Abstract
The phrase “the crisis of higher education” names a mood more reliably than it names a problem. Beneath the single label sit at least seven distinct failures—cost and debt, access and equity, value skepticism, demographic enrollment decline, credential inflation, administrative cost growth, and chronic non-completion—each with its own causes, its own evidence base, and its own plausible remedies. This paper disaggregates the bundle, traces the structural mechanisms that drive its two best-documented dimensions (the cost engine and the completion gap), and then maps the University of the People (UoPeople) model onto the resulting taxonomy. The mapping distinguishes three categories: dimensions the model addresses directly by design, dimensions it addresses only indirectly or contingently, and dimensions it does not address and was never built to address. The paper closes by deriving the evaluative criteria that the remaining six papers in this section will apply. Its claim is methodological rather than promotional: a reform model can be judged honestly only against the specific failures it targets, and the first task of any serious assessment is to refuse the comfort of a single name.
1. The Problem of the Single Name
Every generation announces a crisis in education, and the announcement is usually right in its mood and wrong in its diagnosis. The mood is one of legitimate alarm. The diagnosis tends to fail because it treats a cluster of separable problems as one ailment with one cure. When a commentator says that higher education is “broken,” the listener cannot tell whether the complaint is about price, about debt, about who gets in, about whether the degree still pays, about half-empty lecture halls, about administrative overgrowth, or about the students who enroll and never finish. These are not facets of a single condition. They are different conditions that happen to share a patient.
The distinction matters because remedies are dimension-specific. A reform that lowers price may do nothing for completion. A reform that widens access may worsen completion if it admits students the model cannot then support. A reform that restores public confidence in the degree’s value may leave the cost structure entirely intact. Conflating the failures produces a predictable analytic error: a model is praised or condemned wholesale, when the accurate judgment is partial—strong on two dimensions, weak on a third, silent on a fourth. This paper exists to prevent that error for the rest of the suite. Before any one institution can be evaluated, the criteria of evaluation must be built, and they can be built only by taking the crisis apart.
The institution this section examines, the University of the People, is useful precisely because its three headline claims—tuition-free study, regional accreditation, and borderless admission—press on three different pressure points at once. That makes it a clean instrument for demonstrating why the dimensions must be held apart. A model that touches cost, legitimacy, and access simultaneously will reward a disaggregated scorecard and punish a single verdict.
2. Methods Note
This is a framework-building paper, and its method is conceptual and documentary rather than empirical in the primary sense. It proceeds in three moves. First, it constructs a taxonomy of the failures commonly folded into “the crisis,” defining each dimension narrowly enough to be assessed on its own terms and grounding each in published data or in the established scholarly literature on the relevant mechanism. Second, it maps the UoPeople model onto that taxonomy, sorting the dimensions into direct targets, indirect or contingent effects, and out-of-scope concerns. Third, it converts the taxonomy into a set of evaluative criteria—a scorecard schema—that White Papers 2 through 7 will apply with their own evidence.
The evidence base for the taxonomy draws on three kinds of source: national statistical series (student-debt totals, tuition indices, enrollment and completion data) reported by federal agencies and clearinghouse research centers; the peer-reviewed and book-length scholarly literature on the underlying mechanisms (cost disease, the revenue theory of cost, credential signaling, demographic demand); and the institution’s own communications regarding accreditation, enrollment, and mission. Figures attributed to UoPeople are self-reported and are treated as claims to be tested in later papers, not as settled findings here. The paper makes no causal estimate of any reform’s effect; it builds the rubric against which such estimates will later be read.
3. Disaggregating the Crisis: A Taxonomy of Seven Failures
3.1 Cost and debt
The most visible dimension is price and the borrowing that price compels. Total outstanding student-loan debt in the United States stood at roughly $1.84 trillion at the end of 2025, a figure that includes both federal and private loans. The per-borrower average has climbed in parallel; average federal student-loan debt reached about $39,000 per borrower by 2025, close to twice the average of 2008. These are stock figures, and they understate the lived problem, which is the timing of repayment against early-career earnings: debt service falls hardest in the years when income is lowest and household formation most expensive.
Two clarifications keep this dimension distinct from the next. First, debt is a consequence variable; it is downstream of price, aid design, and time-to-degree. Second, the sticker price and the net price diverge widely, so “cost” must be specified as cost to whom—the listed tuition, the discounted tuition after institutional aid, or the all-in cost of attendance including foregone earnings and time. A model that claims to solve cost must be asked which of these it actually moves.
