🧮 Methodology: How We Calculate University Survivability
The Collegiate Survivability Index (CSI) is a data-driven model designed to forecast which private, nonprofit universities are most at risk of closure by 2028.
It combines financial, academic, and demographic indicators to produce a percentile ranking for each institution. A higher CSI percentile score indicates a stronger likelihood of long-term sustainability.
🧱 The Three Core Components
Each institution is scored across three key categories using public data from IRS Form 990s, IPEDS, and the U.S. Census.
1. Adjusted Financial Health (40.02%)
We created an improved version of the traditional Composite Financial Index (CFI), known as the adjusted CFI (aCFI). This revision accounts for financial realities that the original CFI overlooks:
- Depreciation is removed, since it distorts short-term viability for tuition-dependent schools.
- Restricted assets are included, recognizing that institutions often access donor-restricted funds in times of crisis.
- All values are standardized using Z-scores, making comparisons fair across institutions of different sizes.
This gives a more realistic picture of an institution’s solvency over the next 3–5 years.
2. Market Saturation (39.99%)
We evaluate how crowded a school's recruitment territory is by calculating:
- The 15–19-year-old population in the state
- The number of higher ed institutions in the region
- Each university’s first-time, full-time enrollment
This creates a saturation score that reveals how fiercely a university must compete for traditional students — especially as birthrates decline.
3. Academic Efficiency (19.99%)
Academic efficiency measures whether a university is delivering strong outcomes with sustainable instructional resources. Key indicators include:
- Graduation rate
- Student-to-faculty ratio
- Degrees awarded per faculty member
- Instructional spending per student
While elite schools may intentionally maintain small classes, many low-ranked institutions show inefficiencies due to weak retention and program bloat.
🔁 Forecasting the Future
Each component is projected to 2028 using Holt-Winters exponential smoothing, a forecasting technique that accounts for trends and seasonality. To fine-tune the model:
- We apply Winsorization to control outliers
- We run a Sobol Sequence sensitivity analysis to balance the influence of each variable
- We use Bayesian optimization to adjust weights dynamically as new data becomes available
📌 A Model That Learns and Updates
The CSI is not static. It is updated continuously as:
- New Form 990s are released
- IPEDS enrollment and graduation data are refreshed
- Colleges announce closures or mergers
Institutions that were initially excluded due to missing data will be added as information becomes available through open records requests or direct outreach.
🧠 Why CSI Matters
Unlike the federal CFI — which failed to flag at-risk schools like Limestone University and St. Andrews University — our CSI model correctly identified both as high-risk months before closure announcements.
This makes the CSI a powerful, transparent early-warning system for:
- University leadership and boards
- Donors and accreditors
- Parents and prospective students
- Journalists and policymakers