Part 1

Costs and Productivity in Higher Education

AS MY WIFE keeps reminding me, I have a Don Quixote–like tendency to flail away at windmills—to take on topics such as race in America and affirmative action; the insidious problems with college sports at all levels, including Division III and the Ivy League (which cause me to cringe whenever the NCAA refers to its legions of “student-athletes”); and, yes, the un forgiving economics of labor-intensive industries, such as the performing arts and higher education. But, my DNA is what it is, and so I am now adding to this list the potential implications of online learning for college costs.

Context matters, and I will begin by outlining as succinctly as I can aspects of the economics of higher education that are relevant to my topic:

•   trends in costs, the “cost disease,” and how to think about changes in productivity;

•   other forces, some deeply ingrained in the fabric of higher education, that also push up costs; and

•   growing worries about affordability, especially in the public sector, where reductions in public support have been coupled with significant increases in tuition.

Then, in the second part of the book, I will discuss what I think—or, better said, what I suspect—about the potential impact of the variety of approaches to online learning that are everywhere present, including, of course, at Stanford University and at Stanford spin-offs such as Coursera and Udacity. Is there, as President Hennessy has suggested, a tsunami of some still ill-defined kind coming? Is it realistic to imagine that online learning is a “fix” (at least in part) for the cost disease? Throughout, I will maintain a system-wide perspective, since it will not do to think about these large questions solely from the perspective of individual institutions.

Cost Trends, the “Cost Disease,” and Productivity in Higher Education

It is fitting that I gave these Tanner Lectures in close proximity to Clark Kerr’s neighborhood, since it was President Kerr, in his capacity as chairman of the Carnegie Commission on the Future of Higher Education, who commissioned a study of mine in the mid-1960s that became The Economics of the Major Private Universities. In that study I documented the seemingly inexorable tendency for institutional cost per student (which is, of course, different from tuition charges) to rise faster than costs in general over the long term. Kerr christened this finding Bowen’s Law, although he was, he said, “originally skeptical about it.”¹

What is important today is not the exact numbers contained in that study (which were based largely on a detailed examination of the experiences of the University of Chicago, Princeton University, and Vanderbilt University between 1905 and 1966) but the underlying pattern, which has been found to hold for public as well as private universities, and for colleges too. I reproduce here, as something of a historical relic, a figure from my 1960s Carnegie study (figure 1). The figure shows that, excepting war periods and the Great Depression, which require separate analysis, cost per student rose appreciably faster than an economy-wide index of costs in general. The consistency of this pattern suggested to me then, as it does today, that we are observing the effects of relationships that are deeply embedded in the economic order.

Figure 1 Direct costs per student, compared with an economy-wide cost index

Running through all the factors at play (and there are many, as I will indicate shortly) is a key proposition that my teacher and lifelong friend, William J. Baumol, and I first articulated in our study of the performing arts, which also dates from the mid-1960s.² The proposition is known to this day in the literature as the “cost disease.” The basic idea is simple: in labor-intensive industries such as the performing arts and education, there is less opportunity than in other sectors to increase productivity by, for example, substituting capital for labor. Yet markets dictate that, over time, wages for comparably qualified individuals have to increase at roughly the same rate in all industries. As a result, unit labor costs must be expected to rise faster in the performing arts and education than in the economy overall.

Robert Frank of Cornell University provided this succinct explanation of the cost disease as recently as March 2012: “While productivity gains have made it possible to assemble cars with only a tiny fraction of the labor that was once required, it still takes four musicians nine minutes to perform Beethoven’s String Quartet No. 4 in C minor, just as it did in the 19th century.”³ In short, productivity gains are unlikely to offset wage increases to anything like the same extent in the arts or education as in manufacturing; hence, differential rates of increase in costs are to be expected—a finding Baumol and I reported for major orchestras at about the same time that my Carnegie study of higher education was under way.⁴

About a decade after the Carnegie study, I reported a similar pattern in my 1976 President’s Report at Princeton: “While prices in general have risen about 50% [over the previous 10 years alone], the most widely used price index for higher education has risen about 70%.”⁵ And in 2012, three and a half decades later, Sandy Baum, Charles Kurose, and Michael S. McPherson reported basically the same pattern. In their paper “An Overview of Higher Education,” presented at Prince ton University, they cite a careful study using data from the Delta Cost Project that shows that “educational expenditures per FTE student increased at an average annual rate of about 1% beyond inflation at all types of public institutions from 2002 to 2008.”⁶ There is no need to burden this argument with more data about trends in institutional costs, which are notoriously hard to interpret, in part because they often involve aggregations of various kinds. It is easy to get mired in the underbrush, and we do well to remember the admonition of the architect Robert Venturi: “Don’t let details wag the dog.”

There is, however, a final big point to note about trends—namely, the reversal that has occurred in the last decade or so in the respective positions of private and public institutions. When I wrote my 1976 report, from the perspective of the president of a private university, there was widespread concern about the widening gap in charges between the privates and the publics (with the privates becoming ever more expensive relative to the publics). In those years, the privates were hit especially hard by the stagflation of the time, with its dampening effect on stock market values that, in turn, affected both returns on endowments and private giving. Today, it is the publics that have suffered more than most of the privates (and certainly more than the most selective privates), largely as a result of sharp cutbacks in state appropriations.

During a discussion session the day after I originally made these remarks at Stanford, President Hennessy contrasted trends in tuition and student aid in the public sector with the recent experience at Stanford. He observed that while Stanford’s “sticker price” has continued to increase, as it has throughout almost all of higher education, Stanford has had the financial wherewithal to increase its outlays on student aid by even more than the increases in its tuition and has chosen to spend some part of its resources in this highly commendable way. Only a small number of other wealthy private institutions have been able to do the same thing, and the fortunate circumstances of these relatively well-off “outlier” institutions should not be allowed to obscure the general pattern pertinent to the public colleges and universities that educate three-quarters of this country’s college students⁷—or, for that matter, the trends pertinent to the large number of private colleges and universities that have also been compelled to raise tuition faster than they have been able to raise student aid. I will return in due course to the broad subject of increasing stratification in higher education and its implications.

I am aware that thus far I have been using an important word—productivity—without defining it. Put simply, productivity is the ratio of outputs to the inputs used to produce them. But this formulation conceals at least as much as it reveals, since it is maddeningly difficult in the field of education to measure both outputs and inputs—even within a single institution, never mind across institutions serving different missions. If only we produced standardized widgets or harvested blueberries!

As one illustration of how treacherous this terrain is, the National Academy of Sciences released, in 2012, a massive report of over two hundred pages devoted to the measurement of productivity in higher education. A major virtue of the report, which in turn cites a voluminous literature, is that it debunks the idea that productivity in higher education is unidimensional. It warns against a multiplicity of dangers that lurk behind the use and misuse of (inevitably) simplified measures. The report insists that “quality should always be a core part of productivity conversations, even when it cannot be fully captured by the metrics.”⁸ It also emphasizes the complications stemming from joint production of outputs such as teaching and research, and the need to recognize a complex mix of inputs, including capital and student time.

In thinking about the implications of these myriad complications for the ways in which technology might impact the cost disease, I have been helped greatly by the authors of an article in the New England Journal of Medicine (NEJM), who have captured quite skillfully factors that explain what is known as the IT productivity paradox—the apparent tendency, noted by Robert Solow of MIT in 1987, for computerization to fail to improve standard measures of productivity. Solow noted famously, “You can see the computer age everywhere but in the productivity statistics,” an observation said to have launched more than two decades of research into the sources of the paradox.⁹

The authors of the NEJM article argue that explanations for the IT productivity paradox fall into various categories. Under the heading of “mismeasurement,” they note that “important dimensions of service output such as accessibility and convenience—factors that are greatly improved by IT—are difficult to quantify and are rarely captured by productivity metrics.”¹⁰ For example, ATMs increased consumer convenience in banking, but this increase in convenience, and all the time saved by customers, was not captured by traditional measures of productivity.

The authors go on to point out: “In terms of ‘mismanagement,’ the introduction of new technologies usually forces reexamination of the assumptions that underpin less productive processes.” They give a telling example concerning the introduction of electricity in manufacturing: early on, “factories simply swapped large electronic motors for waterwheels and steam engines but retained inefficient belt-and-pulley systems to transmit power from the central power source. Real productivity gains came only after manufacturers realized that many small motors distributed throughout a factory could generate power where and when it was needed.”¹¹

This discussion in the NEJM, aimed at implications for the health industry, resonates with the uses of IT in education. It is easy to think of examples, including the tendency in the early days of online teaching simply to mimic typical classroom teaching methods, often by videotaping lectures, rather than re-engineering the teaching process as a whole.