3.2 The cost engine: why tuition outran inflation
Underneath the debt totals sits a structural puzzle: tuition has risen faster than general prices for decades. Over the two decades ending in 2022, consumer prices rose roughly 39% while published college tuition rose about 68%, and by longer-horizon measures tuition has climbed nearly 180% over twenty years. Three competing explanations dominate the literature, and they are not mutually exclusive.
The first is Baumol’s cost disease. In labor-intensive sectors where productivity gains are hard to capture—a string quartet still needs four players—wages must nonetheless track those in sectors where productivity rises, so unit costs climb relative to the rest of the economy (Baumol & Bowen, 1966; Baumol, 2012). Teaching, on this view, resists the automation that cheapens manufactured goods, so its relative price rises by structural necessity rather than mismanagement.
The second is the revenue theory of cost, often called Bowen’s law: institutions raise all the money they can and spend all the money they raise, because in a prestige competition there is no natural ceiling on desirable expenditure (Bowen, 1980). Cost rises to meet revenue rather than the reverse.
The third is the Bennett hypothesis: subsidized credit expands what students can pay, and institutions capture part of that expanded capacity in higher prices (Bennett, 1987). Empirical work has lent this real support; research on the expansion of federal aid finds that increases in loan availability were associated with tuition increases at affected institutions (Lucca et al., 2019).
These three mechanisms point to different remedies, which is the analytic payoff. Cost disease implies that the only durable relief comes from changing the production function—substituting capital and self-directed or peer-mediated learning for salaried instructional labor. Bowen’s law implies that relief requires a binding external constraint on revenue. The Bennett hypothesis implies that relief requires decoupling price from subsidized credit. A model that removes tuition entirely, as UoPeople claims to, is in effect a wager on the first mechanism; whether the wager holds is the subject of White Paper 2.
3.3 Access and equity
Access is the dimension most often invoked and least often specified. Barriers to entry are at least fourfold and behave differently. Financial barriers exclude those who cannot pay even discounted tuition or cannot forego earnings to study. Geographic barriers exclude those who cannot relocate to or commute to a campus. Political barriers exclude those displaced by conflict, denied recognition, or barred by their own jurisdictions—most sharply, refugees and women under regimes that forbid their education. Personal barriers—caregiving, employment, disability, prior academic record—exclude those whose lives will not bend to a fixed residential schedule.
Holding these apart is essential because a model can demolish one barrier while leaving another untouched, or even raise a new one. Removing tuition addresses the financial barrier but not the bandwidth barrier; admitting across borders addresses geography but not credential recognition in the student’s home labor market. Access, in short, is not one gate but several, and a model’s access claim must be audited gate by gate.
3.4 The legitimacy problem: value and ROI skepticism
Distinct from price is the public’s growing doubt that the degree delivers what it once did. This is a perception problem with a partial basis in fact. The signaling literature holds that much of a degree’s labor-market payoff comes from what it certifies about the holder rather than from learning added, with sorting and screening doing work that instruction is often credited for (Caplan, 2018; and, on the historical sociology of credentials, Collins, 1979). Direct measures of learning gain have been discouraging in places; influential work found limited measurable gains in critical reasoning over the first two years of college for a large share of students (Arum & Roksa, 2011). Whether or not one accepts the strong signaling claim, the legitimacy dimension is real as a belief, and beliefs about value drive enrollment decisions and willingness to pay. A model can be cheap and accessible yet still fail this test if employers and the public do not credit its degree—which is why accreditation, the sector’s central legitimacy signal, carries the weight it does (White Paper 3).
3.5 Demographic and enrollment decline
A structural contraction in the traditional-age population now bears on the sector. Falling birth rates during and after the 2008 recession produce a shrinking pool of eighteen-year-olds reaching college age in the latter half of this decade; the Western Interstate Commission for Higher Education projects roughly a 13% decline in high-school graduates by 2041, and demographic modeling anticipates a steep drop in the number of eighteen-year-olds entering college between 2025 and 2030 (see also Grawe, 2018, 2021). Undergraduate enrollment had already fallen about 6.6% from its pre-pandemic level by the mid-2020s, more than a million fewer students, though aggregate postsecondary enrollment ticked upward again in 2025.
This dimension is largely exogenous to any single institution—no school can raise the birth rate of eighteen years prior—and it is regionally uneven, falling hardest on tuition-dependent institutions in the Midwest and Northeast. Its relevance to the present case is oblique but real: a model whose addressable population is global and whose marginal cost per student is low is structurally insulated from a domestic demographic contraction in a way that a place-bound, tuition-dependent regional college is not.
