Gainful Employment and the Earnings Test: What US Colleges Face in 2026

For more than a decade, a provost at a public or nonprofit university could file gainful employment under "for-profit colleges and certificates" and leave it to financial aid. That filing system stopped working on July 1, 2026, when the Department of Education published its final earnings accountability rule. Almost every kind of program that takes federal Direct Loans, from an associate degree in general studies to a master's in counseling, now sits under one pass-or-fail earnings test.

The test is easy to describe and hard to manage. In the fourth tax year after completion, do a program's graduates out-earn working 25 to 34 year olds whose highest credential is a high school diploma (a bachelor's, for graduate programs)? Fail in two out of three years and the program loses Direct Loan access. First results arrive in 2027, the first loan losses in July 2028, and the career center can still move the number.

Related reading: US Community College Employer Partnerships in 2026: Funding and Workforce Pell · Career Services Staffing Ratios in 2026: 1,381 Talents per Professional

From gainful employment to an earnings test for almost everyone

Start with 2023: it's still the reporting machinery your financial aid team lives with this fall. The Financial Value Transparency and Gainful Employment rule, published in the Federal Register on October 10, 2023, took effect on July 1, 2024. It calculated a debt-to-earnings rate and an earnings premium for most programs but only penalized GE programs: certificates, plus nearly everything at for-profit schools. Degree programs at publics and nonprofits got disclosure, not consequences.

Then Congress acted. Section 84001 of the One Big Beautiful Bill Act, signed on July 4, 2025 as Public Law 119-21, amended section 454 of the Higher Education Act to tie Direct Loan eligibility to an earnings test for undergraduate degrees, graduate and professional degrees, and graduate certificates. (The Department's 2026 documents now call the law the Working Families Tax Cuts Act. Same statute, new name.)

The Department finalized its implementing rule on Monday, June 29, 2026 and published it on July 1 as 91 FR 40136. It went further than the statute. The debt-to-earnings metric is gone, a revised earnings premium stays, and the same loan consequence applies to GE and non-GE programs alike, so undergraduate certificates, which the law left out, are pulled back in through the Department's gainful employment authority. Inside Higher Ed's June 29, 2026 coverage described the 641-page rule as largely unchanged from the April proposal.

How the earnings premium works

In plain terms:

• The Department takes a program's completers from one award year and gets their median earnings from federal tax data for the fourth tax year after they finished. The first run, in 2027, uses 2025 earnings for people who completed in the 2020-21 award year.
• Undergraduate programs are compared with the median earnings of working adults aged 25 to 34 whose highest credential is a high school diploma and who aren't enrolled, using Census Bureau American Community Survey data.
• Graduate programs are compared with working 25 to 34 year olds who hold only a bachelor's degree, using the lowest of several state and field-of-study benchmarks.
• Your state's benchmark applies if at least half of the institution's enrollment is in-state. Otherwise it's the national one.
• Programs with fewer than 30 completers get pooled, first with up to three earlier award years of the same program, then with sibling programs in the same four-digit CIP code and credential level.

Two details get missed. First, "working" is a low bar. The Department's own analysis treats anyone with positive earnings as working, and when commenters asked for a floor (one suggestion was $15,000), it declined. The graduate picking up 12 hours a week of shift work while studying for a licensure exam is in your median. Second, the July 2026 preamble says the Department "strongly disagrees" with letting colleges appeal using alternative earnings data such as state wage records or graduate surveys. Appeals are limited to errors in the calculation, and you have 30 days to file one.

Fail in two out of any three consecutive years and the program becomes a low-earning outcome program. It loses Direct Loan eligibility for at least two years, and you can't simply relaunch it, because programs at the same credential level that share its four-digit CIP code and an overlapping occupation code are blocked during that period. There's an institution-level trigger too: at least half of your Title IV recipients and half of your Title IV dollars must come from programs that aren't low-earning. Miss that in two of three years and you're on provisional certification, and your low-earning programs lose all Title IV aid, Pell included, unless you stop Direct Loan borrowing in them.

The dates on the calendar

The lag is the whole story. (Dates come from the 2023 and 2026 Federal Register rules, Federal Student Aid's August 2026 guidance and Inside Higher Ed's July 2026 reporting.)

The first penalties in July 2028 rest largely on people who finished in 2020-21 and 2021-22, long before your current team could have helped them. But the cohort you're advising this fall, the 2026-27 completers, will be measured on their 2031 earnings, in a calculation the Department would run around 2033. Career outcomes work is a slow lever. That's exactly why most institutions will pull it too late.

Which programs are exposed?

The regulatory impact analysis in the July 2026 rule counts roughly 61,900 covered programs. That's only about 30% of the programs in its 2026 program performance data (the rest mostly take no federal loans or are too small to measure), but they enroll about 79% of Title IV recipients. It expects about 5.2% of them to fail and puts the first-year count at about 3,300 programs (the April proposal said 6,520). In people, that's 3.0% of Title IV enrollees in failing programs in the first sanction year (from July 2028), rising to 4.3% once the tipped-field delay expires. It also estimates 831,000 people will need formal warnings.