From the standpoint of our interest in the cost disease, it is critical to keep in mind that the productivity ratio has both a numerator and a denominator. Productivity improvements can be either output-enhancing (raising the numerator) or input-conserving (lowering the denominator). It seems evident that information technology has been extremely consequential in higher education over the last twenty-five years, but principally in output-enhancing ways that do not show up in the usual measures of either productivity or cost per student. It is important to distinguish between at least two broad types of educational “output”: research findings and student learning outcomes. We should also recognize that there is a consumption component in the output numerator. The veritable revolution in information technology has had an especially large impact on research output. Data management systems and powerful number-crunching capacities have permitted research that would have been simply impossible otherwise. Work in particle physics and studies of the human genome are but two examples from the physical and life sciences. To cite a much more mundane example from the social sciences, the work that Derek Bok and I did on the effects of race-sensitive admissions would have been impossible without the construction of the large College and Beyond database.¹² More generally, advances in communications, and the development of networks and systems for managing text and exchanging perspectives with colleagues at a distance, have revolutionized the way papers are prepared and revised—again and again! Yet these innovations do not show up at all in the usual measures of output.

Technology has also led to dramatic improvements in the scholarly infrastructure. If I may again cite activities that I know well, the creation of JSTOR (a highly searchable electronic database of scholarly literature) has changed fundamentally the way scholars use the back issues of journals and has had profound effects on libraries. Similarly, ARTstor (a digital repository of high-quality images) now permits art historians to study, for example, images of a Bodhisattva on the wall of a cave in Dunhuang, an oasis town on the Silk Road, alongside images of the same Bodhisattva on a silk painting at the Guimet Museum in Paris.¹³ It is worth emphasizing that these benefits generally do not accrue to the institutions that made the invest ments necessary to realize them. For example, the extra ordinary time savings for scholars made possible by both JSTOR and ARTstor do not prompt the institutions that employ the scholars to harvest these savings by, for example, increasing teaching loads (unimaginable!).

Although faculty and students have certainly benefited in many ways from easy Internet access, relatively little has happened with respect to classroom teaching—until quite recently. In the second part of this book, I will suggest that we are only at the beginning of the kind of re-engineering that could in time transform important parts—but only parts—of how we teach and how students learn. Most fundamentally, I will argue that we need to improve productivity in two ways: (1) through determined efforts to reduce costs—that is, we need to focus more energy on lowering the denominator of the productivity ratio; and (2) through new ways of increasing the student-learning component of the numerator of the ratio, principally by raising completion rates and lowering time-to-degree.

Factors Other Than the Cost Disease Pushing Up Educational Costs

As important as I believe the cost disease to have been (and to be) in putting upward pressure on instructional costs, I certainly do not think that it is the sole villain. Let me now mention ever so briefly three other forces behind the rise in costs. I recognize, of course, that this list is by no means comprehensive.¹⁴

Inefficiencies

I am not one of those who looks with disdain at how poorly managed colleges and universities are often alleged to be. (I have seen too much of other organizations in all sectors of the economy, including the for-profit sector.) It is at least mildly annoying when, with no attempt to provide evidence, a business publication such as Forbes blithely asserts on its cover (of November 19, 2012) that “no field operates more inefficiently than education.”¹⁵ I wonder if that is really true. Perhaps the time has come—if it is not past—when we should cease making such sweeping pronouncements (recognizing that colleges and universities in general have had much greater staying power than many for-profit businesses, which have been seen to fall by the wayside in surprisingly large numbers).

Still, it is hardly surprising that the severe financial pressures of our time have led to renewed calls for more business-like approaches. One consulting firm, Bain & Company, has found that universities such as the University of California, Berkeley, the University of North Carolina, Chapel Hill, and Cornell University are complex, decentralized institutions that could save money by simplifying oversight structures and centralizing functions such as human resources, information technology, and purchasing.¹⁶ In my view, just as it is wrong for business-oriented writers to assert that inefficiency is rampant, it is also unwise for academics to dismiss studies such as these out of hand as contributing to a “corporate mindset” in higher education. Universities do have to become more business-like in relevant respects at the same time that they have to retain their basic commitments to academic values.

I would, however, caution against too facile an acceptance of the assumption that when universities devote increasing shares of their resources to non-instructional infrastructure costs, they are demonstrating inefficiency. That may or may not be true. In some instances, regulatory requirements compel universities to spend more on compliance. There are also situations in which it makes excellent business sense to substitute lower-cost inputs for faculty time spent on mundane tasks that others can do at least as well. Of course, none of this is to deny that wasteful “bureaucratic creep” can occur.¹⁷

It is also true that educational institutions are good at adding things but not good at subtraction. Fixed costs are often truly fixed (such as the costs associated with cutting-edge scientific laboratories in narrowly defined fields). Moreover, universities are collections of highly specialized talents that cannot be readily shifted from, say, teaching Russian to teaching Spanish. Institutional rigidities are facts of life that in many, though hardly all, cases derive from the very nature of the academic enterprise. It is harder, however, to defend antiquated organizational structures such as “centers” of one kind or another, which are notoriously difficult to dismantle even when they have ceased serving their purposes. A good rule of the road is to use flexible structures such as workshops or experimental colleges that do not take on lives of their own.¹⁸

Still more controversial aspects of alleged inefficiencies on the academic side of the house are the scope of program offerings, the use of cross-subsidization to support lowenrollment programs, and the reluctance to use differential tuition pricing to ration costly offerings and encourage students to go into less costly areas. The value propositions at issue are vigorously contested, and I can do no more here than recognize the importance of this debate.¹⁹

An Ingrained Desire to “Buy the Best”

Institutional proclivities are a powerful factor of a very different kind. Charles Clotfelter, in his detailed case studies of costs at elite universities, found that there was a determination to spend whatever it took to excel.²⁰ There is, indeed, a deep-seated commitment to enhancing institutional reputation. Given this mindset, the availability of resources is a strong driver of costs.²¹ Lawrence S. Bacow, former chancellor of the Massachusetts Institute of Technology (and former president of Tufts), has said that at MIT, “the mentality was to do what we needed to do to make sure our students mastered the material, regardless of cost.… We looked to reduce class size, increase teacher-student contact, do more hands on learning, etc. All of these drive costs up and productivity down.”²² Moreover, faculty and students often collaborate to create inefficiencies. An example is Friday classes, which neither students nor faculty want; it is very difficult for presidents to prevent the demise of these classes when students and faculty agree on such an objective.

Competitive juices are everywhere evident, and I confess that I am conflicted in how I feel about this undeniable source of upward pressure on costs. In company with other economists, I believe that competition to “be the best” drives up quality and is basically a good thing. The competitive (entrepreneurial) nature of American higher education stands in sharp contrast to what one often finds elsewhere and is, I believe, a key reason why many American research universities are the best in the world. In recent years, however, I have come to wonder whether, in some situations, there can be too much competition for the societal good. We have seen more and more stratification within higher education, with the wealthiest institutions distancing themselves from other very good, but not so wealthy, places.²³ I believe this combination of increased stratification and a determination to buy the best can have some pernicious effects. For instance, wealthy institutions routinely make huge investments in the start-up costs of faculty hires in the sciences. This puts great pressure on other places that think of themselves as peers to match such outlays, even if they have to divert funds from needy fields such as the humanities.

I worry, too, that the financial aid policies of wealthy institutions apply too much de facto pressure on other institutions to be extremely generous, thereby encouraging “quasi-merit-aid-wars” of dubious societal value. Students and their families complicate all of this by applying pressure of their own for more and more amenities (elaborate student centers and fitness facilities, dormitories that sometimes have features that 99 percent of the population can’t enjoy, and so on). Institutions feel that they have to satisfy the desires of fullpaying affluent families who (not surprisingly) want more and more of everything, including customization.²⁴ This is hardly surprising in a society in which it is now possible to order highly customized clothing by clicking online. But, of course, the multiplication of choices is expensive. Still another complicating (I would say “aggravating”) factor is the U.S. News rankings, which encourage institutions to put too much weight on maximizing their yields and keeping up their average SAT scores even as more and more evidence casts doubt on the predictive value of these scores.²⁵

There is a conundrum here. Institutions have an un derstandable interest in always improving themselves, even if the pursuit of immediate institutional self-interest cuts against larger societal interests. Still, the most privileged places should think hard about the ramifications of their actions. When I spoke at the installation of Morton O. Schapiro as president of Williams College, I used a quotation from the Midrash Tanhuma: “The rich should ever bear in mind that his wealth may merely have been deposited with him to be a steward over it, or to test what use he will make of his possessions.”²⁶

There is a stewardship responsibility. Moreover, American colleges and universities are so fiercely competitive that consideration has to be given to benign forms of collusion and even some regulation. Reluctant as I am even to mention the NCAA in any kind of quasi-favorable light, we should acknowledge that there is value in obligatory academic requirements (minimal as they are) for participation in intercollegiate sports.²⁷ I also think that some years ago the U.S. Department of Justice did us all a disservice in applying simplistic notions of antitrust regulation to well-designed efforts to ensure that limited financial aid resources were in fact distributed on the basis of agreed notions of financial need.

