← The New Commonwealth

Paper 06: Jobs Without the Degree Tax

Signed 1 October 2026.

Plank: No one should have to buy a four-year degree to prove they can do a job that has nothing to do with one.

1. What we want

No American should have to buy a four-year degree to prove they can do a job a degree has nothing to do with. We want the federal government to stop requiring degrees for its own jobs and to give employers a lawful way to hire on tested skill instead of a diploma. We want states to stop using a license as a general-purpose barrier to work and to recognize a license earned in one state when a worker moves to another. We want public money that now flows automatically to colleges through their endowments to flow instead, in part, to the student directly. None of this bans the degree. It removes the legal and regulatory reasons employers reach for one when they don't need to.

2. The evidence

What actually predicts whether someone can do the job

The most cited study in personnel selection is Frank Schmidt and John Hunter's 1998 meta-analysis of 85 years of hiring research, published in Psychological Bulletin. It put general mental ability at roughly .51 correlation with job performance, the highest of any single predictor they measured, and years of education near the bottom of their list, at roughly .10. That paper has been the field's reference point for a quarter century.

It has since been revised, and we lead with the revision rather than let an opponent bring it up first. Sackett, Zhang, Berry and Lievens re-examined the same territory in 2022, in the Journal of Applied Psychology, and argued that the older meta-analyses had overcorrected for a statistical problem called range restriction. Their updated estimates: structured interviews at .42, job knowledge tests at .40, work sample tests at .33, and general cognitive ability at .31. General mental ability is no longer the single best predictor of job performance. Structured interviews are.

What doesn't change between the two studies is where education sits. It was weak in 1998 and nothing in the 2022 revision rehabilitates it. Testing whether someone actually knows the work outperforms asking whether they finished a degree by a wide margin in both versions of the literature. The claim here isn't that tests are perfect. It's that American hiring built its main gate out of one of the worst predictors available and made it close to mandatory.

What employers' own numbers show

In 2017, Joseph Fuller and Manjari Raman of Harvard Business School, working with Accenture and Grads of Life, published "Dismissed by Degrees." They compared employer job postings to the people employers already had doing those jobs. For production supervisor positions, 67 percent of postings required a bachelor's degree. Sixteen percent of the people actually working as production supervisors had one. That is a company's own hiring criteria checked against its own successful employees, and it doesn't hold up. The finding is specific to production supervisors, not supervisory work generally, and it should be cited that way.

Contrary evidence: does dropping the requirement actually change hiring?

It mostly hasn't, and we say so plainly. A 2024 study by the Burning Glass Institute and Harvard Business School's Project on Managing the Future of Work looked at more than 11,000 employers who publicly announced they were dropping degree requirements between 2014 and 2023. The share of non-degree hires rose by only 3.5 percentage points on average. Fewer than one in 700 new hires resulted from the announced changes. Only 37 percent of the employers that announced a change, the group the report labels "Skills-Based Hiring Leaders," actually hired differently afterward; almost all of the movement in the data came from that smaller group. If a state or company removes the degree line from a job posting and keeps screening resumes the same way it always did, nothing improves. That is the strongest evidence against this paper's premise, and it means the ask below has to change more than a posting requirement. It has to change what employers are legally protected for doing when they screen on skill instead.

Sheepskin effects

Wage data on college completion shows a jump at the point of finishing a degree that is larger than the gain from any single year of study before it. If a degree purely built skill, the wage return would rise roughly evenly with each year completed. It doesn't. That completion jump is the cleanest evidence that part of the wage premium is the piece of paper itself, not the learning behind it.

The competing explanation, stated fairly

Labor economists Claudia Goldin and Lawrence Katz, in The Race Between Education and Technology (2008), explain the rising value of a degree without reference to any law at all: technology raised the demand for skilled labor faster than the education system could supply it, and the wage premium is the market clearing price for that gap. Autor and Katz's work on skill-biased technical change tells a similar story. We are not claiming the legal story below is the only cause of degree inflation. We're claiming it's a real and specific contributor, on top of the economic one, and that it's the part a legislature can actually fix. The Fuller and Raman finding above doesn't depend on this argument being right, so nothing is lost by stating the limit honestly.

3. Why it happened

What were employers doing before they all switched to degree screens, and what made them stop?

