Editors’ Note: Bruce Kimball casts the light of historical analysis on the two major theories explaining cost escalation in U.S. higher education. His post draws upon the following sources: Bruce A. Kimball and Jeremy B. Luke, “Historical Dimensions of the “Cost Disease” in U.S. Higher Education, 1870s–2010s,” Social Science History 42 (2018): 29-55; Bruce A. Kimball and Jeremy B. Luke, “Measuring cost escalation in the formative era of U.S. higher education, 1875–1930,” Historical Methods: A Journal of Quantitative and Interdisciplinary History 49 (2016): 198-219; and Bruce A. Kimball, “The Rising Cost of Higher Education: Charles Eliot’s “Free Money” Strategy and the Beginning of Howard Bowen’s “Revenue Theory of Cost,” 1869−1979,” Journal of Higher Education 85 (2014): 886-912.
Since the 1960s, the escalating cost of higher education in the United States has attracted intense scrutiny from many policy-makers and scholars. Economists, in particular, have proposed several models to explain the phenomenon, but two economic theories have predominated over the last fifty years. In 1968, economist William Bowen advanced the “cost disease” theory, and many distinguished economists have endorsed this model based on analyzing large quantitative data sets. In 1980 economist Howard Bowen (not related to William) proposed the “revenue theory of cost,” also based on study of large quantitative data sets. Many scholars and leaders of higher education, including a few economists, subsequently embraced this view. This post shows how new historical research on the finances of U.S. higher education informs the debate over the contested validity of these economic theories.
In this discussion, cost—sometimes called the “production cost” by economists—refers not to what students or parents pay but to what colleges and universities spend to produce the education. Cost escalation means that the expense of higher education is growing faster than the cost of living or the national income. When aggregate cost escalates, higher education consumes an increasing fraction of national income, and this trend becomes a serious problem over time, as with the skyrocketing cost of health care. Nevertheless, aggregate cost escalation may be justified, even salutary, if the population of students is expanding.
The more important issue is per-student cost. If that rises persistently, as during most of the last century, then the trend becomes very worrisome. Yet, even this trend might be justified by the improving quality of education, by the need for remediation of entering students, or by the increasing cost of goods and services to obtain the same output from a college or university. Nevertheless, the escalation of per-student cost is generally taken to mean that higher education is becoming less efficient and consuming national income at a faster rate than the number of students is growing.
Over the long period between 1929 and 1981, educational expenditures per student increased at an average annual rate of about 1.7 percent in constant dollars. That rate then accelerated to about 2.75 percent until 1995, when it slowed, but the cost continued to grow through the Great Recession of 2008-09 and thereafter. These rates may seem small, but a growth rate of merely 1.4 percent means that the per-student cost in constant dollars would double every fifty years. Hence the question persists: why does the cost of higher education keep escalating?
Two Economic Theories
Over the past five decades, economists have considered several theories to explain the phenomenon, but two economic models predominate. Following the lead of economist William Bowen in 1968, who later became the President of Princeton University and then of the Mellon Foundation, many distinguished economists have endorsed “cost disease” (CD) theory.
This model begins with a fundamental distinction between manufacturing or goods-producing industries and personal-services industries. That distinction has important implications for economic productivity, according to CD theory. Over time, goods-producing industries have greatly increased their productivity, or output per worker-hour, due to advances in labor-saving technology. It takes fewer workers and fewer hours to build a house than in the past, for example. In contrast, personal-services industries cannot increase their productivity in this way. One professor still teaches a class of twenty-five students, notwithstanding improvements in information technology. Hence, personal-services industries are among “stagnant industries,” which do not increase their productivity.
In order for the living standard of service-providers to keep pace with others in society, the compensation cost of service providers must therefore grow faster for the same output per labor-hour than that of workers in goods-producing industries. Faculty compensation per student will thus grow faster, driving up the overall cost of higher education relative to goods-producing industries. Furthermore, “there is nothing in the nature of this situation to prevent educational cost per unit of product from rising indefinitely at a compound rate of this sort,” emphasized William Bowen in his canonical 1968 study.
Since that time, Bowen and other distinguished economists repeatedly affirmed and claimed to validate CD theory as the explanation for cost escalation in higher education. These repeated affirmations have conferred great authority on the theory, but recent research has identified fundamental problems in attempts to prove it. In fact, cost disease scholarship provides little validation that the theory explains cost escalation in higher education over the last century, although it may explain certain periods within that time span.