As another example, President Bacow has suggested that universities might consider limiting tenure to some number of years. The objective would be to combat the costly and sometimes corrosive effects of the end to mandatory retirement.²⁸ I should confess that when I was a beginning graduate student I was one of those who objected to mandatory retirement, which was legal at the time. I was in the last class that the brilliant economist Jacob Viner ever taught at Princeton on the history of economic thought. Summoning up all my courage, I went into Professor Viner’s office to complain about his impending retirement. Viner gave me one of his most piercing looks and said, with a twinkle in his eye: “Mr. Bowen, most of what you say is true. I am at the peak of my powers, smarter than all of my colleagues, and it would be a shame if future Princeton students were deprived of the opportunity to learn from me. But,” he added, “your conclusion is wrong. I should be forced to retire. I’ll tell you why. My colleagues are good and compassionate people, and they will never distinguish me from all of the other faculty members who should have retired years ago! Either all of us go, or none of us goes. It is much better that all of us go.” Here is the end of the story: Professor Viner did have to retire from Princeton, but he went on to teach at leading universities all over the world until his death.

There is a place for well-considered rules, especially when they allow markets to work (as in Viner’s case). More generally, I believe that there is a need for a thoughtful study of situations in which some collusion is a good thing, as well as situations in which collusion is injurious.

Back to the implications of the relentless pursuit of reputation. One specific problem—a definite source of upward pressure on costs that I attribute in no small degree to status wars—is the proliferation and at times excessive support of graduate programs of middling status in fields such as physics. Neil Rudenstine and I discussed this problem at length in a book we wrote some years ago (In Pursuit of the PhD), and there is no evidence that it has done anything but become more serious since.²⁹

Robert M. Berdahl, when he was president of the Association of American Universities (AAU), courageously asked: “How many research universities does the nation require? … I do not know how many we should have. But it is a serious question, worthy of examination.”³⁰ Berdahl’s probing question led to a two-year, congressionally mandated assessment of financial threats to the nation’s research universities. The study did not, however, answer Berdahl’s central question—which is, to be sure, highly sensitive. William (“Brit”) Kirwan, chancellor of the University System of Maryland, has called this a missed opportunity “to address that very point more explicitly.”³¹ I agree.

During my time at the Andrew W. Mellon Foundation, I tried a slightly different tactic—namely, to encourage, with the carrot of substantial grant funding, some universities with PhD programs that were not highly ranked to substitute less expensive yet stronger postdoctoral programs for them. I was dismayed to find that many presidents agreed privately with my assessment of what made sense but were unwilling to take the political heat that would have been generated by an effort to dismantle any PhD programs. Worries by faculty about potential loss of status in the profession overwhelmed all else, and presidents who had other battles to fight were unwilling to risk struggling with faculty on this issue, with all of its symbolic overtones.³²

Supply-Side Problems and Mismatching

Less controversial and every bit as fundamental are two systemic issues: (1) ineffective supply-side provision of higher education by some institutions, combined with weak incentives for students to finish programs in a timely way; and (2) what is known as the mismatching problem.

Sarah Turner and colleagues, in an important and under-appreciated paper, have documented a marked increase in time-to-degree (TTD) over the last three decades.³³ If it takes longer for students to complete their degrees, and if large numbers never finish, the implications for productivity are clear. As someone observed, “The most expensive degrees are those that are never obtained”—or, one might add, the ones that require five, six, or more years to obtain.³⁴ A lengthening of time-to-degree could, of course, be the result of an influx of poorly prepared students, but Turner and her colleagues have demonstrated rigorously that this is not the main source of the problem. Indeed, they found that “the increase in TTD is localized among those who begin their postsecondary education at public colleges outside the most selective institutions.” A combination of declines in resources at these less selective public institutions and the tendency for students to work more hours for pay (at the expense of finishing their studies) is at the root of the problem.

There is abundant evidence that undergraduate students who fail to graduate in four or four and a half years often take more credits than they need, in part because of inadequate guidance, starting and stopping majors, and a lack of places in gateway courses.³⁵ Student attitudes are another part of the problem. A recent graduate of a highly selective flagship university in our Crossing the Finish Line study said that at his university, graduating in four years was like “leaving the party at 10:30 P.M.”³⁶ But we are starting to see reports that schools are now addressing the problem of long TTD more aggressively, by altering the way they charge for credits and pushing “super-seniors” to graduate in a timely way. There is less willingness to tolerate five- or six-year graduation rates.³⁷ Easing transfer paths from two-year to four-year institutions would also make a considerable difference, and places like the City University of New York are making active efforts to facilitate flow through the system.³⁸

There is also strong evidence that both lower completion rates and longer time-to-degree are caused in no small measure by the failure of surprisingly large numbers of well-qualified students to enroll at colleges and universities for which they are qualified, ending up instead either at less challenging institutions or at no postsecondary institution at all. The primary source of this problem is at the application stage: large numbers of students, and especially students from poor families and some minority groups, simply do not apply to institutions at which students with their qualifications do well. The University of Chicago Consortium on Chicago School Research played a pivotal role in introducing into this discussion the concept of “match”—that is, the match between the student’s qualifications and the selectivity level of the college he or she actually attends. The Chicago Consortium has described in detail the frustrating experience in its city of watching students who had worked hard and successfully in high school then fail to take advantage of the potential collegegoing opportunities that they had earned. In Crossing the Finish Line, our research team found strong evidence of this same “undermatching” phenomenon in North Carolina.³⁹

The serious consequences of this persistent pattern are related directly to the by now well-documented empirical relationship between completion rates and the selectivity of colleges and universities. Even after controlling carefully for differences in the observable qualifications of entering students, evidence shows that students who attend institutions that enroll high-achieving students are themselves more likely to graduate, and to graduate in four years, than are comparable students who attend less selective institutions. This finding may seem counterintuitive at first—shouldn’t it be easier to graduate from less selective (and presumably less rigorous) schools than from those that are more selective and more rigorous? But the finding is correct. Presumably peer effects, differences in expectations for graduation, opportunities to work closely with faculty, and institutional resources such as libraries and laboratories are very important.⁴⁰

Two major research projects are now under way to study alternative ways of alleviating the mismatch problem: one is directed by the research organization MDRC in New York, and one is led by Professors Caroline Hoxby from Stanford University and Sarah Turner from the University of Virginia.⁴¹ Success on this front would reduce disparities in outcomes related to socioeconomic status by raising timely completion rates for well-qualified students, often from disadvantaged backgrounds, who currently mismatch. It is also possible that improvements in match would raise overall completion rates somewhat, though this is conjectural and there is no hard evidence of which I am aware. Much depends on the elasticity of the supply of places at moderately selective institutions and on what would happen as a result of any reshuffling of the student population that would occur because of better matching.⁴²

Affordability

The word affordability has achieved iconic status; it became a part of the ad wars in the 2012 presidential campaign. Is higher education affordable today for students and their families? Will it be affordable tomorrow? These are key questions to ponder, but they do not lend themselves to simple answers. This is a murky terrain, and I hope you will be pleased to know that I intend to ride roughshod over it.⁴³ For my purposes, it will suffice to note commonly cited numbers generated by others and emphasize a limited number of basic points.

At the root of much of the discussion of affordability is the well-known fact that state appropriations per student have declined sharply in recent years, both in absolute terms and relative to other sources of revenue.⁴⁴ According to one study, the state appropriations share of the total receipts of public colleges and universities fell from 44 percent in 1980 to 22 percent in 2009.⁴⁵

Particularly in recent years, the price students pay for college at public universities has risen faster than per-student costs (never mind prices in general)—at the same time that these universities have experienced reductions in state and local support.⁴⁶ Public systems seeking to avoid cutbacks in enrollment and to maintain quality have had little choice but to raise charges. A 2012 College Board report states: “Over the 30 years from 1982–83 to 2012–13, the increase for in-state students [in average published tuition and fees] at public four-year institutions was 257%, from $2,423 to $8,655”;⁴⁷ furthermore, “the growth in published prices at public four-year institutions has been higher over the past decade (averaging 5.2% per year beyond inflation) than over either of the two preceding decades.”⁴⁸ Net charges have increased less rapidly because of both efforts to augment financial aid and the substitution of some federal dollars, including stimulus aid, for state dollars. Still, net tuition as a percent of total educational revenue in public higher education rose from 23 percent in 1986 to 43 percent in 2011 (figure 2).⁴⁹

Furthermore, as Joseph Stiglitz, the 2001 Nobel laureate in economics and former chief economist of the World Bank, has emphasized: “Parents’ ability to pay without resorting to debt is declining.… The income of the typical American family, adjusted for inflation, declined from 2007 to 2010. Their wealth was down almost 40%. Separate data show that household income is back to the levels of a decade and a half ago.”⁵⁰ Economic conditions have indeed taken a toll, and those who complain that college costs are rising faster than incomes should recognize that stagnation of median family incomes is definitely one blade of this scissors.

