Personal project · Higher education

College Closure & Financial Distress Forecaster

This model is a survival-analysis pipeline that predicts which US private nonprofit colleges will close 1 to 4 years ahead, using only public data. It is a continuation of my interest in higher-ed economics that began with my senior thesis looking at test-optional admissions.

The project has three parts: a college distress index that predicts how likely each institution in the sample is to close in the next four years; a college report card that grades each institution on its command of the market and its financial resilience; and a demographic look at the higher-ed landscape, showing where in the country students have been going over the past two decades and which type of institution, public or private, has been the big winner.

The distress index was the foundation of the project; the report card and the enrollment analysis came after I read the Wall Street Journal’s article on Syracuse University’s enrollment and budget struggles. While that article was directionally correct about many national demographic trends (a shrinking high school class, a shift away from private, northern colleges toward public, southern schools, and shrinking international enrollment), it is unclear whether Syracuse University will close any time soon. Although the models’ information is two years delayed because of the lag in IPEDS data, Syracuse is still one of the safest schools on the distress index and earns a B on the report card. The article is more illustrative of the fact that even regional powerhouses like Syracuse need to pay attention to the changing landscape in higher education.

Explore the tools
Interactive · Closure risk

College Distress Index

Every private nonprofit four-year college ranked by its modeled chance of closing within four years. Compare the models and trace any school’s risk back to 2008.

Open the index
Interactive · Grades

College Report Card

An A+ to F grade built from two pillars, command of the market and financial outlook, scored on absolute thresholds and traced year by year since 2010.

Open the report card
Interactive · Market

Enrollment by State

Where undergraduates are gaining and losing ground, public versus private, year over year or across the last decade, with campus mergers and reporting changes netted out.

Open the map

Data through fall 2024. Each tool opens as its own page.

Data

The model runs on an institution-year panel of roughly 1,200 colleges from 1998 to 2024, assembled from IPEDS (via the Urban Institute API and raw NCES files), FSA/PEPS closure records, WICHE cohort projections, and Census demographics. Closure-versus-merger labels are hand-verified against roughly 20 known events, since the two look identical in most administrative data but mean very different things.

Approach

Three models: a discrete-time hazard logit for interpretable coefficients and calibrated probabilities, a Cox proportional-hazards model as a robustness check, and a class-weighted LightGBM for nonlinearities. All three are evaluated by rolling-origin backtest in which every feature respects its real publication lag. This matters more here than in most settings: IPEDS finance data arrives two years late, so a model that ignores the lag is predicting closures with data that would not have existed at decision time.

Results and caveats

The full model beats the Department of Education's own financial-responsibility score from public data alone: ROC 0.85 versus 0.74 on identical rows. Two caveats are worth stating. A two-variable benchmark, enrollment trend and tuition dependence, still wins in the extreme tail. And the demographic cliff is not yet detectable in closures through 2024 once calendar time is controlled for.

The pipeline is refreshable annually as new IPEDS data publishes, with a CLI that scores the live universe of institutions. Two of the model's 2024 top-25 highest-risk institutions, Fontbonne and Northland, closed in 2025, after the model's information window.