M04 — When a program crosses from passing to failing
+
+ How the rule reads single-year results, what 'two failures in three
+ years' actually means, and why every measurement looks back four
+ tax years.
+
+
+ ●
+ Public data only · No private or institutional data leaves your browser
+
+
+
+
+
+
+
+
The short version
+
+ The earnings-premium test produces a single-year pass-or-fail
+ arithmetic for each measured program. One single-year failure
+ attaches a disclosure obligation but no Title IV consequence. Two
+ single-year failures within any three consecutive measurement
+ years take the program out of Direct Loan eligibility for new
+ students.
+
+
+ Each measurement year reads earnings four tax years after the
+ measured graduating cohort. That four-year lag is built into the
+ statute itself. The first measurement year a graduating class
+ ever appears in is therefore four years after that class
+ graduates — and a program does not feel any pass-or-fail signal
+ until that lag has already passed.
+
+
+
+
+
The single-year pass/fail arithmetic
+
+ For each award year a program is measured in, the rule computes
+ two numbers — the program's median earnings (graduates who are
+ working and not still in school) and a benchmark figure drawn
+ from federal Census data — and asks whether the median is
+ greater than or equal to the benchmark.
+
+
+ Pass
+ The program's median earnings are greater than or equal to the
+ benchmark figure. The program is reported PASS for that
+ measurement year. No federal consequence attaches; the program
+ remains fully Title IV eligible.
+
+
+ Fail
+ The program's median earnings are less than the benchmark
+ figure. The program is reported FAIL for that measurement year.
+ Disclosure obligations attach (M07). Title IV eligibility is not
+ yet affected by the single failure.
+
+
+ Not measured
+ The program does not have enough graduates, or enough IRS
+ earnings records, to produce a published median. The program is
+ reported NOT MEASURED for that year. NOT MEASURED is not a pass
+ and not a fail; no consequence attaches.
+
+
+ The single-year arithmetic is straightforward. The state machine
+ on top of it — the rule that decides what consequence attaches
+ and when — is what makes the test stateful across years.
+
+
+
+
+
The two-failures-in-three-years rule
+
+ OBBBA § 84001(c)(1) sets the program-level Title IV consequence
+ on a two-failure trigger inside a rolling three-year window. The
+ window slides forward each year. A program that fails in year
+ one and again in year three loses Direct Loan eligibility on the
+ year-three release; a program that fails in year one and again
+ in year four does not lose eligibility on the year-four release,
+ because year one has rolled out of the three-year window by then.
+
+
+ The trigger reads only across measurement years. A NOT
+ MEASURED year does not count as a pass and does not count as a
+ fail. It is a year without a result. A program with measurement
+ gaps can take longer to accumulate two failures inside any
+ three consecutive measurement years, but the gap does not erase
+ earlier failures — they remain on the running tally until they
+ age out of the three-year window.
+
+
+ The federal data product the Department published in advance of
+ the rulemaking carries pre-computed pass/fail flags for each
+ measurement year. The tool reads the published flag as
+ authoritative and surfaces local re-computation alongside it as
+ a cross-check (M01).
+
+
+
+
+
The four-year earnings lag
+
+ The statute defines program earnings as median earnings four tax
+ years after the measured graduating cohort completed. That lag
+ is structural: the federal data the Department uses is built
+ from IRS records (W-2 and Schedule SE filings), and four tax
+ years is the lag that produces stable post-completion earnings
+ for graduates who entered the workforce after their degree.
+
+
+ The lag has two practical consequences. First, the first
+ measurement year a graduating class ever appears in is the
+ fourth tax year after that class graduates. A class that
+ completed in academic year 2023-24 first appears in measurement
+ year 2027-28 (the federal data release that reports tax-year
+ 2027 earnings).
+
+
+ Second, any institutional response to a failure — a curriculum
+ change, a program closure, a new program, a CIP-code shift —
+ takes at least four years from the entering cohort onward to
+ register on the test. A program that admits a new entering
+ cohort under a redesigned curriculum will not see that cohort's
+ earnings appear in the federal data for at least six years
+ after admission (two-to-three years to graduate plus four tax
+ years post-completion).
