning.

Episode three. Why employers say they want skills, and still hire for degrees.

The signal 0:36

On the first of July 2026, a US law took effect that lets low-income students spend federal Pell Grants on short job training programs. The bar is high. Programs must show completion and job placement rates of at least seventy percent each, and their tuition and fees may not exceed their graduates' value-added earnings. By the first of October, twenty-one programs had been approved, in six states.

Virginia has tied funding for noncredit workforce credentials to credential attainment since 2016, and is often held up as the model. When it ran its full slate of two hundred and fifty credit and noncredit workforce programs through the new filters, only six were eligible. That is according to the state's Randall Stamper, as reported by Washington Monthly.

Supply is not the problem. Credential Engine counts one point eight five million unique credentials on offer in the United States. In the second quarter of 2026, unemployment among recent US graduates was about five point six percent, and forty-two percent were in jobs that typically do not need a degree.

A job posting can lose the word degree in an afternoon. But the hiring manager still needs evidence that is quick to read and safe to trust, and most skill claims do not supply it. Our reading: the missing piece is not skills. It is proof.

The evidence 2:19

Start with what employers did. The most careful study of hiring records comes from the Burning Glass Institute and Harvard Business School. It tracked about eleven thousand three hundred roles at large US firms that dropped a degree requirement from 2014 onwards, and matched them to real hires using the career histories of sixty-five million workers.

After the requirement came off, the share of hires without a bachelor's degree rose by about three and a half percentage points. But few roles dropped the requirement at all. Net, that is roughly ninety-seven thousand workers a year, out of seventy-seven million hires. Fewer than one in seven hundred.

The firms split three ways. Thirty-seven percent hired materially more people without degrees. Forty-five percent, the largest group, changed the posting and little else. The other eighteen percent gained early, then slid back. The authors' explanation for the laggards is that managers under pressure to fill a costly vacancy fall back on the degree as a shortcut.

And the postings themselves have edged back. Indeed Hiring Lab reports that nineteen point three percent of US postings on its site required a bachelor's degree or more in November 2025, up from sixteen point six percent two years earlier.

So employers are not leaning on the new credentials. Should they? The American Enterprise Institute and Burning Glass examined more than twenty-three thousand credentials against career histories. About twelve percent delivered significant wage gains that earners would not otherwise have had. Most credentials, the authors say, rest on the provider's own assurance of mastery, or only of completion.

The ones that pay tend to share features. At Virginia's community colleges, earning an industry-recognized, third-party credential raised earnings, by about eight hundred and eighteen dollars a quarter in the published version of the study, as reported by Washington Monthly. And four randomized evaluations of sector-focused programs, which screen applicants and add support services, found earnings gains of fourteen to thirty-nine percent in the year or so after training.

Hiring research agrees. In revised estimates by Paul Sackett and colleagues, structured interviews predict job performance best, with a mean validity of point four two. The top predictors are specific to the job.

And skill information helps when a stranger can check it. In South Africa, assessment results that jobseekers could credibly share with firms raised employment and earnings. Results they could not easily share had smaller effects.

Then there is AI. On the freelance platform Freelancer, employers used to pay more for workers whose applications were tailored to the job. After large language models arrived, they did not. That is one platform, but it suggests that polished self-descriptions may now say less about effort than they did.

The tension 6:05

So is the degree obsolete, and employers just slow? The evidence does not say that.

Hiring records are scarce. The Burning Glass and Harvard study is observational, covers large firms, relies on online career histories, and ends in 2023. Its authors say the results are sensitive to the mix of occupations.

And the degree still pays. Median earnings for full-time workers aged twenty-two to twenty-seven with only a bachelor's degree were sixty thousand dollars in 2025, against forty thousand with only a high school diploma.

The AI evidence is unsettled. A Stanford payroll study finds that twenty-two to twenty-five year olds in AI-exposed occupations hold nineteen percent fewer jobs than if they had tracked less-exposed peers. But the authors call that descriptive, not causal, and some of the divergence predates generative AI. Robert Fairlie and Jane Wu found no spike in recent graduates' unemployment in the summer of 2026, compared with earlier summers.

One field experiment points the other way on writing. Among nearly half a million jobseekers, those given algorithmic help with their résumés were hired eight percent more often, with no evidence that employers were less satisfied.

Most credential returns are non-experimental, so they cannot rule out that motivated people both earn credentials and get raises. The randomized evidence covers intensive programs, not the typical badge. Company figures need care. Coursera's report says ninety-four percent of employers surveyed would pay graduates with micro-credentials a higher starting salary. That is a stated intention, not a hiring record. Coursera's chief content officer, Marni Baker Stein, told Times Higher Education, quote, there's not a great taxonomy of trust.

Even the best hiring tool is modest: structured interviews average point four two, and results vary widely between settings.

The implication 8:33

If removing the degree line does not change hiring, the lever, on our reading, is the evidence behind the alternative. Read together, the research suggests skill information changes hiring when it is specific to the job, produced by someone other than the candidate, and in a form a stranger can check.

That puts weight on observed work. Graduates with at least one internship had forty-eight point five percent lower odds of underemployment. The economist Amanda Pallais argues that firms hire too few inexperienced workers, because the information a first job reveals benefits the worker, who cannot pay the firm for it.

Public money is starting to ask for proof, but it is early days. Carrie Warick-Smith, who oversees federal policy at the Association of Community College Trustees, was quoted by Washington Monthly: July 1 is not a floodgate. It is a start point of the marathon.

The WAC view 9:41

In our white paper, we propose a Proof-of-Skill Ladder. It sorts skill claims by two questions a hiring manager rarely asks aloud. Who says so? And can I check? At the bottom is claimed: a skill listed on a résumé. Then attended: a completion certificate or badge. Then assessed: an exam graded by the provider. Then verified: a supervised test by an independent body. And at the top, demonstrated: observed work, such as a placement, an apprenticeship or a first job.

A credential climbs the ladder by passing four tests. Is it assessed under supervision against a stated standard? Is it comparable, so another assessor would reach the same result? Is it verifiable by an employer in a minute? And is it valued, with employment and earnings published against a comparison group?

And technology does not move a credential up a rung. A cryptographically signed record of attendance is still rung two.

For universities, that means sorting every award by rung, building at least one assessed placement or employer-set project into each program, and publishing outcomes for each short credential. For educators, it means shifting weight from work AI can produce to supervised performance. For policymakers, it means tracking who is hired, not only what postings say.

One honest caveat. Better evidence may not be fairer. Work sample tests show the second-largest Black-White score gap among the tools Sackett and colleagues compare. And observed work depends on someone offering a first chance. If it is rationed by connections, the ladder will reproduce the advantage it is meant to replace.

What to watch 11:48

One. Whether Workforce Pell approvals grow beyond the first twenty-one programs, and whether providers game the thresholds through selective enrollment or temporary jobs, as one commenter on the rule warned.

Two. Whether verifiable digital credentials change who gets hired. Our authors found no hiring-record study that tests this.

Three. Whether the entry-level shortfall widens for later graduating classes, as Fairlie and Wu warn it might.

The sources for every figure in this episode are listed on the episode page, along with the full white paper, The Proof-of-Skill Gap, at worldassessmentcouncil.org.

WAC Signal is produced by the World Assessment Council, with AI-assisted research and AI voices, and every claim is checked against its source before publication.

This has been WAC Signal. Evidence and ideas for the changing world of learning.