They were testing. On July 2, 1965, the day Title VII of the Civil Rights Act of 1964 took effect, Duke Power Company added a new requirement for its higher-paying jobs: a high school diploma and passing scores on two aptitude tests. Duke Power had run openly segregated departments before that date. Black employees sued, and in Griggs v. Duke Power Co., 401 U.S. 424 (1971), the Supreme Court ruled unanimously that an employment practice which is neutral on its face but produces a disparate outcome by race is unlawful unless the employer can show it's genuinely necessary for the job. Duke Power's tests failed that standard. The context matters: this was a company that had discriminated openly and then reached for a facially neutral instrument the moment discrimination became illegal, and the ruling fit that case. Our argument is that the ruling has outlived its case. Unvalidated aptitude tests became a legal hazard for any employer, everywhere, for any reason, whether or not that employer had ever discriminated.

The Supreme Court eased that standard eighteen years later, in Wards Cove Packing Co. v. Atonio, 490 U.S. 642 (1989), which shifted the burden of proof back onto workers and lowered the bar employers had to clear. It didn't matter for long. Congress passed the Civil Rights Act of 1991, which restored the tougher Griggs standard for disparate impact claims and, in Section 106, added something new: a flat ban on adjusting employment test scores by race, a practice called race-norming.

Race-norming existed because the federal government had built a real alternative to the diploma and then quietly altered it to keep passing its own disparate-impact test. The Department of Labor's General Aptitude Test Battery ran through state employment offices for decades, giving workers without a degree a validated way to show what they could do. Starting in 1981, the government began reporting GATB scores as a percentile within the test-taker's own racial group rather than as a raw score, specifically to keep the pass rates across groups from diverging. When that became public in the late 1980s, it set off a controversy that the 1991 law resolved by banning the practice outright. The Department of Labor suspended the GATB in the early 1990s, and no comparable federal instrument replaced it.

In 1971, using an unadjusted aptitude test became a legal risk for any employer. In 1991, the adjustment employers had been using to manage that risk became illegal, and the government's own validated alternative was taken off the table in the same stroke. The changes, twenty years apart, left the diploma as the legally uncomplicated option, already familiar to HR departments and requiring no employer to defend a validation study to anyone. Nobody designed a system that would load $1.7 trillion in federal student debt onto entry-level hiring. That system grew from a liability rule that made every other option riskier than the diploma, which was already available for employers to use.

One piece of the old infrastructure still runs today. The military's ASVAB is administered to more than a million people a year through the Student Testing Program and at entrance processing stations, which makes it the largest validated aptitude-testing operation in the country by a wide margin. The capability this paper is asking the federal government to extend to civilian hiring already exists and already runs at scale. Custom, not any technical limit, is what keeps it restricted to military purposes.

4. What to change

Federal. Create a statutory safe harbor under Title VII for employers who hire using aptitude and job-knowledge instruments validated against a published federal standard. The federal government's role should be the standard and the legal shelter, not the test itself: multiple private testing companies should be able to certify their instruments against that standard and compete on accuracy and cost. A single government-run test would hand whoever administers it, fifteen years from now, effective control of hiring in America, and that is a worse outcome than the status quo.

Second, the federal government should stop requiring a degree for its own jobs, the same way several states have already stopped requiring one for theirs. This is entirely within the executive's control for the classified federal workforce and doesn't need to wait on the safe harbor to take effect.

Third, require federal contractors to use validated aptitude and job-knowledge instruments in place of degree screens, through a new executive order on contractor hiring screens. Executive Order 11246 shows both halves of what that lever can and can't do: signed in 1965, it spread a new hiring standard through the private economy by tying it to federal contract dollars, and by the Department of Labor's own estimate it once reached federal contractors employing roughly a fifth of the American workforce. It was also revoked by a new executive order in January 2025, and contractors wound down compliance within months [21]. An order can redirect contractor hiring practice fast, and an order lasts exactly as long as the administration that issued it, no longer, which is the argument for making the safe harbor above a statute rather than stopping at this executive order.

Fourth, expand the federal tax on large university endowments and direct the new revenue to students rather than the general treasury. The 2025 tax law already moved federal policy in this direction: it replaced the old flat 1.4 percent excise tax on private university endowment income with a tiered rate that reaches 8 percent for the wealthiest schools, starting in 2026. That revenue currently goes to the general fund. Our ask is to route a defined share of it into a direct payment to the student, structured so it reduces what a student borrows rather than what a university spends.

State. Remove degree requirements from state employment and promotion, for positions where the requirement doesn't test anything the job actually needs. More than a dozen states have already done some version of this by executive order since 2022, starting with Maryland; the case here is for doing it by statute, since an order lasts one administration and a statute doesn't. This is the most immediately fileable piece of the whole paper, and Section 5 below describes the bill.

