If you searched for “colleges most likely to close,” you probably want a straight list.
College Closure Watch is not going to invent one.
There is real public data about college finances, enrollment, federal oversight, accreditation and closures. That data can identify institutions under pressure. It cannot support a precise, reliable public ranking of which individual college will close next—at least not without making assumptions that are easy to hide and hard to validate.
So this page gives you the useful part: the strongest public warning signals, a way to find schools displaying them, and a clear line between evidence of stress and a prediction of closure.
Why “most likely to close” is harder than it sounds
A college can look weak in historical financial data and then receive a major gift. It can lose students and then merge with a stronger institution. A public university can run deficits but have state support that a private college does not. A small religious college may have backing from a denomination. A for-profit chain can close locations while the parent organization continues.
Even the outcome needs a definition. Does a merger count as a closure? What if the campus remains open but the old institution stops awarding degrees? What if one branch closes? What if the institution winds down over three years?
A prediction model has to answer all of those questions before its probability means anything.
What the public data can tell you
Enrollment direction
Large, sustained enrollment declines can create real pressure. College Closure Watch currently uses a 15 percent three-year undergraduate decline or a 25 percent five-year decline as a review threshold.
That flag is descriptive. It says enrollment crossed the threshold. It does not say a closure is more likely by a specific percentage.
Federal financial-responsibility status
Financial-responsibility composite scores and related requirements can identify institutions that did not meet the standard federal benchmark without alternative arrangements or oversight.
Useful? Yes. Real-time? Not always. Sufficient for a closure forecast? No.
Heightened Cash Monitoring
Federal Student Aid may place a school on HCM for financial or compliance reasons. HCM can add operational pressure, particularly at the more restrictive level, because it changes the timing and documentation around federal aid funds.
But HCM is not a government failure prediction. Treat it as an oversight signal and read the reason when available.
Accreditation and state actions
An accreditor show-cause order, probation or withdrawal can be highly consequential. A state can also impose conditions or revoke authorization.
These actions should carry significant weight in due diligence because they can affect the institution's ability to operate or students' program pathways. Yet even here, the precise action and effective date matter more than a generic “accreditation trouble” label.
Program cuts, layoffs and restructuring
Cost cutting can signal financial pressure, but it can also be rational management. The context matters. A university eliminating one low-enrollment major is not equivalent to an institution announcing a broad academic retrenchment after several years of enrollment decline.
The site's program tracker keeps confirmed program actions separate from institutional closure.
Why a simple risk score can mislead
Imagine giving every signal points: two for enrollment decline, three for HCM, four for accreditation probation. Add the points and rank the colleges.
It looks scientific. The problem is that the weights are arbitrary unless they have been validated against a clearly defined historical outcome. Signals can also be correlated. A financial-responsibility problem can lead to HCM, so counting both as fully independent evidence can double-count one underlying issue.
There is also missing data. Private institutions, public institutions and for-profits do not always expose the same information. A school with fewer public signals could simply have less observable data.
That is why College Closure Watch currently displays signal counts as review priority, not as a probability of failure.
A better way to evaluate a college
Step 1: look for multiple independent signals
One decline flag is a prompt to investigate. Several current signals from independent sources deserve more attention.
Step 2: read the dates
A 2021 concern and a 2026 concern are not equivalent. The direction of travel matters. Has the institution resolved the issue or accumulated new problems?
Step 3: read the primary documents
Do not stop at a red badge. Open the federal report, accreditor action or state notice. The underlying language is often more nuanced than a summary.
Step 4: ask what happens to your program
Institutional stability matters, but students experience risk through programs. Is your major large and central to the school? Is it professionally accredited? Have nearby programs been cut? Would another school take your credits?
Step 5: compare the cost of being wrong
If two colleges are similar in price and academic fit, but one has several unresolved official warning signals, stability can reasonably become a tie-breaker. If the school with the signals offers a unique program and a large scholarship, you may choose it—but do so with a contingency plan.
What historical closure data can eventually teach us
College Closure Watch has a growing historical closure database. That creates an opportunity to study what patterns were visible before past closures.
A responsible retrospective analysis could compare closed institutions with similar institutions that remained open. It could test how often large enrollment declines, federal financial signals or regulatory actions preceded closure. It could also measure false positives.
That is much stronger than looking only at colleges that failed and declaring every common trait predictive.
If the site later publishes a statistical model, it should disclose the outcome definition, training period, data vintages, missing-data treatment, validation results and error rates. Until then, the watchlist should remain what it says it is: an evidence index.
So which colleges should you review most closely?
Use the national watchlist and filter for institutions with multiple current official signals. That is the closest responsible answer to the search intent.
Then open the institution record and decide whether the evidence is relevant to your choice.
Do not confuse a watchlist with a closure list
For schools that have actually closed, use Closed Colleges. For authoritative announcements of closures this year, use Colleges Closing in 2026. For forward announcements, use Colleges Closing in 2027.
Those pages require a closure record or announcement. The watchlist does not.
Frequently asked questions about predicting college closures
Why not publish a top 20 list of colleges most likely to close?
Because a ranked list implies a level of predictive confidence the available public data may not support. It can also create serious reputational consequences for institutions and anxiety for students. A responsible site should distinguish observed warning signals from a validated statistical forecast. College Closure Watch currently publishes the former.
Could machine learning predict college closures?
Potentially, but a credible model would require careful outcome definitions, historical training data, comparable institution records, treatment of mergers and branch closures, and genuine out-of-sample validation. It would also need to report false positives and false negatives. A model that simply fits past closures and ranks current schools without those safeguards could look sophisticated while being unreliable.
Which signal has the strongest historical relationship with closure?
That should be measured rather than assumed. Large enrollment declines and financial distress commonly appear in closure narratives, but a valid analysis needs a comparison group of institutions that displayed the same signals and did not close. College Closure Watch's historical database creates a foundation for that research, but the site should not declare a “best predictor” without publishing the method and results.
Is a school with four signals necessarily riskier than one with two?
Not necessarily. The four signals may be correlated versions of the same underlying issue, while the two signals at another institution could be more severe and recent. Signal count is an efficient research filter, not a calibrated risk probability.
What is the safest way to use a closure-risk watchlist?
Use it to identify questions. Open the source behind each flag, compare the dates, look at the full enrollment trend and ask the college for its explanation. Then consider how difficult it would be to transfer or complete the degree if conditions changed. The watchlist should improve due diligence, not replace judgment.