Figure 2 Net tuition as a percentage of total educational revenue in U.S. public higher education, fiscal 1986–2011

Note: In calculating these figures, net tuition revenue used for capital debt service is included in net tuition but excluded from total educational revenue. Years are state fiscal years, which commonly start July 1 and run through June 30 of the following calendar year. For example, FY 1986 is July 1985 through June 1986.

Source: State Higher Education Executive Officers, SHEF—State Higher Education Finance FY 2011, 2012, http://www.sheeo.org/resources/publications​/shef-%E2%80%94-state-higher-education-finance-fy11.

It is important to recognize that these trends do not appear to have led most students and their parents to conclude that college is not for them or is simply beyond their reach. Indeed, 83 percent of college students and parents participating in the most recent Sallie Mae / Ipsos survey strongly agreed that “education is an investment in the future,”⁵¹ and a majority said they were “willing to stretch [themselves] financially” to make this education possible.⁵² Strong demand for higher education appears to be ever present, but it would be helpful to have more hard evidence than is available now as to the actual effects on student behavior of increases in tuition and changes in financial aid policies at public universities.⁵³ It would be a mistake to assume that the apparent lack of big effects to date guarantees that there will be no adverse effects in the future.

Figure 3 Trends in student debt and credit card debt, 2003–2012

Source: Federal Reserve Bank of New York, “Household Debt and Credit Report,” September 30, 2012, http://data.newyorkfed.org/householdcredit/.

We get closer to the core of affordability concerns when we recognize that the combination of upward trends in charges and deteriorating family circumstances has led to a large increase in student debt—which, as has been widely reported, now exceeds credit card debt (figure 3).⁵⁴ For many students, borrowing has become the only option. Student debt has risen sharply—though nothing like as astronomically as the incredibly inept story in the Sunday New York Times of May 13, 2012 suggested.⁵⁵ The Times originally reported that 94 percent of bachelor’s graduates leave college with educational debt. The correct number is around two-thirds, as Sandy Baum and Michael S. McPherson pointed out in their devastating commentary on the Times article.⁵⁶ An equally troubling aspect of the Times article (again quoting Baum and McPherson) is that it “focused on a student who has more debt than almost every other college graduate and who chose to enroll at an institution, Ohio Northern University, where average debt levels exceed those at almost every other college in the country.”⁵⁷ As a colleague of mine now at the Brookings Institution, Matthew M. Chingos, wryly observed: “Share of student borrowers with >$54k debt: 10%; share of grads interviewed by NYT with >$54k debt: 100%.”⁵⁸ Gross misreporting and fear-mongering do not encourage thoughtful consideration of a complex issue.

Professors Christopher Avery and Sarah Turner have made a commendable effort to create an analytical framework that can be used to think through borrowing decisions.⁵⁹ They ask whether (and when) college students borrow too much, and whether (and when) college students borrow too little. Much depends, they explain, on the aptitudes/talents of an individual, choice of major, institution attended, likelihood of actually getting a degree, career interests and prospects, and so on. An important conclusion of their research is: “The claim that student borrowing is ‘too high’ across the board can—with the possible exception of for-profit colleges—clearly be rejected.”⁶⁰ There are many cases in which students elect (often unwisely, I believe) not to borrow modest sums needed to finish degree programs in a timely way, choosing instead to work so many hours on off-campus jobs that they either delay completing their programs or do not complete them at all.⁶¹ U.S. Department of Education data show that most students have been graduating with what seem like manageable debt loads. According to data from the Department of Education, three-quarters of four-year graduates owe less than $33,857 on earning a degree—often much less.⁶² The Pew Research Center recently reported that the average outstanding student loan balance was $26,682 in 2010 among all households with student debt; the Pew study also shows (not surprisingly) that the relative burden of student debt is greatest for households in the bottom fifth of the income distribution.⁶³

In correcting overwrought worries about student debt, I do not want to go to the other extreme. Many students surely borrow too much and have their lives affected adversely. There is evidence that high debt may make students less likely to choose lower-paying jobs which might well prove more satisfying, and this is one reason why I favor combining loan programs with forgiveness features that take account of job choices.⁶⁴ An irony is that many students attending highly selective, wealthy universities such as Princeton and Harvard should be the best candidates to borrow at least modest sums to pay part of the costs of their education—yet these are the very institutions that, for what are surely praiseworthy reasons, have elected to adopt grant-only financial aid programs. As I suggested earlier, an unfortunate consequence is that less wealthy colleges and universities can feel pressured to adopt financial aid policies that are unwise for them.

Reluctance to take on even a modest amount of debt may also have a sizeable impact on college choice and contribute to the mismatch problem described earlier. Where one goes to school is by no means the be-all and end-all, but it can be important. I am reminded of an experience I had in the aftermath of the publication of the book Derek Bok and I wrote on affirmative action.⁶⁵ I was at a gathering in Washington, DC, when a white woman stood up and said that surely there are many fine schools in America, and she couldn’t understand why minorities made such a fuss about getting into a place like Stanford. An African-American woman stood up and replied: “Wait a minute. Are you telling me that all those white folks fighting so hard to get their kids into Stanford are just ignorant? Or are we supposed to believe that attending a top-ranked school is important for the children of the privileged but shouldn’t matter to minorities?” There was dead silence. Interestingly, the evidence we presented in The Shape of the River shows that the gains associated with attending the most selective schools are, if anything, greater for minorities than for whites.⁶⁶

Is There a Serious Problem—Even a Crisis?

There are certainly reasons to think so. Among measures of educational outcomes, more and more attention is being focused on completion rates.⁶⁷ Yet, in spite of President Obama’s exhortations,⁶⁸ various Department of Education initiatives, and vigorous efforts by the Gates and Lumina Foundations, among other private players, there is no evidence that levels of educational attainment in the United States are rising to match the progress made in other countries.⁶⁹ Moreover, serious questions have been raised regarding the capacity of America’s higher education system to deliver on a second core mission: to enhance mobility and serve as a powerful equalizer—as an engine of opportunity. Scholars have found evidence that achievement gaps between rich and poor children have been increasing, not closing.⁷⁰

There are also numerous voices saying that many colleges and universities, including even prestigious places, are on financially unsustainable paths. Moody’s recent “Higher Education Mid-Year Outlook” paints a grim picture of the future of higher education. One of our most esteemed leaders in higher education, Brit Kirwan, has been warning for years that we are indeed in perilous times. Speaking before the AAU in the spring of 2010, he said, “We are also in a period of fiscal famine, experiencing unprecedented resource trauma that threatens the ability of many, if not most of our institutions to carry out their core missions.”⁷¹ I agree with Chancellor Kirwan’s assessment. But I would add (and I don’t think he would disagree) that it is easy, and wrong, to underplay the staying power and resiliency of colleges and universities—a lesson that history teaches us. We should avoid that mistake.

Nor should we blame the “inexorable” workings of the cost disease for whatever grim prospects seem to lie ahead. In a new book, William J. Baumol explains clearly that the same economy-wide increases in productivity that are at the root of the cost disease raise overall wealth and generate additional resources that could be used to pay the rising relative costs of activities in labor-intensive sectors such as education, if we were to choose to spend them in this way. As Baumol notes in his introduction, this proposition about “possibilities” was first explained to him by the renowned Cambridge economist Joan Robinson many decades ago, but even Baumol did not immediately recognize its full implications. Future prospects come down to a matter of priorities. Could is not the same as will. The key question, then, is whether we will choose, collectively, to invest the fruits of overall productivity gains on goods such as quality education.⁷²

My verdict: not likely. To be sure, in November 2012 voters in California passed “a tax hike that averts what many called potentially disastrous cuts” in state support for higher education.⁷³ But it is well to remember that this action comes after a series of sharp cuts in state funding, tuition increases, and the imposition of enrollment caps.⁷⁴ This is also a special case in that the dire consequences for a great many people, and perhaps for the state overall, of failing to avoid further cuts were spelled out in great detail—and yet the margin of victory was still fairly narrow. Previous cuts were not rescinded, and a bit of breathing room is all that has been achieved.