3.6 Credential inflation
As more of the population holds a given credential, that credential certifies less, and the threshold for a given job rises—the bachelor’s degree comes to occupy the social position the high-school diploma once held (Collins, 1979; Labaree, 1997). Credential inflation is partly a positional arms race: education delivers relative advantage that erodes as others acquire the same signal, even as the economy’s genuine demand for skill also rises (Goldin & Katz, 2008). The dimension cuts two ways for any access-widening model. By extending degrees to populations previously excluded, such a model serves equity; but to the degree it adds to the supply of a credential whose value is partly positional, it participates in the very inflation that erodes the degree’s worth. A reform cannot resolve credential inflation, because the dynamic is systemic; it can only decline to make it worse, or accept that it adds to it in exchange for distributional gains.
3.7 Administrative cost growth
A specific and well-documented contributor to the cost engine is the expansion of non-instructional staff. Across recent decades the administrative ranks of American colleges grew faster than either enrollment or faculty; non-instructional spending outpaced instructional spending between 2010 and 2018, and institution-level analyses find administrator headcount rising far ahead of faculty headcount even as student numbers held roughly flat (for one documented case, a 78% rise in administrators against a 40% rise in faculty over twenty years; the foundational account is Ginsberg, 2011). This dimension matters to the present case because it identifies a cost layer that a lean, campus-free model removes almost entirely by construction—an advantage examined in White Paper 2—while raising the question of whether the functions those administrators performed (student support, compliance, retention services) are eliminated or merely shifted onto students and volunteers.
3.8 Completion failure
The last dimension is the most under-discussed and arguably the most damaging: students who enroll, borrow, and leave without a credential bear cost without benefit. Nationally, about 61% of students who began college in fall 2019 had earned a credential within six years, while nearly 30% had stopped out. The mechanisms behind non-completion include under-preparation, financial shock, and the thinning of per-student resources as enrollment expands (Bound et al., 2010). Completion is the dimension where access and quality collide most directly: widening the door without strengthening the path to graduation can convert an access gain into a debt-without-degree loss. Any model that lowers entry barriers must therefore be asked not only who it admits but who it graduates—a question that bears especially hard on self-directed, low-touch designs.
4. How the Failures Interact
The dimensions are separable but not independent, and several of their interactions are worth naming because they will recur in the scorecard. The cost engine (3.2) feeds debt (3.1) and is fed by administrative growth (3.7). Debt and price feed value skepticism (3.4): the more a degree costs, the more its payoff is scrutinized. Value skepticism interacts with demographic decline (3.5): a shrinking cohort that also doubts the degree’s worth produces a compounded enrollment threat. Access (3.3) and completion (3.8) stand in the sharpest tension, since the easiest way to improve measured completion is to admit only the already-advantaged, and the easiest way to widen access is to admit students the model may then fail to support. Credential inflation (3.6) sits across the whole system as a slow erosion that no single institution controls.
Recognizing these couplings guards against a second analytic error, the assumption that improvements travel together. They do not. A model can lower cost and widen access—the two UoPeople most plausibly addresses—while leaving completion and credential-value exposed precisely because of the interaction between access and completion. Coherence among a model’s strengths is a real feature; the assumption that strengths imply the absence of weaknesses is a fallacy the scorecard is designed to block.
5. Mapping the UoPeople Model onto the Taxonomy
With the dimensions defined, the model can be placed against them. UoPeople’s relevant features, as the institution reports them, are these: tuition-free study funded by a non-profit model with assessment fees rather than tuition; open educational resources in place of licensed content; no physical campus; a lean, largely volunteer instructional staff; asynchronous peer-cohort pedagogy; admission across borders by default; a completed move from national (DEAC) to regional (WSCUC) accreditation, with the regional recognition announced in February 2025 and the national relationship withdrawn at the end of that year; and an enrollment the institution reports at more than 170,000 students from 213 countries, including the largest refugee enrollment of any university and more than 5,000 Afghan women, supported by partnerships with UNESCO, UNHCR, and UNICEF (University of the People, 2025). These figures are self-reported and are tested, not assumed, in later papers. Sorting the eight dimensions yields three categories.
Addressed directly, by design. Cost and debt (3.1) and the cost engine (3.2) are the model’s primary targets: it removes, rather than discounts, the three most expensive line items—licensed content, physical plant, and a large salaried faculty. Administrative cost growth (3.7) is addressed as a near-automatic consequence of a campus-free, lean structure. Access, in its financial and geographic forms (3.3), is addressed directly by zero tuition and borderless admission. These are the dimensions on which the model’s claim is strongest and most plausible on its face, and they are the subject of White Papers 2 and 4.
