It's Already Too Late

Was this forwarded to you by a friend? Sign up, and get your own copy of the news that matters sent to your inbox every week. Sign up for the On EdTech newsletter. Interested in additional analysis? Upgrade to the On EdTech+ newsletter.

One of the central changes in the One Big Beautiful Bill (OB3) is Do No Harm (institutional accountability) that judges academic programs based on whether graduates produce enough earnings premium. At the program level, failing the accountability metrics in two out of three years has real consequences. It’s not just a public label; it’s a trigger that can remove access to federal student loans at a minimum, and for some workforce-oriented schools even access to Pell Grants. Either way, for a large share of failing programs this is a survival-level issue—not a compliance nuisance.

Under Secretary Nicholas Kent described the goals at the American Council on Education (ACE) summit in late February.

Virtuous cycles and downward pressure in education policy require feedback loops, however, and over the next 6-10 years there will be no opportunity for specific programs to adjust design and behavior based on the Do No Harm earnings metrics. If a program’s results fall short in terms of completer earnings, the institution should have time to respond: improve program quality, tighten alignment to labor market needs, adjust costs, and, yes, improve student success. Over time that improves outcomes and avoids sanctions.

The problem is that the draft regulations implementing the Do No Harm statutory language are implemented retroactively, holding institutions accountable for earnings of student cohorts that were enrolled well before the actual earnings metrics were even defined. And it will take many years before any feedback loop enabling improvement is in place.

The timing mismatch is structural

Institutions do have levers that influence the earnings premium metric. Some are slow and indirect (curriculum changes, employer partnerships, student supports), but one lever is immediate and powerful: admissions. If an institution believes it will be judged based on earnings outcomes, it can change the profile of who it admits. That’s not a moral judgment; it’s simply how incentives work.

Even in the case of open access programs, there is the admissions question of how many students to enroll in order to meet local workforce demands.

But here’s the reality: admissions decisions today won’t show up in Do No Harm earnings data anytime soon. There’s the time to complete the program—one year for a certificate, four to five years for a bachelor’s—and then there’s the earnings observation window (earnings four years post-completion). Add the reporting and processing lag that comes with using federal earnings data (one to two years). You’re quickly in a world where the students who drive a program’s “future” accountability result were admitted 6-10 years before the program ever sees the metric that will judge them.

Keep in mind that the metrics were first defined with the OB3 bill that was signed into law July 2025. But until accountability hits a steady state in 2032 and beyond (at the earliest), there is no feedback loop to allow institutions to make changes and improve program performance before losing access to federal student loans.

Consider a one-year undergraduate certificate for the first year of official reporting and a two-year cohort measurement. And consider the first round of official earnings data that is due February 2027.

• Students entered the program in 2019-20.

• Students completed the program in 2020-21.

• Program earnings are calculated for tax year 2025.

• Do No Harm official metrics are published in February 2027.

• For programs that fail the first two years, they lose access to loans in mid-2028.

The institution will be judged on admissions and enrollment decisions that happened back in 2019 and 2020 when no one knew of the metric until mid 2025. There’s no realistic “fix” available in the near term, because a school can’t go back in time and re-admit a different cohort, and you can’t re-run the economy for those completers.

It will already be too late.

Bachelor’s programs are even longer

Now consider a bachelor’s degree under the same conditions. The pipeline is longer, the cohorts are farther back, and the time between “institution can act” and “metric reflects the action” stretches into a decade.

• Students entered the program in 2016-17.

• Students completed the program in 2020-21.

• Program earnings are calculated for tax year 2025.

• Do No Harm official metrics are published in February 2027.

• For programs that fail the first two years, they lose access to loans in mid 2028.

The institution will be judged on admissions and enrollment decisions dating back to 2016 or earlier, at least nine years before OB3 was passed and 11 years before the first programs lose access to loans.

In the early years, Do No Harm will not function as an improvement incentive. It will function as a delayed penalty system. It will end up failing programs that—by the time they’re identified—may already have changed, improved, shrunk, merged, or been redesigned.

Even a proactive institution can’t outrun the lag

Even if an institution began planning the moment OB3 became a serious possibility and even if leadership tried to change behavior immediately, the first realistic admissions cycle that could fully reflect a deliberate response is Late 2025 for undergraduate certificates and Fall 2026 for bachelor’s degrees.

If proactive executives started admissions changes as soon as they learned of OB3 details, when would that show up in an earnings-based accountability determination?

• For undergraduate certificates, we’re looking at 2031-33.

• For bachelor’s programs, we’re into the mid-2030s.

We can quibble about precise years, but the reality is that any improvement strategy that relies on the OB3 Do No Harm standards will not take effect for 6 - 10 years.

Requests to remove the retroactive nature of the rules

It’s already too late for many programs

Do No Harm is being framed as accountability that will push institutions to improve outcomes. But in the first several years of determinations, the system isn’t rewarding improvements—it’s judging historical cohorts. The people who can “respond” to the signals are being judged on students they admitted long before the signals existed.

That doesn’t mean the policy will have no effect. It will create pressure and shape behavior simply due to implied threats and the nature of bully pulpits. But much of that behavior will be defensive and immediate (restrict access, reduce risk, change program mix), because the longer-horizon “improve outcomes and let the data reflect it” pathway is simply not available on the timeframe the policy operates.

Which leads to the uncomfortable conclusion: for many programs, by the time the first official metrics arrive, it’s already too late.

Unless there are changes made during the final public comment period this spring.

The main On EdTech newsletter is free to share in part or in whole. All we ask is attribution.

Thanks for being a subscriber.