Understanding the rising cost of producing higher education starts with the fact that the wealthiest, most elite colleges and universities, which set the norms in higher education, are admitted “cookie monsters,” whose insatiable and competitive need for revenue has long been viewed as beneficial — until recent decades.
Questions about higher education’s production cost — the cost of producing the education — have persisted for a century, and new modes of generating revenue and wealth to pay that rising cost have ignited sharp criticism during the last four decades as student debt has swelled. The history of these closely related developments over the past 150 years is explained in our forthcoming book, Wealth, Cost and Price in American Higher Education: A Brief History (Johns Hopkins University Press).
Annual alumni funds, national fundraising drives and endowment building emerged at the beginning of the twentieth century to help pay the production cost of higher education. But these innovations also led to higher education’s troubling wealth stratification today, which parallels the growing wealth inequality in the nation.
Tuition and academic fees are sometimes likened to prices in business, which are set by a mark-up over production cost. But the price charged to students in higher education has historically covered less than half — often merely a third — of the total production cost. The remainder has been paid by subsidies from gifts, grants, contracts, endowment income, and government appropriations.
Since 1980, the proportion of production cost paid by those appropriations has gradually decreased, impacting public colleges and universities especially. Between 2008 and 2018, federal, state, and local per-student appropriations and subsidies declined proportionally across all public four-year institutions, falling from about 69 percent to 57 percent of revenue at public doctoral universities, for example. Tuition and academic fees have risen to cover the shortfall, particularly in the public sector, and the burden of paying the production cost has shifted from taxpayers to the individual student since 1980, as economists have observed.
The steady increases in list price (the announced tuition) have fueled public resentment over the last four decades. But only wealthy students and families pay the list price, except at the least expensive colleges where many students pay the announced tuition. For the vast majority of students, the list price is reduced by grant aid, and students pay the difference (the net price) from savings, work and loans.
And the average net price plateaued between 2006 and 2020, according to College Board pricing studies. After deducting grant aid from the list price, the average net price of tuition and fees held steady over those fourteen years. In constant 2020 dollars, the annual average net price remained at about $15,900 for private, non-profit, four-year colleges and about $3,200 for in-state students at public colleges.
Why then did student debt rise to about $1.6 trillion? Different reasons apply to the various kinds of borrowers and we estimate that about two-thirds of the debt is owed by graduate students (including those in medical, law or business school) and by students who attended often-predatory, for-profit colleges and universities. Our focus is on undergraduates at non-profit colleges. A major reason for the burdensome debt of undergraduates at both public and private nonprofit colleges is neither the rising list price nor the steady net price. Rather, their ancillary expenses — food, housing, travel, technology, books, clothing, recreation and so forth — have risen with the cost of living and, since 1980, grown faster than the wages and salaries of their families. This growing shortfall over the past four decades, combined with the declining proportion of production cost paid by government subsidies, forced undergraduates at non-profit colleges to borrow more, we maintain.
While undergraduates’ net price of tuition plateaued even as their debt ballooned in recent decades, the proportion of revenue that colleges actually received from tuition increases (their net tuition revenue) declined. Through the 1960s, we explain, almost all the grant aid awarded to students comprised scholarships and fellowships funded by colleges’ revenue from gifts, grants, or endowments. Colleges’ tuition revenue virtually equaled their listed tuition through the 1960s.
Then, during the stagflation of the 1970s, a few colleges began to discount tuition — cut their list price — in order to attract specific students. Subsequently, tuition discounting accelerated, particularly at private colleges and universities, as they competed to attract the best students, diversify the student body or simply fill their seats. Net tuition revenue therefore diverged from listed tuition.
Colleges are generally reluctant to divulge how much they discount, and less than a quarter of the private non-profits report the amount, while discounting at public institutions is murky and little studied. Non-reporters obscure their discounts by folding them into “institutional grant aid,” which is reported in many datasets. By 2009, this grant aid (including scholarships and discounts) covered about 55 percent of the average list price at private, four-year colleges, which do most of the discounting and compose more than half of all four-year colleges and universities. The great majority of private colleges, lacking huge endowments and current gifts, realized diminishing returns from raising tuition because they had to increase their discounts at the same time. Therefore, the proportion of net tuition revenue available to pay their production cost decreased.
Analyzing production cost across higher education begins by distinguishing between aggregate cost and per-student cost. Growth in aggregate cost is usually salutary for the nation. Educating more citizens at a higher level fosters social mobility and improves decision-making in a democratic society, as well as generating more wealth for society by increasing people’s intellectual capital and making them more productive. However, if aggregate cost grows faster than the economy, then higher education is consuming an increasing fraction of the national income. This trend, called “cost escalation,” can become a serious problem, as with health care today.
But aggregate cost escalation may be justified if enrollment is expanding, because educating more students costs more money and generates more wealth for society in the long run. Here arises the issue of per-student cost, on which most scholars have focused over the past 70 years.
In addition, scholars have distinguished between “instructional” production cost, which contributes directly to the education of students, and production cost that does not. For example, faculty compensation is clearly an instructional cost; fitness centers may not be. This fundamental distinction grew fraught as the budgets of colleges and universities expanded, particularly during the 1950s and 1960s when massive federal grants and contracts began flooding into higher education. Is the cost of building a remote cyclotron or telescope educational or not? If so, how much? For that matter, how about the campus security force? Or university legal staff? Some scholars jettisoned the distinction. Others still apply it.
Relying on these two distinctions and the elegant economic theory of “cost disease,” an economist at Princeton University, William Bowen, published in 1968 what many scholars and policymakers came to regard as the orthodox interpretation of historical, per-student, instructional production cost. Bowen, later the president of Princeton and the Mellon Foundation, maintained that this cost had escalated inexorably since the beginning of the twentieth century and even since the start of the Industrial Revolution.