As an alternative, in 1980, Howard Bowen, who served as president of Grinnell College, the University of Iowa, and Claremont Graduate University, proposed what he called “the revenue theory of cost” (RC). Based on his study of aggregate quantitative data over the period from 1929 to 1979, Bowen maintained that the cost of operating colleges or universities is set primarily by the revenue available rather than the requirements of their operations. Colleges and universities never give money back and never stop asking for more, he observed. Fundraising, grant-seeking, and tuition-increases are ceaseless, and identifying new needs follows in order to justify them.
RC theory thus reversed the conventional assumption that the “cost of higher education…is determined by the needs of institutions,” Howard Bowen maintained. Consequently, the RC theory has been summarized by John Thelin and Richard Trollinger in the aphorism: “Colleges raise all they can—and spend all they raise.”[i]
This theory rapidly became the predominant explanation for cost escalation among scholars and leaders of higher education, including some economists. In 2008 Dutch scholar Franz van Vught maintained that “Bowen’s famous law of higher education still holds” for universities throughout the world.[ii]
Many distinguished economists, however, challenged RC theory, asserting that the model is tautological or circular or inverts cause and effect. Some objected that Howard Bowen’s explanation is not really an economic theory at all, because it addresses only higher education and does not identify an objective factor that can be tested and validated. A few economists consider RC theory to be a description of the incentives or motivations of higher education institutions and leaders, rather than an explanation for their behavior. In fact, Howard Bowen himself did not validate the theory or even attempt to do so, and no subsequent scholar has actually proven RC theory.
Consequently, over the past fifty years, economists have not validated either theory as an explanation for cost escalation in higher education, though they have attempted to do so for cost disease theory and claimed they were successful, while generally dismissing the possibility of doing so for revenue theory.
The important purpose of developing and testing theory is, of course, to explain cost escalation by identifying its cause in the hope of developing policy to stem it. Complicating that purpose is the normative issue of responsibility: who, if anyone, is responsible? This issue has generally been sublimated because economists have preferred the impersonal, technical explanation of CD theory and dismissed the normative issues entailed in RC theory. Nevertheless, the normative implications exist and magnify the significance of each theory and the importance of determining each one’s validity.
Cost disease theory entails three normative implications. First, the theory exculpates higher education institutions and their leaders. Personal-service industries cannot prevent their cost from rising, according to the theory. Their stagnant productivity stems from the intrinsic, labor-intensive nature of teaching. “Though it is always tempting to seek some villain to explain such cumulative cost increases, there is no guilty party here….the cost increases are not caused by criminal neglect, incompetence, or greed, but rather…the essentially irreducible quantity of labor entailed,” stated cost-disease economist William Baumol in 2012. Hence, if CD theory holds, then higher education is not responsible for its rising costs.
Second, CD theory implicitly identifies a culprit, because it posits that faculty compensation is the root cause of cost escalation in higher education. That proposition also magnifies the significance of validating CD theory. Fingering the compensation of faculty as the problem, even if they are blameless, suggests that faculty have little basis for complaint about their salaries.
Third, CD theory may justify the steady growth in hiring contingent or adjunct faculty in higher education, who are paid relatively little for teaching. This “adjunctification” improves the “productivity ratio” of higher education. This ratio of outputs/inputs measures the productivity of an industry, and economists, including William Bowen in 2012, have argued for “lowering the denominator of the productivity ratio” through “determined efforts to reduce costs.” Since, according to CD theory, the main driver of cost is faculty compensation and since the growing reliance on adjuncts slows the growth of overall faculty compensation and thereby raises the productivity ratio of higher education, the validation of CD theory would potentially justify the increasing reliance on adjunct faculty.
Cost disease scholarship therefore seems entangled in a curious paradox regarding normativity and validity. The theory holds higher education blameless for cost escalation, so society cannot complain. It identifies rising faculty compensation as the root cause, so faculty cannot complain. And it appears to justify the much-criticized adjunctification of the faculty. Yet, cost disease scholars have not studied specifically whether increases in faculty compensation could account for growth in the cost of higher education.
Unlike cost disease economists, Howard Bowen explicitly announced the normative dimensions entailed in RC theory. Instead of absolving colleges and universities of responsibility, he considered them culpable. Though citing the quest for “educational excellence,” he gave equal weight to institutional “vanity,” “competition,” or “prestige” as drivers of revenue-driven spending. In addition, he maintained that high-minded goals like excellence actually mask pecuniary aims by conferring on a college or university “a sort of hunting license which enables it to gather funds wherever it can find them and to obtain the maximum amount possible.”
Acquiring “the maximum amount possible” then leads colleges and universities to “waste” money on “less important purposes,” such as building “a more opulent physical plant” or providing “overlays of administrators,” according to Howard Bowen. In addition, he maintained that “the personal ambitions of the administrators and professional workers” often fuel institutional aspirations.
