The STATS earnings accountability rule is currently being framed as a 2027 compliance issue.

The rule is moving quickly. It was published in April, the comment deadline is May 20, and the framework will be finalized by July. Most institutions are treating this as a downstream compliance exercise tied to 2027 outcomes.

The more immediate issue is how those outcomes will be calculated. Cohort selection, early earnings, and state benchmarks will determine results well before anything is published. Under the current timeline, the first calculation will draw on IRS earnings data four years after completion, pointing back to the COVID cohort: 2021 graduates.

If that holds, the baseline reflects a disrupted labor market. This is the part of the rule that is still open to influence. After May 20, it is not.

This deep dive covers:

Based on the rule’s current timeline and data dependencies, the first STATS calculation is highly likely to anchor on 2021 completers, with earnings measured in 2025 and published in 2027. This is not a modeling assumption, but a function of how IRS earnings data lags program completion by four years.

That choice matters more than any other element of the rule.

The framework assumes that earnings four years after completion provide a clean signal of program value. For the 2021 cohort, that signal is structurally distorted. These graduates entered the labor market during the most disruptive period in recent history, with hiring freezes, sector shutdowns, and compressed entry-level opportunities across multiple fields. The result is well documented: graduates entering recessionary labor markets experience measurable early-career earnings losses that can persist for years.

That distortion does not fully resolve by the fourth year. In many cases, the fourth-year earnings snapshot captures the tail end of that scarring effect, not a normalized trajectory.

There is a second, less visible effect embedded in the benchmark itself.

The rule compares program earnings against the median earnings of high school graduates. Post-pandemic wage dynamics pushed up wages at the lower end of the labor market faster than at the median. In practical terms, programs may be evaluated against a benchmark that was temporarily elevated, while their own graduates’ earnings were temporarily suppressed.

The result is a compression from both sides:

The first set of STATS outcomes will not simply reflect program quality, but a specific economic moment.

The implication is straightforward. If 2021 remains the baseline, the initial classification of programs as passing or failing will be shaped as much by cohort timing and labor market conditions as by institutional performance.

Exposure is not evenly distributed, and most institutions cannot yet identify where it sits in their own portfolio.

ED’s own preliminary analysis shows that risk concentrates in specific program types and credential levels, most visibly in short-duration and lower-wage fields.

Under the proposed framework, roughly 29 percent of undergraduate certificate programs would fail the earnings test, compared to low single-digit failure rates for most bachelor’s and master’s programs.

At the field level, the pattern is consistent:

These are not outliers. They are programs where early-career wages are structurally low, either because of public-sector pay schedules, nonprofit employment, or career paths that ramp over time.

The issue is not underperformance, but that the metric captures wages and not the full set of compensation elements that shape career value over time.

The earnings benchmark is set at the state level. That means identical programs can face materially different thresholds depending on where they are located. Within-state variation compounds this further. Programs serving rural or lower-wage regions are benchmarked against statewide medians that reflect urban labor markets their graduates may never access.

In practice, this creates scenarios where a program passes comfortably in one state and fails in another, with no difference in quality.

The data required to model STATS exposure does not exist in a usable form internally. Institutions would need to connect:

Even ED has acknowledged it cannot fully model the outcome:

That leaves institutions in an unusual position. They are expected to make portfolio decisions, such as program redesign, pricing, or potential closure, based on a metric they cannot fully replicate or test in advance.

For most leadership teams, the question is whether the institution can identify programs at risk before the rule does.

The structure of the rule is largely set and unlikely to change at this stage. What remains open are a small number of technical inputs that will determine how that structure operates in practice.

The current timeline uses 2021 completers as the baseline, but that choice is not set in stone. How ED defines or aggregates cohorts will directly determine how much of the COVID-era distortion carries into the first set of outcomes. ED has discretion over how it defines or aggregates cohorts, particularly when single-year cohorts yield unstable or unrepresentative results. A shift to multi-year averaging, or an adjusted initial cohort window, would materially change early outcomes.

The use of state-level medians introduces structural distortions, particularly for institutions serving rural or lower-wage regions. ED has already acknowledged concerns about geographic variation and data limitations in the American Community Survey, and is explicitly soliciting input on alternative approaches.

Institutions are expected to produce program-level cost and outcomes data on a compressed timeline that many are not operationally prepared to meet. Implementation timing will determine how quickly institutions are forced into program-level decisions they are not yet prepared to make.

What is unlikely to change is the underlying premise of the rule: early-career earnings will serve as the primary signal of program value, and failure against that signal will carry eligibility consequences. Institutional negotiators attempted to push on several aspects of the earnings test, including geographic fairness and the scope of appeals, with limited success.

For institutions, this means the practical question is no longer whether the rule will change, but how to operate within it as currently designed.

This is the distinction that matters. The rule itself is not being reconsidered. The question is whether the inputs used to operationalize it reflect the realities of institutional programs and labor markets, or whether they introduce distortions that will be difficult to unwind later.

After May 20, those inputs move from contestable to fixed.

Higher Education Leadership Intelligence is for presidents, provosts, CIOs, and institutional decision-makers leading through enrollment, funding, and tech disruption.

This is one of our six education and learning-related publications spanning K-12, Higher Education, and Workforce. Our education newsletters reach tens of thousands of senior decision-makers across the U.S. and key international markets.

Ping us at [email protected] if you’d like to learn more, explore Enterprise Subscriptions, or would like to partner in other ways.

The Intelligence Council is a next-gen B2B media and business intelligence platform built for people who make strategy, allocate capital, and carry operating risk.