Skip to main content

Which Colleges Are Most at Risk?

We built a financial stress score from three variables (enrollment trend, tuition dependency, and cash reserves) and applied it to 1,516 institutions from the public-data cohort. With only 6 observed closures, the score can be inspected but not treated as a validated forecast.

Stress Score Distribution

Where Do Closures Cluster?

The histogram below shows the distribution of financial stress scores for operating institutions (blue) and those labeled closed (rose). The six closures do not form a clean high-score cluster: one is in the Moderate tier, three are High, and two are Critical.

Stress Score Distribution: Operating vs. Closed
Finding
The selected rule detected 0 of 6 observed closures and produced 59 false positives. The real-data output does not validate this threshold.

Financial stress score: enrollment trend (5-year compound annual growth rate) + tuition dependency (tuition revenue / total revenue) + days cash on hand. Scored 0–100, higher = more stressed.

Risk Tier Breakdown

How Many Are in the Danger Zone?

We grouped institutions into four tiers based on stress score. The Critical tier contains 548 institutions and 2 observed closures; the High tier contains 3, the Moderate tier 1, and the Low tier none.

Institutions by Risk Tier

The tiers overlap too much to support a confident cutoff. Closed labels appear in three of the four tiers, while most Critical institutions remain operating in the current records. The tier is a stress indicator, not a closure prediction.

Watchlist

Top 20 At-Risk Operating Institutions

These are the 20 currently-operating institutions with the highest computed stress scores. They are a data-quality and screening watchlist, not a prediction that any named institution will close.

Institution State Sector Enrollment Score Tuition Dep. Days Cash

Source: College Scorecard and IPEDS (Integrated Postsecondary Education Data System) 2015–2023 data. Stress score computed from enrollment trend (5-year compound annual growth rate), tuition dependency ratio, and days cash on hand.

Model Validation

How Good Is This Simple Rule?

At the selected threshold, the public output reports 0% recall and 0% precision. Those values are based on only six positive cases, so they describe this retained cohort rather than stable future performance.

Transparency is useful, but it does not rescue a failed validation. The score remains a reasonable screening prototype because its inputs and logic are visible. Before operational use, it needs a broader outcome set, threshold redevelopment, and external validation.