It seems to me, as to many others, that people in general are fed up with rising costs, and especially rising student charges, however understandable the reasons for them may be. As Baum, Kurose, and McPherson put it in an early version of their “Overview” paper: “The anger and resentment expressed toward college leaders appear to be growing, despite the limited ability of those leaders to make college cheaper quickly without lowering quality in ways that will disappoint the same people who decry higher prices.” They added: “Americans as a whole seem extremely reluctant to accept the idea that they should pay more in order to provide more education to more students. Instead the prevalent view seems to be that colleges and universities, especially those in the public sector, should simply find ways to do more with less. If nothing else, sheer political prudence requires colleges to redouble their efforts to accomplish just that, and to undertake those efforts in the most visible possible way.”⁷⁵

Otherwise, or in any case, politicians are likely to seek to tie state support to performance goals that may in some cases be overly narrow or short-term in their focus. For example, the governor of Wisconsin has suggested rewarding students for earning degrees “in jobs that are open and needed today,” however relevant (or irrelevant) such training for today’s jobs may be in the long run.⁷⁶

No part of higher education is immune from the consequences of ignoring this rising tide of anger and resentment. Public perceptions matter, and even seemingly sacrosanct programs such as National Institutes of Health funding for research could be affected if there is spreading distrust of higher education and disbelief in its willingness (commitment?) to “do more with less.” Thus, there are self-serving reasons for even privileged institutions such as Stanford and Princeton to pay close attention to these issues. There are also, of course, nobler instincts at play, and I believe, as I will say in the second part of this book, that thoughtfully developed system-wide efforts have the potential not to cure the cost disease but to ease its harshest effects. This will be far from easy. There are no silver bullets in sight. But there is promising work to be done, if only we can muster the will to meet challenges that are at least as much organizational and philosophical as they are technical. As John Doar used to say to me in the context of the Nixon impeachment inquiry, which he led, “We will know more later.”

Notes

1. Clark Kerr, foreword to William G. Bowen, The Economics of the Major Private Universities (New York: McGraw-Hill, 1968).

2. William J. Baumol and William G. Bowen, Performing Arts, the Economic Dilemma: A Study of Problems Common to Theater, Opera, Music, and Dance (New York: Twentieth Century Fund, 1966).

3. Robert H. Frank, “The Prestige Chase Is Raising College Costs,” New York Times, March 10, 2012. As several commentators have pointed out, it is clear that Frank was referring to only the first movement of Beethoven’s String Quartet No. 4 in C minor, Op. 18, which is approximately nine minutes long. The quartet in its entirety is approximately twenty-five minutes long.

4. See Baumol and Bowen, Performing Arts, the Economic Dilemma, especially chapters 8 and 9, which document trends in the overall cost of performance, performers’ salaries, and other cost components.

5. See William G. Bowen, “The Economics of Princeton in the 1970s: Some Worrisome Implications of Trying to Make Do with Less,” Report of the President, February 1976. This report also cites data for Australia (and see Bowen, The Economics of the Major Private Universities, for data on the United Kingdom). The cost disease is no respecter of national boundaries. In this report, I also restated the basic cost-disease proposition a bit more fully: “The central economic fact of life is the very nature of the processes of education and scholarship. To be done well, particularly at advanced levels, they require a degree of personal attention and personal interaction that simply do not allow the same opportunities for technological change, mechanization, and, if you will, increases in ‘output per unit of labor input,’ that characterize the production of such goods as feed grains and calculators. As a result, we must expect the costs and prices of educational services to rise more rapidly than prices in general over the long run.” (Bowen, “The Economics of Princeton in the 1970s,” 5.) In retrospect, I am especially pleased that I emphasized that the cost disease is a particularly intractable problem at “advanced levels” of teaching.

6. Sandy Baum, Charles Kurose, and Michael S. McPherson, “An Overview of Higher Education,” paper presented at Future of Children Postsecondary Education in the United States authors’ conference, Princeton, NJ, April 26–27, 2012. The study they cite, based on Delta Cost Project data, is Donna M. Desrochers and Jane V. Wellman, “Trends in College Spending, 1999–2009” (Washington, DC: Delta Cost Project, 2011). The “Overview” paper includes a veritable wealth of data on almost every aspect of American higher education over the last fifty years and should be a standard reference. It also contains yet another explanation of the cost disease, and one that is especially detailed—but the paper’s authors do not, as I do not, put the onus for rising costs solely on this phenomenon.

The Delta Cost Project, cited extensively in Baum, Kurose, and McPherson’s “Overview,” has worked with aggregate data available publicly through the Integrated Postsecondary Education Data System. These data in their raw form are not consistent over time, and a real contribution of the Delta Cost Project has been to “clean” the public data and make them consistent. The project has looked at both revenues and expenditures, focusing on public research universities, public master’s universities, public community colleges, not-for-profit private research universities, not-for-profit private master’s institutions, and not-for-profit private bachelor’s institutions; the for-profit sector has not been included because it has been too difficult to obtain consistent data on trends. A good summary of the most recent data is provided in Wellman’s testimony before the House Subcommittee on Higher Education; see Jane V. Wellman, Statement to the House, Subcommittee on Higher Education and Workforce Training, Committee on Education and the Workforce, Keeping College Within Reach: Discussing Ways Institutions Can Streamline Costs and Reduce Tuition, Hearing, November 30, 2011 (Serial No. 112–48).

Another well-known source of data on costs is a set of studies known as the Delaware Studies. In response to a congressional mandate in the 1998 Higher Education Act, the National Center for Education Statistics (NCES) published three reports, the last of which is Michael F. Middaugh, Rosalinda Graham, and Abdus Shahid, A Study of Higher Education Instructional Expenditures: The Delaware Study of Instructional Costs and Productivity (Washington, DC: U.S. Department of Education, National Center for Education Statistics, 2003). Since the publication of that third report, the University of Delaware has continued to survey institutions that volunteer to participate and to give participating institutions their own data, as well as data for sets of comparable institutions (defined by Carnegie classification) so that they can benchmark their own data. These studies are concerned with academic disciplines (such as mathematics). The Delaware Studies also focus on direct instructional expenditures and the factors associated with calculations of direct instructional expenditures per student credit hour (enrollment multiplied by course credits) at four-year colleges and universities. A main finding is that differences among institutions are heavily driven by discipline mix, with costs higher in the sciences than in the humanities. Costs are also related to Carnegie classification (and are higher at research universities) and to variables such as scale.

7. See Christopher Newfield, “Democrats Need a Huge Push to Fix Public Higher Education,” Chronicle of Higher Education, Conversation blog, November 19, 2012.

8. Teresa A. Sullivan et al., eds., Improving Measurement of Productivity in Higher Education: Panel on Measuring Higher Education Productivity: Conceptual Framework and Data Needs (Washington, DC: National Academies Press, 2012), 2. There are also many papers that discuss ways to improve productivity in higher education. See, for example, Davis Jenkins and Olga Rodriguez, “Access and Success for Less: Improving Productivity in Broad-Access Postsecondary Institutions,” The Future of Children (forthcoming); and Douglas N. Harris and Sara Goldrick-Rab, “The (Un)Productivity of American Higher Education: From ‘Cost Disease’ to Cost-Effectiveness,” WISCAPE Working Paper, December 2010. But these papers have more to say about forces driving up costs (the subject of my next section) than about the measurement of productivity per se.

9. See Spencer S. Jones, Paul S. Heaton, Robert S. Rudin, and Eric C. Schneider, “Unraveling the IT Productivity Paradox—Lessons for Health Care,” New England Journal of Medicine 366 (June 14, 2012): 2243–45.

10. Ibid., 2243.

11. Ibid., 2244.

12. See William G. Bowen and Derek Bok, The Shape of the River: Long-Term Consequences of Considering Race in College and University Admissions (Princeton, NJ: Princeton University Press, 1998).

13. Kevin M. Guthrie, founding president of JSTOR and now president of ITHAKA (which encompasses JSTOR), points out that, thanks to JSTOR, there is now some work that gets done that never would have been done before—as a result, for example, of finding old literature that otherwise would have been inaccessible—thus adding to the numerator of the productivity ratio. Similarly, universities in places like South Africa now have access to both literature (via JSTOR) and art images (via ARTstor) that otherwise never would have been possible.

14. During the discussion session following the presentation of this lecture, President Hennessy contributed a short list of other factors that he believed, on the basis of his experience at Stanford, to have contributed to rising costs. These included increases in state Medicaid obligations as a proportion of state budgets, which can crowd out appropriations for higher education (see Thomas J. Kane, Peter R. Orszag, and David L. Gunter, “State Fiscal Constraints and Higher Education Spending: The Role of Medicaid and the Business Cycle,” Tax Policy Center, May 2003, http://tax​policy​center.org​/publications/url.cfm?ID=310787); increases in expenditures on student mental health costs; increases in the proportion of low-income and first-generation college students at universities such as Stanford; and increases in initiatives like community centers and ethnic-themed dorms that Hennessy believes have helped improve Stanford’s completion rate.

15. Forbes, November 19, 2012.

16. For a summary of Bain’s work on these campuses, see Kevin Kiley, “Where Universities Can Be Cut,” Inside Higher Ed, September 16, 2011. Also see Jeff Denneen and Tom Dretler, The Financially Sustainable University (Bain & Company and Sterling Partners, 2012), www.bain.com/Images​/BAIN​_BRIEF​_The​_financially​_sustainable​_university.pdf. The McKinsey con sulting firm has been active in this area as well; see, for instance, Byron G. Auguste, Adam Cota, Kartik Jayaram, and Martha C. A. Laboissière, Winning by Degrees: The Strategies of Highly Productive Higher Education Institutions (McKinsey & Company, November 2010), http://mckinsey​on​society.com​/downloads​/reports​/Education​/Winning​%20by​%20degrees​%20report​%20full​report%20v5.pdf.

