US college enrolment falls 50%

Current status

• Baseline (B): The Clearinghouse's Final Fall Enrollment Trends 2025 (Jan 15, 2026) counts "over 19.4 million" enrollments in fall 2025 (16.2M undergraduate + 3.2M graduate), up 1.0% on the year. Its coverage is 97% of Title IV degree-granting enrollment. A YES needs a coverage-adjusted fall total of about 9.7 million or less. https://nscresearchcenter.org/final-fall-enrollment-trends/
• Latest reading: Final Spring Enrollment Trends 2026 (June 4, 2026) shows 18.6M, up 1.0% (+192k). https://nscresearchcenter.org/enrollment-insights/
• Fall 2026 so far: institutions keep announcing records (UNC System above 260k on Sept 16, 2026; University of Illinois System above 102k; Texas Tech, UTEP, UIC). The Clearinghouse reported that college-going and persistence among recent high school graduates rose (Sept 17, 2026). Hechinger Report (Oct 2026): opinion of college has soured (31% call it "very important"), but enrollment has been "remarkably steady" and has risen since 2022. Enrollment also tends to rise when the job market weakens. Recent-graduate unemployment shows no AI-driven spike yet (UCLA Anderson Review; Washington Post, Sept 28, 2026).
• Calendar: The preliminary report comes out in November and is excluded. Final fall reports come out in January (Jan 2027, 2028, 2029, 2030, 2031), so five fall terms can count (2026–2030). The fall 2031 report would only count if it came out early, in December 2031. December releases happened in 2019 and 2020. Each year's probability therefore lands mainly in Q1, with a little in Q4.

Reference class and base rate

I downloaded NCES Digest Table 303.10 (total fall enrollment, 1947–2022; https://nces.ed.gov/programs/digest/d23/tables/dt23_303.10.asp) and computed:

• Worst one-year fall: −7.9% (1950→51).
• Worst two-year fall: −14.0% (1949→51).
• Largest peak-to-trough drop: −14.0%. The 2010→2022 slide was about −11.6%.

The Clearinghouse's own worst year was fall 2020 (pandemic): total −2.5%, community colleges −10.1%, first-time students −13.1% ("unprecedented"; CTEE Fall 2020, https://nscresearchcenter.org/wp-content/uploads/CTEE_Report_Fall_2020.pdf).

From memory, and not checked: even the WWII draft cut total enrollment by only about a fifth to a quarter. A 50% fall in five years has never happened in the US record. The demographic cliff is slow: high school graduates are projected to fall about 13% by 2041 (WICHE, via Fortune, Aug 31, 2026). Metaculus puts a decline of more than 10% in undergraduate enrollment from 2024 to 2030 at about 29% (search snippet only; the page returned 403). Halving is five times that threshold, far out in the tail.

Paths to YES

1. Very fast AI disruption that wipes out the payoff from degrees by about 2028, followed by mass non-enrolment and dropout. This is still hard to reach. New freshmen are only about 2.5M of 19.4M, so halving the total within 1–2 years means continuing students leave in large numbers too. Dual-enrolled high schoolers and graduate students would cushion the fall, and mass unemployment has historically raised enrollment.
2. Catastrophe (engineered pandemic, war). Here the Clearinghouse, based in Herndon, VA, would probably stop publishing, which resolves NO.
3. Gross data or methodology error. Past errors (the 2022 freshman errata, the 2024 preliminary freshman miscount) were a few points in subgroups, not a halving of the national total. Method changes need a comparable baseline.

Forecast construction

Probability that each fall term first shows a halving and gets reported:

• 2026: 0.01% (the term is already underway with records)
• 2027: 0.04%
• 2028: 0.08%
• 2029: 0.12%
• 2030: 0.15%
• 2031: 0.18%, times about a 15% chance of a December 2031 release

About 15% of each year's probability goes to Q4 (early release) and the rest to Q1. Q2 and Q3 stay flat. The cumulative probability is about 0.4% by Q1 2031 and 0.44% by end-2031.

Strongest case against: forecasters who expect transformative AI by 2027–28 might argue for 1–2%. But even in that world, behaviour lags opinion, institutions and state funding are sticky, and the Clearinghouse has to keep publishing comparable data. I would comfortably take 99:1 against YES, so I keep the estimate below 1%.

Question. By the last day of each quarter, has the National Student Clearinghouse Research Center reported total US postsecondary enrollment for a fall term at least 50% below its fall 2025 level?

Definitions.

• Total US postsecondary enrollment: the Research Center’s coverage-adjusted national total of undergraduate and graduate enrollment combined, across all institutional sectors in its final fall enrollment estimates. Use its enrollment-count measure, not full-time-equivalent enrollment, entering-student enrollment, degree awards, or the number of students in a particular major. Include international students and other student groups included in the published national total; “US” refers to the institutions, not student citizenship.
• Fall 2025 level (B): the national total in the Research Center’s Final Fall Enrollment Trends 2025 report, published January 15, 2026, including its official dashboard and Data Appendix. The report summarizes this as over 19.4 million; use the most precise published total, not that rounded summary. If the qualifying later release supplies an explicitly revised or methodologically restated fall 2025 total, use that comparable total instead. Otherwise use the latest official correction to the baseline available when the qualifying release is published.
• At least 50% below: the later fall-term national total E satisfies E ≤ 0.50 × B. Use the most precise official counts available. A rounded headline alone does not establish crossing if its rounding range straddles the threshold.

Resolves YES if an official final fall enrollment release, or an official correction to one, reports a qualifying national total for a fall term ending after the opening date. A renamed successor report counts if it publishes the same national enrollment measure. If coverage or methodology changes materially, the Research Center must provide a comparable fall 2025 baseline or explicitly establish the decline on a like-for-like basis. An explicit official statement that the comparable national total fell by at least 50% from fall 2025 also qualifies.

Exclusions (resolve NO).

• No qualifying final fall release by the quarter-end.
• Preliminary fall estimates, spring or summer enrollment, annual enrollment totals, projections, and hypothetical scenarios.
• Declines confined to undergraduate enrollment, a sector, state, demographic group, credential type, or major.
• A decline in raw submissions or reporting-institution coverage without a corresponding decline in the coverage-adjusted national estimate.
• Publications that only repeat or revise observations for periods ending on or before the opening date. No attribution of the decline to AI is required.

Resolving source. National Student Clearinghouse Research Center-authored releases only: Final Fall Enrollment Trends, its official dashboard, Data Appendix, corrections, and equivalent successor publications in the Clearinghouse Enrollment Insights series. The final fall report, formerly the fall Current Term Enrollment Estimates report, is normally published in January. This is a narrower source requirement than Public confirmation (A2). Apply Resolution date (A3), the Resolution window, and the Observation period conventions. A qualifying correction resolves on its publication date, not the date of the original release. Once YES, later revisions do not reverse it.