The reason, he argued, is that higher education is a service industry, and service workers — the faculty — do not realize the productivity gains of workers in manufacturing industries. In the latter kind of industry, the per-worker output increases due to technological advances, but the number of students personally taught by a professor remains constant over time in high-quality education. Because faculty compensation must grow with that of manufacturing workers, the compensation cost of faculty escalates, causing the instructional production cost to escalate. This cost-disease explanation led William Bowen to conclude that cost escalation in higher education was “inexorable” and unpreventable because it resulted from the very nature of the higher education industry, which he thus absolved of responsibility for the cost escalation.
In 1980, another economist, Howard Bowen (no relation to William) propounded a different, often misunderstood, explanation that eventually became famous. We clarify and name his view the “revenue-cost theory.”
Attending to the broad expenses of colleges and universities, Howard Bowen, formerly the president of Grinnell College and the University of Iowa, maintained that colleges and universities seek and spend all the revenue they can get. Like leaders of charitable, ecclesiastical and social service organizations, those in higher education believe altruistically that the need for revenue is insatiable because additional funds can always be put to good use. In addition, intense competition for prestige and influence drive the infinite need, Howard Bowen said. Through this strategy, he maintained, higher education deliberately increased its production cost, contrary to William Bowen’s exculpatory account.
Not only do colleges and universities, particularly the wealthiest, seek ever more revenue, they spend less and less of the increases on instruction, maintained Howard Bowen. Some of the non-instructional spending is justified by unfunded government mandates, for example. But much of the non-instructional spending reflects an “administrative expense bias” that increases the fraction of revenue spent on administration as the total revenue grows. Directors, deans, vice-presidents and their assistants multiply.
In addition, Howard Bowen showed that financial data do not support the cost-disease thesis. Drawing on large datasets that William Bowen had not, Howard Bowen revealed that per-student instructional cost declined significantly during the 1930s, 1940s, and 1970s. Our study of earlier decades reveals that the total per-student production cost scarcely escalated at all relative to the price of commodities between 1875 and 1930. In fact, this per-student cost became about three times cheaper relative to the per-capita national income over that period, largely because the number of regular students rose astronomically. Historical financial data therefore do not support William Bowen’s historical account of inexorable, unpreventable cost escalation.
Nevertheless, most scholars of higher education finance endorsed cost-disease theory and neglected Howard Bowen’s revenue-cost theory through the late 1990s. At that point, David Breneman and other prominent economists of education began to affirm that the latter was a credible alternative explanation.
Meanwhile, higher education costs escalated in the 1980s, when a study of more than 2,000 colleges found that per-student “educational and general expenditures” grew 2.5 percent annually above inflation. But gaps and inconsistencies in data sets and the growing complexity of higher education finance made it increasingly difficult to analyze production cost over time. The problem worsened in 1995, when accounting changes made it impossible to compare the costs of public and private institutions after 1995 or to analyze longitudinal trends of cost in the private sector before and after 1995, as economist Robert Toutkoushian has observed.
Notwithstanding these difficulties, it is clear that the total, per-student, production cost of higher education (excluding hospital expenses) has continued to escalate. In the first two decades of the twenty-first century, this production cost of public, two- and four-year colleges has grown, on average, about 73 percent from $14,400 to $24,900, and the production cost of private, non-profit colleges has risen about 28 percent from $36,000 to $46,000, all in constant 2020 dollars.
While production cost escalated in this way, the public attitude toward higher education, particularly its finances, shifted dramatically, we explain in our history. During the century from the 1870s to the 1970s, higher education was highly esteemed for its benefits, and spending money on higher education was generally applauded. Even the dramatic tuition hikes during the 1970s were considered justified by stagflation. Higher education was the victim, not the villain. Then, in the 1980s, the insatiable need for money began to attract criticism, which hardened into resentment over the last four decades.
To explain this resented escalation in production cost, many scholars in the early twenty-first century continued to invoke William Bowen’s cost-disease theory deferentially while they simultaneously neglected the theory in their analyses. At the same time, paradoxically, scholars increasingly adopted tenets of the revenue-cost theory of Howard Bowen, whom they often did not cite.
This paradox occurred, in part, because William Bowen’s, original, 1968 study had significant flaws, which were identified only in 2016. Conversely, some economists did not consider Howard Bowen’s explanation a legitimate theory because it attributed cost growth to intentions and motivations of higher education leaders that could not be tested empirically, it was said.
The paradox was demonstrated in a celebrated book by the distinguished Cornell University economist Ronald Ehrenberg, who wrote that higher education leaders must and do act like “cookie monsters.” In pursuit of “maximizing value,” higher education leaders “seek out all the resources that they can get their hands on and then devour them.” Every selective college and university is therefore “engaged in the equivalent of an arms race of spending to improve its absolute quality and . . . its relative stature in the prestige pecking order.”
This analysis fits closely the revenue-cost theory of Howard Bowen, whom Ehrenberg never cited while invoking William Bowen extensively without actually relying on cost-disease theory. Our history validates this analysis by revealing the late nineteenth-century origin and subsequent proliferation of this strategy and behavior that Howard Bowen identified in 1980. How much is spent and how it is spent are still perplexing questions about production cost. But a consensus on why it grows has therefore emerged in the early twenty-first century.
We conclude that leaders of higher education, particularly the wealthiest elite, are currently perseverating in behavior and strategy that yielded enormous, widely recognized benefits for a century. But this now counter-productive strategy has contributed to the regrettable shift from esteem to resentment for higher education in recent decades.