The End of the Growth Era and the New Economics of Academia.

The Economics of Attrition

For nearly sixty years, American higher education answered almost every strategic problem the same way: by expanding. Universities added academic programs, opened research labs, fielded more athletic teams, and built new campus facilities in pursuit of comprehensive status. That era of debt-funded expansion has ended.

According to Jeff Selingo’s report, The State of Higher Education, the financial pressure on colleges is no longer a cyclical downturn but a permanent structural deficit. Over the past decade, 387 colleges, or one in seven nationwide, lost money cumulatively. Of those, 135 posted operating losses in at least eight of the last ten years. Operating margins at the private colleges Moody’s tracks are the thinnest in two decades. To keep enrollment steady, small private institutions now discount tuition by nearly 55 cents on every dollar, which eats into their own revenue.

The traditional engines of growth are also faltering. Undergraduate enrollment remains volatile. Spring 2026 enrollment rose 1.3% to 15.5 million students, but four-year institutions saw little of that gain, since it came almost entirely from dual-enrolled high school students at community colleges. Overall, undergraduate enrollment is still nearly 400,000 below its 2019 peak. Enrollment by field is shifting as well: Computer & Information Sciences fell 8.4% and English fell 7.7%, while Engineering rose 6.8% and Health Professions rose 6.2%.

The graduate market, once a reliable source of revenue, is weakening too. International graduate enrollment rose 53% after the pandemic and helped prop up university budgets, largely through master’s students from India, but it fell 12% in fall 2025. Domestic graduate students have also shifted toward part-time, working professionals. Education master’s programs are now 62% part-time and Business programs 49% part-time, and these students want flexible, online formats that pull enrollment away from traditional, high-margin residential programs. Programs offered online now award nearly 100,000 more master’s degrees a year than classroom-only programs.

Even so, with revenues shrinking, domestic demographics flat, and student preferences changing, the cost of running a university keeps climbing. The main reason is that higher education is labor-intensive and institutions have been unwilling to cut payroll.

The Department of Education’s College Affordability and Transparency Report shows the same pattern. Federal law requires the 5% of institutions with the largest tuition and net price increases to explain them, and their answers were nearly identical. Of the 446 institutions surveyed, which spent an average of 2.19 burden hours on their responses, most cited fixed staffing costs and higher operating costs carried over from the pandemic era.

The report also shows little appetite for cutting costs. When asked what they plan to do to reduce costs for students, 60% of responding institutions said they have no plans to reduce costs. They cited inflation, state-mandated wage increases, and technology upgrades (often continuing spending that began with HEERF/CARES Act funds) as obstacles. One private institution wrote, "Due to the high cost of living and the increase of minimum wage, salaries can't be reduced."

In addition, 82% of institutions said they do not have sole control over setting their tuition and fees and defer to state legislatures or governing boards. Mission requires margin, but the traditional sources of margin (over-enrollment, strong state appropriations, and growing international master’s programs) have dried up. Institutions are now raising prices to make payroll, adding costly programs to attract fewer students, and relying on the federal government to keep underwriting student debt.

Washington's ROI Mandate

For years, universities covered structural operating deficits by relying on federal student loans. Encouraging students and parents to borrow large sums let institutions pay for administrative growth and expensive campus amenities without immediate consequences. Washington has since moved from funding enrollment growth to requiring proof of outcomes.

Selingo notes that federal loan limits tightened in July 2026. The Grad PLUS program, which had no cap and financed a large expansion of master’s programs, is now capped at $20,500 a year and $100,000 total. Parent PLUS borrowing is capped at $20,000 a year and $65,000 total. This hits private higher education hardest: 54% of Parent PLUS borrowers constrained by the cap are enrolled at private colleges. Without unlimited federal loans to pass costs on to parents, these institutions must discount tuition out of their own shrinking endowments, push students toward private loans, or lose those students.

The Department of Education has also set an earnings floor for every academic program. If graduates of a program do not out-earn high school graduates in the same state, the program loses access to federal loans. While 95% of students are in programs that clear this bar, the failures are concentrated.