Also at the state level: roll back occupational licensing to occupations where a real public safety case exists, and make the licenses that remain portable, so a license earned in one state is recognized in another without the worker repeating training and fees they've already completed. Arizona did this in 2019 with a universal recognition law: a worker licensed for at least a year in any other state, with no pending complaint or disqualifying record, gets an Arizona license without redoing the coursework. More than twenty states have passed some version of it since.

Local. Fund apprenticeship and the trades on the same footing as the college-preparatory track, so a school district's spending doesn't assume every graduate is headed to a four-year degree.

5. What a bill would do

The most immediately fileable piece is the state employment bill, which strips degree requirements out of state hiring and promotion. It finds that educational attainment is a weak predictor of job performance relative to demonstrated skill, and that degree requirements screen out qualified residents without improving the state workforce. It then prohibits any state agency from imposing a degree requirement on a position in the classified service unless the agency files a written determination that the degree is required by law, required by an accrediting body, or that no validated alternative assessment exists for the job, and publishes that determination. In place of the degree, agencies have to accept relevant work experience, apprenticeship completion, military service, or a validated skills assessment for any qualification a degree used to satisfy. It changes who may apply for a job, not the pay or protections that come with holding it, and the bill says so directly, which is what keeps public-sector unions neutral on it.

The state's personnel department has to publish, every year, how many positions carry a necessity determination and what share of new hires hold no postsecondary degree. Without that public count, an agency that dislikes the law can file a one-line necessity determination for every position it wants to keep gatekeeping. That lets it satisfy the law on paper while the reform disappears inside a couple of budget cycles, without anyone having technically broken the statute. This bill belongs with the committee that handles civil service, not education or labor, because it is a hiring-practices bill. It moves slower the moment it gets treated as one about schools.

A second, related state bill covers occupational license portability, on the Arizona model. It would direct every licensing board to issue a license to an applicant who already holds an equivalent license from another state, has held it for a minimum period, has no pending discipline and no disqualifying record, without repeating coursework, apprenticeship hours or exams already completed elsewhere. It would require the board to act within a fixed number of days or the license is deemed granted, and it carries the same kind of reporting clause: each board publishes annually how many reciprocity applications it received, approved and denied, and why, for the same reason the employment bill's reporting clause matters: a board that dislikes the law can approve every application in form while finding a paperwork reason to deny most of them in practice.

The federal safe harbor is a longer bill and a harder fight, because it has to specify what "validated" means, who certifies a testing instrument against that standard, and how a certified instrument earns Title VII protection. That's model-bill work for a separate document; the load-bearing clause is the same idea in a different place, that the federal government sets and enforces the validation standard but does not itself build or own the test.

6. What would prove us wrong

We'll publish the answers to these whether they help us or not.

States that adopt universal license recognition show no measurable increase in the number of licensed workers crossing state lines within five years. If that happens, portability wasn't the binding constraint on labor mobility and the case for this ask is weaker than stated.

Employers given a validated safe harbor keep screening on degrees anyway. The 2024 Burning Glass and Harvard finding, that most employers who dropped a degree requirement didn't actually change who they hired, is the closest evidence we have that this could happen, and it's the outcome we think is most likely to bite. If a legal shelter for skills testing produces the same non-result that removing the posting requirement did, then the diploma screen was never really about legal exposure, and the argument in Section 3 is wrong.

A replication of Fuller and Raman's employer data finds that degree requirements do track job performance after all, once measured against a wider set of occupations than production supervisors. If that holds, Section 2's independent case collapses along with it.

The sheepskin effect disappears under better wage measurement, and the return to education turns out to be roughly linear per year of study rather than concentrated at completion. That would hand the whole story back to Goldin and Katz's supply-and-demand account and take the signaling argument off the table.

Sources

  1. Schmidt, F. L., & Hunter, J. E. (1998). "The Validity and Utility of Selection Methods in Personnel Psychology: Practical and Theoretical Implications of 85 Years of Research Findings." Psychological Bulletin, 124(2), 262–274. https://www.researchgate.net/publication/232564809_The_Validity_and_Utility_of_Selection_Methods_in_Personnel_Psychology_Practical_and_Theoretical_Implications_of_85_Years_of_Research_Findings

  2. Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2022). "Revisiting Meta-Analytic Estimates of Validity in Personnel Selection: Addressing Systematic Overcorrection for Restriction of Range." Journal of Applied Psychology, 107(11), 2040–2068. https://pubmed.ncbi.nlm.nih.gov/34968080/

  3. Society for Industrial and Organizational Psychology. "Is Cognitive Ability the Best Predictor of Job Performance? New Research Says It's Time to Think Again." (Summary of Sackett et al. 2022 validity figures.) https://www.siop.org/tip-article/is-cognitive-ability-the-best-predictor-of-job-performance-new-research-says-its-time-to-think-again/