Addressed indirectly or contingently. The legitimacy/value dimension (3.4) is addressed only through the accreditation transition, which is a necessary but not sufficient condition for public and employer confidence; the model’s value claim rests on outcomes evidence that remains to be weighed (White Paper 3). Access in its political form—the refugee and displaced-learner mission (3.3)—is addressed powerfully but contingently, since the barriers the model removes (geography, tuition) coexist with barriers it cannot remove (bandwidth, language, home-jurisdiction recognition). Demographic decline (3.5) is addressed only obliquely: the model is insulated from a domestic contraction by its global, low-marginal-cost design, but insulation is not the same as solving the sector’s demographic problem.
Not addressed, and not designed to be. Completion (3.8) is the dimension the model does not resolve by construction and may strain, because self-directed, low-touch designs place heavy demands on learner persistence; this is the central risk White Paper 5 must confront. Credential inflation (3.6) is systemic and beyond any single institution’s reach; by extending degrees to new populations the model serves equity while adding, at the margin, to the supply of a positional good. Naming these as out-of-scope is not a criticism—no model resolves every dimension—but it is essential to an honest verdict.
The mapping yields the section’s organizing claim in sharpened form: the UoPeople model is built to resolve cost and financial-geographic access, is contingent on evidence for legitimacy and the refugee mission, and is structurally exposed on completion and silent on credential inflation. A wholesale verdict—”it works” or “it doesn’t”—would obscure exactly this structure.
6. The Evaluative Criteria for the Suite
The taxonomy converts directly into a scorecard schema, which White Paper 7 will populate after Papers 2 through 6 supply the evidence. For each dimension the suite will ask a defined question and identify the evidence that would answer it.
On cost and the cost engine: does the model lower the true cost of a credential, or shift cost onto students’ own time and unpaid labor, and do its unit economics survive growth? (White Paper 2.) On legitimacy and value: does the accreditation transition, together with outcomes and placement data, support employer and public confidence in the degree? (White Paper 3.) On access: which barriers does the model actually remove, gate by gate, and what does the refugee mission demonstrate about access-first higher education at its limit? (White Paper 4.) On completion and pedagogy: does peer-cohort, low-touch instruction graduate the students it admits, and how does its attrition compare with the relevant benchmarks? (White Paper 5.) On sector position and replicability, including credential inflation and demographic exposure: what is proprietary to the model versus transferable, and could a conventional institution adopt its strengths without its costs? (White Paper 6.) The final paper returns to all eight dimensions and grades them, engages the serious critiques, and weighs the forward threats—accreditation durability, financial stability, and the double-edged arrival of generative artificial intelligence. (White Paper 7.)
The governing rule of the scorecard is stated once here and applied throughout: a model earns credit only on the dimensions it targets, loses no credit for the dimensions it never claimed, and is judged most severely on the dimensions where its design and its claims pull against each other—above all, access against completion.
7. Limitations
This paper has the limitations proper to a framework. First, the taxonomy is a choice, not a discovery; the seven-plus dimensions could be cut differently, and a critic might fold administrative growth into the cost engine or split access further. The defense is utility, not uniqueness: these cuts are the ones that let the model be assessed dimension by dimension. Second, the dimensions are presented as separable for analytic clarity, but the interactions in Section 4 mean that any scorecard will involve judgment calls about where a coupled effect should be booked. Third, the UoPeople figures used in the mapping are institutional self-reports and remain to be tested; this paper deliberately defers that testing rather than crediting the claims. Fourth, the framework is built around one institution and one model; its criteria are designed to generalize to other access-first designs, but that generalization is asserted here and demonstrated only later in the volume. Finally, the paper takes no normative position on whether access ought to be the sector’s first principle; it builds the instrument for judging a model that assumes so, and reserves the normative argument for the synthesis.
8. Conclusion
The usefulness of the University of the People as an opening case lies in how cleanly it forces the disaggregation this paper has performed. Because the model presses on cost, legitimacy, and access at once, it cannot be assessed by a single verdict without distortion; it requires exactly the dimension-by-dimension scorecard the taxonomy yields. The provisional shape of that verdict is already visible: the model is strongest where it is designed to be strong—cost and financial-geographic access—contingent where it depends on evidence yet to be weighed, and exposed where its low-touch design meets the completion problem. Whether that shape holds is the work of the six papers that follow. The contribution of this first paper is the discipline that makes their work legible: the refusal to let “the crisis of higher education” stand as one word for many problems, and the insistence that every proposed remedy be asked, plainly, which of those problems it actually answers.