+
+
+
+
+
What this means in practice
+
+ One failure attaches a disclosure obligation that the program
+ must publish on admissions and recruitment materials before the
+ next entering cohort accepts admission (M07). The disclosure
+ obligation is independent of any appeal. Filing an appeal under
+ proposed § 668.603 does not pause the disclosure requirement;
+ the failing program must surface the federal-warning language
+ while the appeal is pending.
+
+
+ Two failures inside any three consecutive measurement years take
+ the program out of Direct Loan eligibility for new students.
+ Existing borrowers are unaffected for prior disbursements (M05).
+ A program restored to compliance after eligibility loss must
+ clear the published procedure under proposed § 668.604; the
+ institution may not restart enrollment without an affirmative
+ determination from the Department.
+
+
+ The proposed appeal scope under § 668.603 is limited to
+ calculation-error claims — arithmetic disputes about the
+ single-year pass/fail computation. Methodological disagreement,
+ IRS-match-quality concerns, and disputes about the cohort-side
+ expansion outcome are outside the appeal scope.
+
+ What the institution must publish on admissions materials when a
+ program fails the earnings test for the first time, when the
+ obligation attaches, and how it persists through any appeal.
+
+
+ ●
+ Public data only · No private or institutional data leaves your browser
+
+
+
+
+
+
+
+
The short version
+
+ A program's first failure on the earnings test produces an
+ immediate disclosure obligation. The institution must publish
+ a federal-warning notice on admissions materials, recruitment
+ communications, and student-facing surfaces for the failing
+ program — before the next entering cohort accepts admission.
+ The disclosure obligation is separate from, and earlier than,
+ the Title IV eligibility-loss consequence (M04, M05).
+
+
+ Proposed § 668.43(d)(1) is the regulatory authority for the
+ first-failure disclosure. The 2023 Gainful Employment final
+ rule's disclosure procedure is the operative procedural model
+ the Department points to for compliance review — the same
+ posting locations, same Department-prescribed warning
+ language, same audit trail.
+
+
+
+
+
The trigger — first-failure release date
+
+ The disclosure obligation attaches on the first-failure
+ release date: the day the federal data file is publicly
+ released showing the program failed the earnings test for that
+ measurement year. The disclosure does not wait for an institution
+ response, an appeal filing, or a Department determination. It
+ attaches at federal-data publication.
+
+
+ Once attached, the obligation runs forward continuously until
+ one of two things happens. Either the program clears the test in
+ a subsequent measurement year and stays clear (which lifts the
+ disclosure when the rolling window of failures has emptied),
+ or the program registers a second failure inside the
+ three-year window and the obligation transitions to the
+ eligibility-loss disclosure under proposed § 668.604 (M04).
+
+
+
+
+
The form — Department-prescribed warning
+
+ The Department prescribes the warning text. The institution is
+ not free to author its own language; the regulation calls for a
+ specific federal-warning form. Proposed § 668.43(d)(1) sets the
+ form for first-failure disclosure, and proposed § 668.43(d)(2)
+ sets a different form for the disclosure during eligibility
+ loss. Both forms are Department-prescribed.
+
+
+ The 2023 Gainful Employment final rule's disclosure procedure is
+ the operative procedural model. That procedure governs:
+
+
+ Posting locations
+ The federal warning must appear on admissions materials,
+ recruitment communications, the program's web page, and any
+ student-facing surface that markets the program. The 2023 GE
+ rule's enumeration of covered surfaces is the operative list.
+
+
+ Prominence
+ The warning must be displayed in a manner the Department
+ considers prominent — not buried in fine print, not gated
+ behind a click-through. The 2023 GE rule's prominence
+ requirements carry forward.
+
+
+ Audit trail
+ The institution must be able to demonstrate compliance: dated
+ posting evidence, archived versions of admissions materials,
+ and a record of when the warning first appeared on each
+ covered surface.