More than 644,000 students, representing $2.7 billion in loans, are enrolled in failing programs. At the certificate level, 93% of cosmetology programs fail. At the bachelor’s level, theater, fine arts, and religious studies programs have failure rates above 15%. At the master’s level, mental and social health services is the only large field where most programs fail.

The new “Workforce Pell,” which allows Pell Grants to be used for short-term training, comes with the highest performance bar yet attached to federal aid. To qualify, programs must pass three tests: a structural screen (150-599 clock hours), an outcomes screen (70% completion rate and 70% job placement rate), and a value-added earnings screen (tuition cannot exceed the earnings gain). Credential Engine estimates that tens of thousands of programs meet the length requirement, but the federal government says only a few hundred currently pass the placement and earnings tests.

Research Funding and AI

At the top of the hierarchy, the research enterprise is growing, reaching a record $117.7 billion in FY 2024. But the terms are changing. The federal government is pulling back, and universities now cover 26% ($30.2 billion) of research costs from their own funds. The risk is concentrated: 15 of the 18 largest universities with medical centers depend heavily on federal funding aimed almost entirely at the life sciences, which leaves them exposed to decisions at Washington’s health agencies.

Commercialization and tech transfer are not a solution either, despite administrators’ rhetoric. Universities launched 1,095 startups in 2025, but gross licensing income equals only 3% of the sector’s total research spending, and most tech-transfer offices do not cover their own operating costs.

The larger problem in research is a change in what discovery requires. The scarce resource is no longer lab space or researchers but computing power and data storage.

According to Selingo, training a single frontier AI model now costs more than the entire annual research computing budget at most universities. Of 95 notable AI models released in 2025, 93 came from private industry and only two from academia. Federal initiatives such as the NAIRR (National AI Research Resource) and the Energy Department’s $5 billion Genesis Mission aim to close the gap, but faculty are likely to move to the few institutions that have the compute needed for modern science.

Universities also face growing cybersecurity risk. Moody’s notes that education has the highest rate of unresolved, known-exploited vulnerabilities of any sector it tracks. Larger budgets do not prevent this: 97% of institutions with more than $5 billion in revenue had a vulnerability that went unpatched for over a year. The median ransom demand has risen to $420,620, and major infrastructure such as Canvas has already suffered serious cyberattacks.

Campuses are buying AI tools to improve administrative efficiency, but only 29% of Chief Technology Officers say AI has met their ROI expectations. AI mostly saves money by replacing staff. Anthropic data suggest AI is well suited to automating entry-level tasks in management, finance, and office administration, which are the tasks recent graduates use to learn their professions. Higher education institutions have historically been reluctant to eliminate jobs, so they are paying for AI compute out of tight budgets without getting the savings promised.

Governance and Leadership

These financial, structural, and technological pressures are landing on governing boards and leadership structures built for a calmer period. Many boards were assembled to reward major donors or serve political patronage, not to make the hard financial decisions required now. Data from the Fiduciary Effectiveness Quotient (FEQ) show that the strongest boards share three traits: they are small (under 20 members), they attend meetings (92% attendance), and they refresh their membership (27% annual turnover). Weak boards do the opposite.

The executive pipeline is thinning as a result. Average college president tenure has fallen from 8.5 years in 2006 to 5.9 years today. Some 55% of sitting presidents plan to leave within five years, and 25% within two, yet only one in four institutions reports having a viable succession plan.

This instability comes at a time of low public trust. Confidence in higher education has fallen to 38%, driven largely by a sharp decline among Democrats (to 50%), while confidence among Republicans and Independents is lower still (23% and 39%, respectively). Nearly 9 in 10 enrolled students, though, remain confident that their degree will prepare them for the job market, the widest gap between the people using the product and the public judging it.

Administrators are being asked to cut back their institutions under these conditions. Program elimination has gone from a crisis response to an annual budget step. Indiana’s public campuses have eliminated or consolidated 1,279 programs since 2025, and Penn State is effectively closing seven campuses. Leaders are trying to narrow their institutions’ scope while dealing with hostile public sentiment and faculty senates that were designed to slow change, not speed it up.