Notes
¹ The seven-dimension taxonomy is offered as an analytic convenience rather than a settled ontology. Other treatments enumerate the failures differently; the test applied here is whether each dimension can be measured and remediated on its own terms, not whether the list is exhaustive or canonical.
² “Cost disease,” “the revenue theory of cost,” and “the Bennett hypothesis” are presented as competing but compatible explanations. The paper does not adjudicate among them; it observes only that each implies a different remedy, and that the UoPeople model corresponds most closely to the production-function change implied by the cost-disease account.
³ All enrollment, accreditation, and mission figures attributed to the University of the People are self-reported institutional communications as of late 2025 and are treated throughout the suite as claims to be verified, not as independent findings. The accreditation chronology—regional recognition through the WASC Senior College and University Commission announced in February 2025, with the prior national (DEAC) relationship withdrawn at the close of 2025—is examined in White Paper 3.
⁴ National completion figures report credential attainment within six years for a given entering cohort and are not directly comparable to an institution’s own graduation rate without adjustment for student mix, part-time status, and starting sector; the comparison is developed in White Paper 5.
⁵ The interaction between access and completion (Section 4) is the single most important caution the framework issues, because it is the interaction most likely to convert a measured access gain into a debt-without-degree loss for the student.
References
Archibald, R. B., & Feldman, D. H. (2011). Why does college cost so much? Oxford University Press.
Arum, R., & Roksa, J. (2011). Academically adrift: Limited learning on college campuses. University of Chicago Press.
Baumol, W. J. (2012). The cost disease: Why computers get cheaper and health care doesn’t. Yale University Press.
Baumol, W. J., & Bowen, W. G. (1966). Performing arts: The economic dilemma. Twentieth Century Fund.
Bennett, W. J. (1987, February 18). Our greedy colleges. The New York Times.
BestColleges. (2024). College tuition inflation statistics. https://www.bestcolleges.com/research/college-tuition-inflation-statistics/
Bound, J., Lovenheim, M. F., & Turner, S. (2010). Why have college completion rates declined? An analysis of changing student preparation and collegiate resources. American Economic Journal: Applied Economics, 2(3), 129–157.
Bowen, H. R. (1980). The costs of higher education: How much do colleges and universities spend per student and how much should they spend? Jossey-Bass.
Caplan, B. (2018). The case against education: Why the education system is a waste of time and money. Princeton University Press.
Collins, R. (1979). The credential society: An historical sociology of education and stratification. Academic Press.
Edvisorly. (n.d.). How colleges can prepare for the enrollment cliff. Retrieved May 2026, from https://www.edvisorly.com/university-insights/enrollment-cliff
Ginsberg, B. (2011). The fall of the faculty: The rise of the all-administrative university and why it matters. Oxford University Press.
Goldin, C., & Katz, L. F. (2008). The race between education and technology. Harvard University Press.
Grawe, N. D. (2018). Demographics and the demand for higher education. Johns Hopkins University Press.
Grawe, N. D. (2021). The agile college: How institutions successfully navigate demographic changes. Johns Hopkins University Press.
Labaree, D. F. (1997). How to succeed in school without really learning: The credentials race in American education. Yale University Press.
LendingTree. (2026). U.S. student loan debt statistics. https://www.lendingtree.com/student/student-loan-debt-statistics/
Lucca, D. O., Nadauld, T., & Shen, K. (2019). Credit supply and the rise in college tuition: Evidence from the expansion in federal student aid programs. The Review of Financial Studies, 32(2), 423–466.
National Student Clearinghouse Research Center. (2025, December 4). College completion rates hold steady at recent highs. https://www.studentclearinghouse.org/news/college-completion-rates-hold-steady-at-recent-highs/
Progressive Policy Institute. (2023). How to cut administrative bloat at U.S. colleges. https://www.progressivepolicy.org/how-to-cut-administrative-bloat-at-u-s-colleges/
SoFi. (2025). How much has college tuition outpaced inflation? https://www.sofi.com/learn/content/college-tuition-inflation/
Synario. (2025). Higher education enrollment cliff: Financial perspective is everything. https://www.synario.com/resources/blog/enrollment-cliff/
The Daily Pennsylvanian. (2024, February). Penn administrator hiring far outpaces faculty growth, raising concerns of “bloat.” https://www.thedp.com/article/2024/02/penn-hires-more-administration-than-academic-faculty
The Motley Fool. (2026). Student loan debt 2025: Statistics, forgiveness, and outlook. https://www.fool.com/research/student-loan-debt-statistics/
University of the People. (2025). Institutional communications on accreditation, enrollment, and refugee education. https://www.uopeople.edu/