Small percentages, and that's the trap, because failures cluster. Preston Cooper's January 2026 analysis for AEI, built on the Department's preliminary program data, found that nearly 2,000 of America's roughly 5,000 colleges had at least one failing program. Almost all bachelor's programs passed, apart from a handful in the arts and humanities. At the master's level about 4% failed, and mental and social health services was the only large field where most programs fell below the benchmark. Cooper also counted more than $2.7 billion in 2024-25 federal loans to people in failing programs.

A hypothetical plenty of regional publics will recognize: your largest master's program is clinical mental health counseling. Its graduates spend their first years in supervised practice before full licensure, often at community agency pay. In year four, the median sits a few thousand dollars either side of your state's bachelor's benchmark. It could pass one year and fail the next two. The teaching didn't change. Washington now prices the program by its graduates' earnings curve, and the licensure pay bump may land right at, or just after, the measurement year.

What your institution reports, and what it can't fix

You don't submit earnings; the Department gets aggregate medians from federal tax data. You submit everything around them. Under the July 2026 rule (34 CFR 668.406), each program reports its CIP code, credential level, length, accreditation, licensure states and enrollment. For each enrollee you report cost of attendance, tuition and fees, residency, grants and known private loans, with whole-enrollment totals for anyone who completed or withdrew. Reporting is due October 1 each year, and you get 60 days to correct the completer lists the Department sends back.

Plenty of institutions are behind. In an electronic announcement dated August 11, 2026 (GENERAL-26-49), Federal Student Aid said more than 1,900 institutions had not reported, or had under-reported, data from the 2024 and 2025 FVT/GE cycles. They have until January 15, 2027, which the Department calls a final deadline, and the 2026 cycle is due October 1, 2026. For the 2026 cycle only, you can skip data elements the new rule eliminates, and the Department will treat that as early implementation. It also warned of fines, sanctions and administrative capability questions for anyone who misses them.

Then there's the part you can't fix afterwards. Once the Department notifies you that a program could lose loan access, you have 30 days to send a written warning to everyone enrolled in it. Anyone enrolling has to acknowledge the warning before Title IV money is disbursed, and Pell recipients get a statement of their remaining lifetime eligibility. It's far cheaper to move the median before that letter exists.

Why is this a career services problem now?

Because the money moved. Since July 1, 2026, Grad PLUS has been closed to new borrowers, and the new aggregate caps, spelled out in Federal Student Aid's May 2026 loan limits FAQ, are $100,000 for graduate study and $200,000 for professional programs. For many master's programs, Direct Unsubsidized Loans are now the main federal financing left. Lose them and you've lost how most enrollees pay. That's an enrollment problem first and a budget problem a semester later.

Most outcome reporting stops at six months, with a first-destination survey of placement and starting salary. The federal test looks at year four, uses tax data, and won't accept your survey on appeal. So your first-destination survey is now mostly a marketing document: useful for prospective families, close to useless as a compliance instrument. If your outcomes work ends at the six-month mark, you're measuring the wrong year.

What moves a median is the lower half. The graduates who decide whether a program passes are the ones in part-time, off-field or underpaid work at year one and still there at year three. Getting your strongest graduates slightly better offers does nothing for the median. Getting stuck graduates into full-time roles in their field, and helping them change jobs or negotiate in years two and three, can actually shift it. That's alumni-facing career services work, and most centers aren't staffed for it.

The obvious shortcut is closed, too. Commenters asked to exclude completers based on the career path they chose, and the preamble rejects that as an invitation to game the test. Every working completer counts, bar a few narrow exclusions. Tooling helps: platforms such as Talenlio give career teams a readiness dashboard, weekly reports and exportable data across a cohort, so you can see who's drifting before a survey would. Whatever you use, you need a live view of the people who'll be in the median.

A 12-month plan for provosts and career centers

Here's the order we'd work in. If you run the career center, don't wait to be invited.

1. Run a program-level exposure check this term against the Department's 2026 program performance data. Flag every program within roughly 10% of its benchmark, not only the ones already failing.
2. Put October 1, 2026 and January 15, 2027 on the provost's calendar, not only the financial aid director's. Incomplete reporting is now an administrative capability question.
3. Give each flagged program a named owner who can get academic affairs, institutional research, financial aid and career services in one room. Four partial views is how programs fail quietly.
4. Extend outcomes tracking for flagged programs from six months to four years, using alumni check-ins, LinkedIn data and employer partners. It won't help on appeal, but it's your only early warning.
5. Concentrate career services capacity on those programs' final-year cohorts and their first three alumni years: interview practice, employer pipelines in the field, salary negotiation. AI tools can carry some volume; Talenlio's Interview Coach AI and Job Hunter AI Agent (80+ job boards) are one option, with weekly reports for the program owner.
6. Decide your fallback early. A program that has failed but isn't yet low-earning can run an orderly closure for up to three years (or its full-time length, if shorter), or stop Direct Loan borrowing for at least five years to protect institution-level eligibility. Those are board-level calls. Don't make them inside a 30-day appeal window.

By the time the first sanctions land in 2028, the earnings premium will sit next to enrollment and net tuition revenue in every graduate program review, with the career center director in the room. The institutions that come through well will be the ones that put that director there in 2026, while the people who'll be measured are still on campus.

Want to see that live view for one of your cohorts? Book a walkthrough, or start with our university page.