+
+
+
+
+
The timing — before the next entering cohort
+
+ The disclosure must be in place before the next entering cohort
+ accepts admission. For a program with a fall entering cohort,
+ if the federal data release happens in spring, the warning
+ must be posted before the fall admissions cycle begins. The
+ regulatory aim is that prospective students considering
+ enrollment in the failing program see the federal warning
+ before they commit to enroll.
+
+
+ The institution's enrollment calendar drives the operational
+ deadline. A program with rolling admissions has a continuous
+ obligation to post immediately on the first-failure release
+ date. A program with a single annual entering cohort has the
+ window between the federal release and the start of that
+ cohort's admissions cycle.
+
+
+
+
+
The appeal — disclosure does not pause
+
+ Filing an appeal under proposed § 668.603 does not pause or
+ suspend the disclosure obligation. The failing program must
+ publish the federal warning while the appeal is pending. If the
+ appeal succeeds and the failure is reversed on a calculation-
+ error finding, the disclosure can be removed retrospectively;
+ but during the appeal the warning runs.
+
+
+ The proposed appeal scope under § 668.603 is itself narrow —
+ calculation-error claims only (M04). Methodological objections,
+ disputes about which graduates were counted, and concerns
+ about the IRS-match data quality are outside the appeal scope.
+ A failing program with a methodological objection has the
+ public-comment record and, if the rule is finalized, judicial
+ review — not a § 668.603 appeal.
+
+
+
+
+
What this means in practice
+
+ The first-failure disclosure is the rule's earliest visible
+ consequence. It precedes Title IV eligibility loss by at least
+ one measurement year — and possibly more, since a NOT MEASURED
+ year does not advance the running failure count (M04). For a
+ program that fails once and then has a NOT MEASURED year
+ followed by a passing year, the disclosure obligation persists
+ until the failure ages out of the three-year window, even
+ though no Title IV consequence ever attached.
+
+
+ For institutions, the operational planning question on a first
+ failure is not whether to disclose — that is mandated — but
+ how to coordinate disclosure across admissions offices,
+ recruitment materials, the program web page, and any third-
+ party recruitment partners. The audit trail exists at the
+ institution level; demonstrating compliance is the
+ institution's procedural burden.
+
M14 — Named populations and the rule's disparate impact
+
+ The Department's own analysis identifies CIP families that the
+ rule disproportionately affects. The recomputed elevation
+ factors, against the most recent federal data, run higher than
+ the prose summary.
+
+
+ ●
+ Public data only · No private or institutional data leaves your browser
+
+
+
+
+
+
+
+
The short version
+
+ The Department's regulatory-impact analysis (the published
+ analysis of which programs are most at risk) identifies named
+ CIP families whose programs the proposed rule disproportionately
+ affects. The named families include CIP 50.05 (Drama and
+ Theatre Arts), CIP 50.07 (Fine and Studio Arts), CIP 50.09
+ (Music), and CIP 51.15 (Mental and Social Health Services),
+ among others.
+
+
+ The disparate impact identified in the Department's prose
+ summary is itself substantial. When the same elevation factors
+ are recomputed against the most recent federal data the
+ Department uses, the per-CIP ratios run higher than the prose
+ numbers. The empirical pattern is the policy-level argument
+ that arts and humanities graduate programs face a structurally
+ different test than other graduate programs — surfaced by the
+ Department's own analysis.
+
+
+
+
+
The named CIP families
+
+ The Department's analysis names specific CIP families with
+ elevated failure risk relative to baseline. The named families
+ break out roughly as follows:
+
+
+ CIP 50.05 — Drama and Theatre Arts and Stagecraft
+ Theatre BAs, MFAs, and related performance-arts credentials.
+ Named as elevated-risk in the Department's analysis. The
+ field-of-study benchmark for graduate programs in this family
+ is drawn from working adults with a baccalaureate in arts —
+ a low-paying field — which routes graduate Theatre programs
+ through the lowest-of-three rule (M03) to a low benchmark.
+
+
+ CIP 50.07 — Fine and Studio Arts
+ Studio art, art history, and related visual-arts credentials.
+ Named as elevated-risk. The graduate-level Art MFA and
+ Art-History MA programs sit in this family. The same
+ arts-baccalaureate field-of-study benchmark applies.