17. In a lengthy front-page story in the Wall Street Journal, Douglas Belkin and Scott Thurm analyze data for the University of Minnesota to make the case that inefficiencies and a growing bureaucracy have been largely responsible for increasing costs. (Douglas Belkin and Scott Thurm, “Deans List: Hiring Spree Fattens College Bureaucracy—and Tuition,” Wall Street Journal, December 29, 2012.) In considering this argument, it is wise to remember that the University of Minnesota may be an outlier in some respects; for example, as the authors point out, data compiled by the U.S. Department of Education show that, among the seventy-two “very-high-research” public universities in the 2011–12 academic year, the University’s main Twin Cities campus had the largest proportion of “executive/administrative and managerial” employees. Not all public institutions have experienced what is sometimes described as an increase in “administrative bloat.” For example, the University of Nebraska, according to a Chronicle of Higher Education article published less than two weeks later, has seen a 5 percent decline in the number of employees with administrative faculty positions between 2001 and 2012.

At the same time, however, Jane Wellman, currently executive director of the National Association of System Heads and formerly head of the Delta Cost Project, notes that the costs of employing more administrators amount to too small a proportion of universities’ overall budgets to account for recent tuition increases or higher education costs; even the University of Minnesota’s costs of administrative oversight constitute only 9 percent of its expenditures. In addition, it is important to ask what fraction of the increases in non-instructional costs at Minnesota can be attributed to the growth in research programs over the period in question—probably a high fraction. For example, Minnesota’s technology, research, communications, and student-services staffs increased “significant[ly]” over the past thirteen years, and research staff at the University of Nebraska doubled as its research expenditures increased from $136 million to $232 million between 2000 and 2011. (Belkin and Thurm, “Deans List”; Jenny Rogers, “How Many Administrators Are Too Many?” Chronicle of Higher Education, January 7, 2013.)

18. When I was president of the Andrew W. Mellon Foundation, I resolutely opposed the creation of new “centers” unless there was an overwhelming case for them. The University of Chicago has been an especially creative user of the “workshop” model, which stresses the value of temporary groupings of faculty and students. Tufts University has had excellent results with the use of the “experimental college” to test out curricular experiments without making more than one-year commitments.

19. For an extended discussion of these issues, written by two veteran leaders of public-university business schools, see Gary C. Fethke and Andrew J. Policano, Public No More: A New Path to Excellence for America’s Public Universities (Stanford, CA: Stanford University Press, 2012), especially chapters 7, 8, and 12. The controversy at the University of Virginia in spring 2012 over the leadership of President Teresa Sullivan also involved aspects of this debate (how important is it for U.Va. to teach German?). See Scott Jaschik, “Fired for Protecting Languages?” Inside Higher Ed, June 18, 2012. Of course, it is easy to state such issues in overly simple terms. For example, it is one thing to insist that U.Va. must continue to be a leader in the teaching of German but quite another to insist that many separate language departments are essential. Precisely how worthy educational objectives are to be served is a legitimate question, which I raise here without suggesting an answer.

20. See Charles T. Clotfelter, Buying the Best: Cost Escalation in Elite Higher Education (Princeton, NJ: Princeton University Press, 1996). Clotfelter looks at the economic factors that drive actions by institutions of higher education and examines the escalation in spending in the arts and sciences at four elite institutions: Harvard, Duke, the University of Chicago, and Carleton College. He argues that the rise in costs has less to do with increasing faculty salaries or decreasing productivity than with broad-based efforts to improve quality, provide new services to students, pay for large investments in new facilities and equipment (including computers), and ensure access for low-income students through increasingly expensive financial aid.

21. For an early presentation of this line of argument, see Howard Bowen’s discussion of the revenue theory of cost in his The Cost of Higher Education (San Francisco: Jossey-Bass, 1980).

22. Lawrence S. Bacow, personal communication, November 22, 2011.

23. Even among the most prestigious universities in the nation, there is a dramatic difference in endowments, which has grown, in absolute terms, over the past ten years. In fiscal year 2001, for instance, Harvard’s endowment was about $18 billion, and those of Yale and Princeton were $10.7 billion and $8.4 billion, respectively. By contrast, in that fiscal year Columbia had an endowment of $4.2 billion, and the University of Pennsylvania and the University of Chicago had endowments of $3.4 billion and $3.3 billion, respectively. By fiscal year 2011, Harvard’s endowment was $31.7 billion, and those of Yale and Princeton had grown to $19.4 billion and $17.1 billion, respectively. In comparison, the endowments of Columbia, the University of Pennsylvania, and the University of Chicago were each in the $6 billion to $8 billion range. See National Association of College and University Business Officers, NACUBO Endowment Study, 2002, and NACUBO-Commonfund Study of Endowments, 2011, both available at www.nacubo.org​/Research​/NACUBO-Commonfund​_Study​_of​_Endowments​/Public​_NCSE​_Tables_.html.

For an excellent discussion of stratification more generally, and the forces driving it, see Caroline M. Hoxby, “The Changing Selectivity of American Colleges,” Journal of Economic Perspectives 23, no. 4 (2009), especially pp. 95–96 and 98. Hoxby concentrates on stratification by student achievement and notes that only the top 10 percent of colleges were substantially more selective in 2007 than they were in 1962. Most colleges became less selective over this period. Hoxby also explains how student selectivity and educational expenditures interact.

24. President Catherine Hill from Vassar spoke eloquently at the April 2012 Lafayette Conference on the Future of the Liberal Arts College on the pressure that selective institutions feel to meet the perceived needs and desires of students from affluent families. Also telling is a recent article in the Atlantic about former George Washington University president Stephen Trachtenberg, who, between 1988 and 2007, built not only computer and research labs but also a “profusion of comforts [that] didn’t just stimulate students’ minds; they also fulfilled their every whim—a change that drew a more selective, more intelligent group of applicants and sent the admission rate plummeting from 75 percent to 37 percent.” Unsurprisingly, these new amenities caused tuition to double during the term of his presidency (after inflation is taken into account). The article continues: “ ‘It was a matter of competition,’ Trachtenberg says in the way of justifying the suite-style dorms he built, the remote campus he acquired, and the tuition hikes that he began almost immediately. ‘These sorts of facilities were being offered elsewhere. We were either in the game or we weren’t in the game.’ ” (Julia Edwards, “Meet the High Priest of Runaway College Inflation (He Regrets Nothing),” Atlantic, September 30, 2012.)

25. See William G. Bowen, Matthew M. Chingos, and Michael S. McPherson, Crossing the Finish Line: Completing College at America’s Public Universities (Princeton, NJ: Princeton University Press, 2009), especially chapter 6, for striking evidence that performance-based indicators, such as class rank and achievement test scores, are far better predictors than aptitude tests like the SAT of almost everything—except family wealth! Preoccupation with yield is also troubling. I can’t count the number of times at Princeton that I told the admissions office that I didn’t want to hear anything about yield; what matters, I did my best to explain, is the quality and character of the incoming class, not how many promising students elected to go elsewhere.

26. The quotation was provided by a former colleague of mine at the Andrew W. Mellon Foundation, Idana Goldberg.

27. As president of Princeton, I led an effort within the Ivy League to create and apply an “academic index” to govern admissions requirements. It had some good effects, but it also had some undesirable side effects that I did not understand until much later. (See William G. Bowen and Sarah Levin, Reclaiming the Game [Princeton, NJ: Princeton University Press, 2008].)

28. Institutions can no longer require tenured faculty members to retire at age 70 (as was possible before 1994). As a result, “I can stay forever,” Ronald Ehrenberg, a tenured faculty member and director of the Cornell Higher Education Research Institute at Cornell University, recently told Inside Higher Ed. This situation can prove frustrating and expensive for tenure-granting administrators. One university president at a highly renowned private university lamented having to spend millions of dollars on buying out superannuated and unproductive faculty members who are reluctant to retire. (Colleen Flaherty, “Adjunct Leaders Consider Strategies to Force Change,” Inside Higher Ed, January 9, 2013.)

29. William G. Bowen and Neil L. Rudenstine, In Pursuit of the PhD (Princeton, NJ: Princeton University Press, 1992).

30. See Robert M. Berdahl, “Reassessing the Value of Research Universities,” Chronicle of Higher Education, July 13, 2009.

31. See Paul Basken, “Nation’s Research Universities Are Offered Hope of Fatter Budgets—at a Price,” Chronicle of Higher Education, June 14, 2012.

32. During the discussion session following the original delivery of this lecture, President Hennessy was very direct in saying that the United States is producing too many PhD’s and that we are going to have to accept the fact that in the future there will be fewer faculty positions than there are today. That being said, a relatively wealthy university like Stanford will probably be better able to finance its graduate programs in the future than many less-well-off public and private institutions. Still, it is unlikely that Stanford will remain completely insulated from pressures to economize.