  4. Fuller, J. B., & Raman, M. (2017). "Dismissed by Degrees: How Degree Inflation Is Undermining U.S. Competitiveness and Hurting America's Middle Class." Harvard Business School, Accenture, and Grads of Life. https://www.hbs.edu/faculty/Pages/item.aspx?num=53502

  5. Burning Glass Institute & Harvard Business School Project on Managing the Future of Work (2024). "Skills-Based Hiring: The Long Road from Pronouncements to Practice." https://www.burningglassinstitute.org/research/skills-based-hiring-2024

  6. Goldin, C., & Katz, L. F. (2008). The Race Between Education and Technology. Harvard University Press. (Publisher record.) https://www.hup.harvard.edu/books/9780674035300

  7. Griggs v. Duke Power Co., 401 U.S. 424 (1971). https://supreme.justia.com/cases/federal/us/401/424/

  8. Wards Cove Packing Co. v. Atonio, 490 U.S. 642 (1989). https://en.wikipedia.org/wiki/Wards_Cove_Packing_Co._v._Atonio

  9. U.S. Equal Employment Opportunity Commission. "Civil Rights Act of 1991" (Section 106, prohibition against discriminatory use of test scores). https://www.eeoc.gov/civil-rights-act-1991-original-text

  10. "General Aptitude Test Battery." Wikipedia, summarizing the race-norming controversy and the Department of Labor's 1990s suspension of the GATB. https://en.wikipedia.org/wiki/General_Aptitude_Test_Battery

  11. U.S. Equal Employment Opportunity Commission et al. "Uniform Guidelines on Employee Selection Procedures" (1978), 29 C.F.R. Part 1607, the four-fifths rule. https://www.ecfr.gov/current/title-29/subtitle-B/chapter-XIV/part-1607

  12. OfficialASVAB.com. "About the ASVAB" (more than one million test-takers annually through the Student Testing Program and Military Entrance Processing Stations). https://www.officialasvab.com/

  13. "Executive Order 11246." Wikipedia (signed by President Johnson, September 24, 1965). https://en.wikipedia.org/wiki/Executive_Order_11246

  14. U.S. Bureau of Labor Statistics, Current Population Survey. "Data on Certifications and Licenses." 2024–2025 annual averages: 21.6 percent of workers hold a license, 24.0 percent hold a license or certification. https://www.bls.gov/cps/certifications-and-licenses.htm

  15. Kleiner, M. M., & Vorotnikov, E. (2018). "At What Cost? State and National Estimates of the Economic Costs of Occupational Licensing." Institute for Justice. (Licensing share roughly 5 percent of the workforce in the early 1950s, rising to roughly 20 percent by the report's estimate.) https://ij.org/report/at-what-cost/

  16. Harvard University Financial Administration. "FAQs: How the 2025 Net Investment Income Tax ('Endowment Tax') Rate Increase Affects Harvard University." Explains the tiered 1.4/4/8 percent rate structure effective for tax years beginning 2026 under the 2025 reconciliation act. https://finance.harvard.edu/endowment-tax-faqs

  17. Cronkite News. "AZ recognizes out-of-state occupational licenses." April 10, 2019, on Arizona House Bill 2569, the first state universal occupational-license recognition law. https://cronkitenews.azpbs.org/2019/04/10/license-bill/

  18. Education Data Initiative. "Student Loan Debt Statistics." Federal student loan debt approximately $1.7 trillion as of early 2026. https://educationdata.org/student-loan-debt-statistics

  19. Center for American Progress. "The Nondiscrimination Protections of Millions of Workers Are Under Threat" (citing the Office of Federal Contract Compliance Programs' own estimate that Executive Order 11246 covers roughly one-fifth of U.S. civilian employees). https://www.americanprogress.org/article/nondiscrimination-protections-millions-workers-threat/

  20. NPR. "Why states are dropping the 4-year college requirement for some government jobs." August 26, 2024, on the wave of state executive orders since Maryland's in 2022. https://www.npr.org/2024/08/26/nx-s1-5084214/why-states-are-dropping-the-4-year-college-requirement-for-some-government-jobs

  21. Federal Register, "Rescission of Executive Order 11246 Implementing Regulations," July 1, 2025 (EO 11246 revoked by EO 14173, January 21, 2025; contractors directed to wind down compliance by April 21, 2025). https://www.federalregister.gov/documents/2025/07/01/2025-12276/rescission-of-executive-order-11246-implementing-regulations