+
+
+ CIP 50.09 — Music
+ Music BMs, MMs, and DMA credentials. Named as elevated-risk —
+ and named with the highest elevation factor among the
+ performing-arts CIPs in the Department's prose. The Music MM
+ credential is one of the central exposure points the
+ regulatory-impact analysis highlights.
+
+
+ CIP 51.15 — Mental and Social Health Services
+ Counseling, marriage and family therapy, and related licensed
+ mental-health credentials. Named as elevated-risk. This family
+ is structurally different from the performing-arts CIPs: the
+ benchmark routing reflects working adults in the
+ health-and-counseling field, but the cohort-side risk is
+ driven by the wage profile of licensed therapists in
+ early-career years.
+
+
+ The named-population list is the Department's own. It is not
+ an institutional advocacy framing or a third-party analysis;
+ it is the empirical record the Department itself published in
+ the rulemaking docket.
+
+
+
+
+
The recomputed elevation factor
+
+ The Department's prose summary in the regulatory-impact
+ analysis assigns each named CIP family an elevation factor —
+ a ratio describing how much more likely a program in that
+ family is to fail relative to a baseline reference.
+
+
+ When the same elevation factor is recomputed program-by-program
+ against the most recent federal data the Department uses, the
+ per-CIP ratios run higher than the prose summary. The
+ recomputation uses the same underlying federal data file the
+ rule reads from. The methodology is replicable; the recomputed
+ factors are reproducible from public data.
+
+
+ The recomputation does not change the rule's pass/fail
+ arithmetic for any individual program. A program that passes
+ its own measured year continues to pass; a program that fails
+ continues to fail. The recomputed elevation factor is a
+ summary statistic about the named CIP family as a whole, not
+ a per-program adjustment.
+
+
+
+
+
What 'disparate impact' means here
+
+ 'Disparate impact,' as a phrase, has a specific meaning in
+ civil-rights law that this panel does not use. The empirical
+ pattern surfaced here — that the rule disproportionately
+ flags programs in a small set of named CIP families — is a
+ descriptive observation about the rule's published effect, not
+ a legal claim about a protected class.
+
+
+ That said, the descriptive observation is itself
+ policy-relevant. The rule's lowest-of-three benchmark routing
+ (M03) and four-year earnings lag (M04) operate identically
+ across all graduate programs. The reason arts and humanities
+ graduate programs disproportionately fail is structural: the
+ field-of-study benchmark for those programs is drawn from
+ arts-baccalaureate earnings, which are themselves at the low
+ end of the baccalaureate earnings distribution. The rule
+ measures the master's program against the bachelor's-degree
+ earnings in the same field — and for fields where
+ bachelor's-level earnings cluster near the high-school
+ benchmark, the master's-level program has a small margin to
+ clear.
+
+
+ The named-population finding is descriptive, not procedural.
+ It does not create an appeal pathway, a procedural defense,
+ or a stay of the disclosure or eligibility-loss consequences.
+ M09 covers what is and is not appealable. The named-population
+ status is part of the rulemaking record — relevant to public
+ comment, to legislative oversight, and to institutional
+ strategic planning, but not to the per-program pass/fail
+ machinery.
+
+
+
+
+
What this means in practice
+
+ For a dean with named-population programs, the operational
+ read is layered. The per-program verdict on each named
+ program runs through the same machinery as any other program.
+ But the broader pattern — that the rule's structural choices
+ route arts, humanities, and licensed-counseling graduate
+ programs to disproportionately low benchmarks — is itself a
+ documented empirical record. Public-comment letters,
+ institutional advocacy positions, and post-finalization
+ litigation strategies can rest on the Department's own
+ analysis as their factual baseline.
+
+
+ The snapshot-metadata footer on each verdict card reports the
+ federal-data release date and the recomputation date so the
+ dean can see whether the elevation factor reflects the most
+ recent published data. The recomputed factor is itself a
+ public-data artifact: it can be reproduced from the federal
+ data file directly, and disagreement with the recomputation
+ is itself a methodological argument the rulemaking record can
+ take.
+