But some economists in high administrative positions at eminent universities—such as Ronald Ehrenberg of Cornell University, Charles Clotfelter of Duke University, and William Massy of Stanford University—have discounted Howard Bowen’s normative claims.[iii] They have asserted that cost escalates due to the pursuit of “quality” and “excellence” and “value,” not self-serving motives. But these individuals also recognize that every university is “engaged in the equivalent of an arms race of spending to improve its absolute quality and to try to improve its relative stature in the prestige pecking order,” in the words of Ehrenberg. As a result, in higher education, “administrators are like cookie monsters searching for cookies. They seek out all the resources that they can get their hands on and then devour them.” But even if this behavior aims at “quality,” “excellence,” or “value,” Howard Bowen’s thesis remains true that the pursuit is endless, the identified needs are innumerable, and cost escalation is not driven by necessity.
How Historical Research Informs the Validation
All these normative implications magnify the importance of validating the two models, and this effort is informed by new historical research on the finances of U.S. higher education during its formative period from the 1870s to the 1920s. Over these five decades, the U.S. economy expanded faster than any national economy in history due to immense productivity gains in manufacturing industries resulting from technological advances wrought by the Industrial Revolution. During such a period of staggering growth in manufacturing productivity, cost disease theory predicts that the cost of higher education, a personal-services industry, would experience sharply “rising costs” compared to the cost of living and to the national income.
Financial data in the annual reports of the U.S. Commissioner of Education between 1875 and 1930 reveal that the total capital (including property and endowment) and the total annual expense of a cross section of higher education institutions grew much faster than the national income (GNP) in constant dollars. The total capital grew nearly twenty times and the total annual expense more than forty times, while GNP rose only about seven times. Hence, cost escalation occurred in aggregate terms, as the nation made a stupendous aggregate investment in higher education between 1875 and 1930.
But that escalation of aggregate cost can be explained and justified by the expansion of higher education. Between 1875 and 1930, the number of colleges and universities in the nation more than doubled from 591 to 1,432, and the number of their “regular” students enrolled during the academic year rose twelve-fold from about 90,100 to about 1,086,000. Given this, the more salient issue for assessing the two economic theories is the per-capita analysis.
The financial data in those annual reports show that the per-student cost of higher education in a cross section of colleges and universities grew very slowly over this period, scarcely faster than the price of all commodities, and much more slowly than per-capita income. Over the 55-year period, the per-student cost grew cumulatively by 27 percent, while commodity prices rose by 17 percent and per-capita income by 332 percent. Higher education became slightly more expensive in terms of commodity prices and much more affordable compared to per-capita income.
This finding is not consistent with the prediction of cost disease theory. It is, however, consistent with revenue theory because Howard Bowen did not claim that cost escalation or the institutional behavior posited by his theory occurred prior to 1929, the starting-point of his quantitative data set. Yet, if the cost-per-student began to escalate only in the late 1920s and if RC theory is correct, then a set of historical questions presents themselves: when and how did the institutional behavior posited by RC theory originate? Why and how did that behavior proliferate in higher education by 1929 when it began to be manifested in the data that Howard Bowen studied?
My historical research suggests answers to these questions. The behavior began with a new financial strategy that Harvard President Charles W. Eliot formulated during his administration from 1869 to 1909. Eliot’s distinctive strategy focused on attracting, saving, and investing gifts in order to increase the financial capital of Harvard, particularly its “free money” or unrestricted endowment. As he wrote in 1906: “In the competition between American universities, and between American and foreign universities, those universities will inevitably win which have the largest amounts of free money.”
Eliot elaborated his “free money” strategy with a set of precepts that closely approximate the tenets of Howard Bowen’s revenue theory of cost. First, the universities were competing for academic achievement and reputation, and the outcome of the competition would be determined by their financial resources. Second, Eliot held that a university must conserve its wealth and always seek new gifts for buildings and new programs. Third, a university should always run a deficit because this provides the justification for asking for more free money. By implication, therefore, a university’s needs are insatiable. In sum, the competition for academic excellence and prestige was a struggle to accumulate wealth. This was his bedrock principle. “If the primacy of Harvard University among American institutions of education is to be maintained, it must not be surpassed by any other in material resources,” he wrote in 1896. Here was the core financial strategy of higher education institutions that Howard Bowen identified nearly a century later.
If Eliot’s strategy was the origin of the behavior codified by the RC theory, why and how did his singular approach become widespread? One reason was that Eliot trumpeted the strategy in his thirty-nine published annual reports, which were frequently quoted in the newspapers and read by college and university presidents, whom Eliot wished to educate on university management.