33. See John Bound, Michael F. Lovenheim, and Sarah Turner, “Increasing Time to Baccalaureate Degree in the United States,” National Bureau of Economic Research, Working Paper 15892, April 2010. Also see the discussion of this study in Baum, Kurose, and McPherson, “Overview.” Of course, college readiness is also part of the problem. The low rate of college readiness among high school graduates in, for example, New York is certainly troubling. (See Sharon Otterman, “College-Readiness Low among State Graduates, Data Show,” New York Times, June 14, 2011.)

34. Focusing on completion rates and time-to-degree reminds us of the dangers of looking only (or mainly) at measures of “cost per [enrolled] student.” There is much to be said for examining “cost per degree conferred,” though here again care is needed to avoid treating all degrees as the same or assigning no value at all to learning that does not result in a degree.

35. See Bowen, Chingos, and McPherson, Crossing the Finish Line, especially chapter 4. It is Matthew M. Chingos who deserves the credit for this part of the analysis. Similarly, a Chronicle of Higher Education article reports that “many students take far more credits than they need to earn a degree” and thus stay in school longer than necessary. (Eric Kelderman, “Board Suggests Ways for Southern States to Lower College Costs and Increase Degree Production,” Chronicle of Higher Education, September 28, 2010.) A significant source of the problem is blocked passages to degrees. A survey conducted by the chancellor’s office of the California Community Colleges system in late August 2012 revealed that more than 472,000 of the system’s 2.4 million students were put on waiting lists for fall 2012 classes. (See Lee Gardner, “Survey of California Community Colleges Reveals Drastic Effects of Budget Cuts,” Chronicle of Higher Education, August 29, 2012.)

36. See Bowen, Chingos, and McPherson, Crossing the Finish Line, 237.

37. See Daniel de Vise, “Public Universities Pushing Super-Seniors to the Graduation Stage,” Washington Post, June 2, 2012. Other imaginative efforts have been made to attack this problem. Lawrence S. Bacow, former president of Tufts, has reported that Tufts had success with a modest program of grants that allowed the neediest students to attend summer school between junior and senior year (something that wealthier students often do without institutional help).

38. In June 2011 the trustees of the City University of New York (CUNY) approved a resolution creating the Pathways initiative, which is designed to facilitate the transfer process between the system’s two- and four-year colleges. Under this project, all students in the system are required to complete thirty “Common Core” credits; students who are transferring from community to senior colleges are required to take an additional six to twelve “College Option” credits. Individual colleges have substantial flexibility in determining the content of the “Common Core” credits, and, in the case of the senior colleges, the College Option credits. For more information on the Pathways Initiative, see www.cuny.edu​/academics​/initiatives​/degreepathways.html, accessed August 30, 2012.

39. See chapter 5 of Bowen, Chingos, and McPherson, Crossing the Finish Line, for an extended discussion of the concept of matching and the phenomenon of undermatching in North Carolina, and p. 99 in particular for a description of the important work of the Chicago Consortium in that city.

40. Ibid., especially chapter 10 and pp. 233–35. Hard as we worked to control for selection effects in this study, a new paper by Sarah Cohodes and Joshua Goodman controls for selection effects even more convincingly by using a research discontinuity design and working with rich data from a Massachusetts Merit Aid program. The authors find that students induced by this scholarship program to attend less selective colleges were more than 40 percent less likely to graduate. They also find that “students are remarkably willing to forego college quality for relatively small amounts of money.” (Sarah Cohodes and Joshua Goodman, “First Degree Earns: The Impact of College Quality on College Completion Rates,” Harvard Kennedy School Working Paper Series RWP12–033, August 7, 2012.) The classic (crisp) discussion of returns to selectivity is Hoxby, “Changing Selectivity,” especially pp. 114–15. Hoxby emphasizes that colleges and universities that attract high-achieving students also invest more in “student-oriented resources”—which is, of course, an important reason that so many students with strong qualifications go to these schools.

41. For a description of the MDRC study and a policy brief explaining the context of the project, see Jay Sherwin, “Make Me a Match: Helping Low-Income and First-Generation Students Make Good College Choices,” MDRC, March 2012, www.mdrc.org/publications/623/overview.html. For a description of the Hoxby-Turner project, see “About the Expanding College Opportunities (ECO) Project,” Expanding College Opportunities, www.expandingcollegeopps.org/eco/about, accessed August 29, 2012. One reader of a draft of this Tanner Lecture suggested that lessons might also be learned from matching programs for medical students.

42. Some have wondered if improving “match” would just substitute more-qualified students for less-qualified students at certain institutions, leaving overall enrollments (and perhaps overall graduation rates) unchanged. This is a good question, and it is surely true that there would probably be no significant effects on overall completion rates at highly selective institutions. But we are not really interested in the Harvards of the world in this context, and I suspect that there is more elasticity in the capacity of moderately selective institutions than is sometimes understood—especially when we recognize that reducing time-to-degree increases the number of students who can be accommodated with a given number of classroom seats. We should also be aware, however, that these moderately selective institutions might have difficulty meeting the financial aid needs of larger numbers of students from modest family circumstances. More thought also needs to be given to the likely effects of any displacement of some students that might occur as a result of better matching. Since there is evidence that it is precisely the kinds of students currently “undermatched” who benefit most from attending institutions with challenging academic programs, it is certainly possible that some reshuffling of the student population could, in and of itself, improve overall completion rates and time-to-degree—and thus productivity. It is also true that such reshuffling might be hard to manage politically, its appeal on the merits notwithstanding.

43. To move from an analysis of institutional costs in higher education to a discussion of affordability for students and their families requires us to peel several layers off the proverbial onion. First, students in nonprofit institutions of all kinds are almost never expected to pay the full costs of their education. State appropriations, federal grants, private gifts, earnings from endowments, and earned income are other sources of revenue which drive a wedge between costs and tuition. Nearly all students in the nonprofit sector receive subsidies, which are often non-trivial in size. Second, thanks to financial aid and “discounts,” there is often a sizeable difference between quoted tuition (“sticker price”) and what students actually pay (“net tuition”). Third, affordability depends not just on what a student is expected to pay, but on trends in family income and wealth that, in turn, depend on variables external to higher education. Finally, it can be difficult to calibrate the long-run effects of different choices that students and parents make in deciding how to pay their college bills, including how much to borrow and what forms of debt make the most sense. Difficulties involved in making these distinctions are compounded by huge differences in tuition levels across higher education and the tendency of journalists to pay far too much attention to stated charges at elite private institutions that enroll only a small fraction of students. Also, it has proved difficult for prospective students and their families to understand widely differing financial aid policies and to recognize that in many cases they will be asked to pay far less than the sticker price.

44. Aggregating data for all state systems, a report by the State Higher Education Executive Officers tells us: “In 2010, state and locally financed educational appropriations for public higher education hit the lowest level ($6,532 per FTE [full-time equivalent enrollment] in constant 2011 dollars) in a quarter century.… This downward trend continued in 2011 with state and locally financed educational appropriations at $6,290 per FTE, a decline of 3.7 percent over 2010 in constant dollars.” The report adds that appropriations per FTE would have been even lower, “except for budget driven enrollment caps in some states and reductions in state financial assistance.” (State Higher Education Executive Officers, State Higher Education Finance, FY 2011, 2012, 19; see also ibid., 20, figure 3.)

45. See Sandy Baum, Charles Kurose, and Michael S. McPherson, “An Overview of Higher Education,” Future of Children 23, no. 1 (2013).

46. See Rajashri Chakrabarti, Maricar Mabutas, and Basit Zafar, “Soaring Tuitions: Are Public Funding Cuts to Blame?” Federal Reserve Bank of New York, September 19, 2012, http://liberty​street​economics​.newyork​fed.org​/2012/09/soaring-tuitions-are-public-funding-cuts-to-blame.html#.UFnZJAauLvQ.twitter, which reports that increases in net tuition at public institutions have been associated with decreases in state and local appropriations since 2007.

47. See Sandy Baum and Jennifer Ma, Trends in College Pricing (New York: College Board, 2012).

48. Ibid., 7.

49. See State Higher Education Executive Officers, State Higher Education Finance, FY 2011.

50. Joseph E. Stiglitz, “Debt Buries Graduates’ American Dream,” USA Today Weekly, International Edition, July 13–15, 2012. Survey data confirm that over the past four years, students have started to foot an increasing share of their families’ total expenditures on college. Between the 2008–09 and the 2011–12 academic years, the share of family college expenditures paid for by parents’ borrowing, income, and savings has fallen from 40 percent to 37 percent, at the same time as the share of the expenditures contributed by the students’ borrowing, income, and savings has risen from 24 percent to 30 percent. See Sallie Mae and Ipsos, How America Pays for College (Newark, DE: Sallie Mae, 2012), 8; in addition, the pie chart on p. 7 of this report shows the share of expenses paid from various sources, including savings, grants and scholarships, contributions by relatives, and borrowing by both students and their parents.