Another reason was that the strategy worked, and Harvard gradually established itself as the nation’s wealthiest university soon after Eliot’s administration. At various points between 1868 and 1915, Cornell, Johns Hopkins, Stanford, and Columbia possessed the largest endowment among universities. But in 1920 Harvard’s invested financial capital, for the first time, assumed the lead among university endowments that it would never relinquish to the present day. Furthermore, Harvard attained this rank even though each of the other eight wealthiest universities in the nation had received a much larger gift than any single gift that had come to Harvard prior to 1920. Eliot’s free money strategy, not the biggest gifts, made the difference.
A third reason for the proliferation of Eliot’s strategy was the inception of nationally organized, cyclical fundraising. In 1914 Harvard alumni began planning the first nationally organized fundraising campaign conducted by a university, which concluded in 1921. The seven-year national effort attracted widespread publicity and many inquiries from other colleges and universities. Throughout the campaign, the Harvard alumni embraced the financial teachings of Eliot, “whose slightest word they regard as law,” wrote the campaign chairman, Thomas Lamont.
Eliot’s free money strategy proliferated throughout higher education as other colleges and universities initiated their own fundraising campaigns, emulating Harvard. Already in February 1920, some 75 colleges and universities were “following in your wake,” as Princeton President J. G. Hibben wrote to Harvard President A. L. Lowell at the outset of Princeton’s first Endowment Fund campaign. During the ensuing decade, some 500 more colleges and universities followed Harvard in running fundraising campaigns for the first time.
As the endless cycle of such campaigns started rolling, Eliot’s free money strategy became the common sense of how colleges and universities should compete for academic eminence by building wealth. The financial behavior later codified in the revenue theory of cost thus became widespread by 1929 along with the escalation of per-student cost in higher education, although Howard Bowen apparently did not know the historical background.
The Significance of Historical Context
What can be said, in conclusion, about the normative implications of the two economic theories? There seems to be little validation—and significant contrary evidence—for cost disease theory as an explanation of cost escalation in higher education between the 1870s and the 2000s, although the theory may explain cost escalation during certain periods within that time span. This conclusion puts in doubt the exculpation of higher education institutions and leaders, the responsibility of faculty compensation for cost escalation, and the potential justification of the adjunctification of higher education that are implied by CD theory.
What about the revenue theory of cost, which is consistent with the recent historical research, discussed above? Contrary to CD theory, RC theory seems to imply that higher education institutions and leaders are “the guilty party” who, driven by “vanity,” “competition,” and “prestige,” are always seeking more money and deserve blame for cost escalation. Or, at the very least, they are like “cookie monsters” with insatiable appetites, who cannot help themselves in the pursuit of “excellence” and “value.” Does this apply to Charles Eliot, his contemporaries, and his successors?
The normative implications must be interpreted in light of the historical context and may vary over time. There is a striking contrast between Eliot’s earnest and public promotion of his strategy and Howard Bowen’s references to institutional vanity and self-interest, generally covert or sublimated. That contrast may arise from the very different historical circumstances in which they developed their ideas.
Eliot served as a university president during one of the longest and greatest economic expansions in the history of the world, fueled by growth in intellectual capital manifested in the establishment of school systems, universities, and junior colleges. In contrast, Howard Bowen in the late 1970s undertook his study in the middle of an extremely difficult economic period “of stagnation complicated by rapid inflation” and by “the impending decline in the number of eighteen-year-olds.” As a result, policies that Eliot publicly and justifiably advocated to develop the fledgling higher education system within an expanding economy were construed in the 1970s by Howard Bowen as selfish grasping for a larger piece of a shrinking pie during a period of economic “stagflation.”
Ultimately, the responsibility implied in economic explanations for cost escalation depends on the historical situation. Consequently, understanding the historical origins and development of the behavior explained by economic theories not only sheds light on the validity of such theories but also helps us to understand and interpret their normative implications.
Bruce Kimball is a professor in the Philosophy & History of Education Program at Ohio State University and a former Guggenheim Fellow.
[I] J. R. Thelin, J. R. and R. W. Trollinger, “Forever is a Long Time.” History of Intellectual Culture, 9 (2011): 1-12.
[ii] F. van Vught, “Mission Diversity and Reputation in Higher Education.” Higher Education Policy 21 (2008): 151-174.
[iii] See R.G. Ehrenberg, Tuition rising: Why college costs so much (2000, 2002); C.L. Clotfelter, Buying the Best: Cost Escalation in Elite Higher Education (1996); W.F. Massy, Honoring the Trust: Quality and Cost Containment in Higher Education (2003).