51. The results of the most recent administration of the National Survey of Student Engagement (NSSE) are similarly encouraging in this regard: about three-quarters of students in their freshman and senior years agreed that college was a good investment. See National Survey of Student Engagement, Promoting Student Learning and Institutional Improvement: Lessons from NSSE at 13: Annual Results 2012 (Bloomington: Indiana University Center for Postsecondary Research, 2012), 17.

52. See Sallie Mae and Ipsos, How America Pays for College, 14, 40. This is not the place to review the vast literature on returns to education, but I believe many commentators (including, unfortunately, many of those speaking for colleges and universities) put too much emphasis on purely economic returns, important as they are. Years ago, in the midst of the depression of the 1930s, no less a figure than the conservative Chicago economist Frank Knight cautioned against over-emphasis on the virtues of what he called “the business game.” He observed: “However favorable an opinion one may hold of the business game, he must be very illiberal not to concede that others have a right to a different view and that large numbers of admirable people do not like the game at all. It is then justifiable at least to regard as unfortunate the dominance of the business game over life, the virtual identification of social living with it, to the extent that has come to pass in the modern world.” (Frank H. Knight, The Ethics of Competition [New Brunswick, NJ: Transaction, 2009], 58.) Also see Sandy Baum, Jennifer Ma, and Kathleen Payea, Education Pays, 2010 (New York: College Board Advocacy & Policy Center, 2010), for a useful summary of the benefits higher education confers, on individuals and society in general, beyond earnings effects.

53. Economists are strong believers in revealed preferences, and it would be most helpful to see what students actually do, not simply what they say they want to do or even what they say that they will do. In contemplating research on such questions, it is important to recognize that it is not enough to focus just on “average” net tuition. Much depends on the distribution of both need-based aid and merit aid by type of student and type of institution attended. What is required is detailed data at the student level that can be connected to institutional data—and such data are hard to obtain for defined populations of students attending different kinds of colleges and universities in various states. National panel data are generally too highly aggregated to serve this purpose.

54. See Josh Mitchell, “Student Debt Rises by 8% as College Tuitions Climb,” Wall Street Journal, May 31, 2012. This article cites data from the Federal Reserve Bank of New York; the online edition also presents a vivid graphic showing the decline in credit card debt alongside the rapid growth in student debt.

55. Andrew Martin and Andrew W. Lehren, “A Generation Hobbled by the Soaring Cost of College,” New York Times, May 12, 2012, online edition.

56. See Sandy Baum and Michael McPherson, “The New York Times Blunder,” Chronicle of Higher Education, May 17, 2012. As Baum and McPherson point out, citing Sarah Turner, the source of the error was incompetent analysis of Department of Education data (failing to understand a skip pattern and ignoring correct data which was supplied to the authors by the Department of Education). What is most disconcerting is that the number reported in the Times article didn’t pass any semblance of a “smell test”; Baum and McPherson surmise that the “story seemed to be striving for maximum drama rather than for an accurate picture of student debt and the very real problems it creates for too many students.” An even deeper lesson to be gleaned from this fiasco is that there is a terrible lack of sophistication among many journalists (though certainly not all) covering higher education—a point that Nicholas Lemann, dean of the School of Journalism at Columbia University, has made repeatedly.

57. Baum and McPherson, “The New York Times Blunder.”

58. Matthew M. Chingos, personal communication, May 13, 2012.

59. Christopher Avery and Sarah Turner, “Student Loans: Do College Students Borrow Too Much—or Not Enough?” Journal of Economic Perspectives 26, no. 1 (Winter 2012): 1–30.

60. Ibid., 25.

61. See Bowen, Chingos, and McPherson, Crossing the Finish Line, 163–64. Similarly, the most recent NSSE results showed that about 32 percent of freshmen and 36 percent of seniors reported that financial concerns had interfered with their academic performance; this proportion rose to almost 60 percent for full-time seniors who worked twenty-one or more hours per week. About three in ten freshmen and seniors also said they often chose not to purchase required academic materials because of their cost, and more than four in ten reported they often chose not to participate in an activity due to lack of money. (National Survey of Student Engagement, Promoting Student Learning, 17.)

62. This figure comes from the Baccalaureate and Beyond data set, collected by the Department of Education’s National Center on Education Statistics, which provides data on the cumulative student loan balances as of 2009 for the graduating class of 2008. (See also Mitchell, “Student Debt Rises by 8%,” and Jennifer Cohen and Jason Delisle, “Focusing the Student Loan Conversation on the Average Borrower, Not the Average Loan,” Ed Money Watch, New America Foundation, May 15, 2012, http://edmoney.newamerica.net​/blogposts​/2012​/focusing​_the​_student​_loan​_conversation​_on​_the​_average​_borrower​_not​_the​_average​_loan-6.) NCES data also show that, of students who began their undergraduate education in 2003–04 and who had attained a certificate, associate’s degree, or bachelor’s degree by 2008–09, 62 percent had borrowed for their undergraduate education, and the average cumulative amount borrowed by those students was $21,700 (in 2012 dollars). For more information about student debt, see Christina Chang Wei, Lutz Berkner, and C. Dennis Carroll, Trends in Undergraduate Borrowing II: Federal Student Loans in 1995–96, 1999–2000, and 2003–04, NCES 2008–179rev (Washington, DC: U.S. Department of Education, National Center for Education Statistics, Institute of Education Sciences, 2008), http://nces.ed.gov/pubs2008/2008179rev.pdf.

63. Richard Fry, A Record One-in-Five Households Now Owe Student Loan Debt, Pew Research Center, September 26, 2012, www.pewsocialtrends.org/2012/09/26/a-record-one-in-five-households-now-owe-student-loan-debt/.

64. See Jesse Rothstein and Cecilia Rouse, “Constrained after College: Student Loans and Early Career Occupational Choices,” Journal of Public Economics 95, no. 1–2 (2012): 149–63, cited in Avery and Turner’s “Student Loans.” As part of the College Cost Reduction and Access Act of 2007, the government enacted an income-based repayment (IBR) program for students who have high levels of debt relative to their income and/or who are pursuing careers in fields with relatively low salaries, such as public service. IBR caps students’ monthly repayments on federal student loans according to their discretionary income level. The maximum repayment period under this program is twenty-five years, after which students’ remaining debt is forgiven. The 2007 law also established a loan-forgiveness program particularly for students who pursue careers in public service. Under this program, any remaining debt is discharged after borrowers have worked full-time in public service for ten years and have made 120 monthly payments on an eligible Federal Direct Loan. Unlike IBR’s twenty-five-year forgiveness program, the ten-year public-service forgiveness program is taxfree. For more information, see Mark Kantrowitz, “Income-Based Repayment,” FinAid, 2012, www.finaid.org/ibr; and Mark Kantrowitz, “Public Service Loan Forgiveness,” FinAid, 2012, www.finaid.org/loans/publicservice.phtml.

65. See Bowen and Bok, The Shape of the River.

66. Ibid., introduction to the paperback edition, especially p. xxxix.

67. Matthew M. Chingos, Michael S. McPherson, and I would like to claim some credit for this shift in emphasis from enrollment to degree completion. See Crossing the Finish Line.

68. See Coffin Eaton, “At White House Meeting on Affordability, a Call for Urgency, Innovation, and Leadership,” Chronicle of Higher Education, December 5, 2011; “Reining in College Tuition,” New York Times, February 3, 2012. Several of the university leaders present at the White House meeting as well as the Times editorial board agree with President Obama that “the federal government must do more to rein in tuition costs at the public colleges that educate more than 70 percent of the nation’s students” (“Reining in College Tuition”).

69. In 2009, for instance, only about 40 percent of 25- to 34-year-olds in the United States had attained some form of tertiary education, giving the United States a rank of sixteenth in the world, according to the Organization for Economic Co-operation and Development. By contrast, among adults between the ages of 55 and 64, the United States’ rate of higher-education attainment was also about 40 percent, giving it a rank of third in the world for this age group and making it virtually the only G20 nation whose rate of attainment had not grown between the older and the younger cohorts. See Organization for Economic Co-operation and Development, Education at a Glance, 2011, September 13, 2011, doi: 10.1787/19991487.

70. See studies by Sean F. Reardon at Stanford and Susan M. Dynarski and Martha J. Bailey at the University of Michigan, among others, in Whither Opportunity? Rising Inequality, Schools, and Children’s Life Chances, ed. Greg J. Duncan and Richard J. Murnane (New York: Russell Sage Foundation and Spencer Foundation, 2011). A recent article in the New York Times offers poignant anecdotes of three young women in Texas whose precarious financial circumstances, in combination with other challenges in their personal lives, have interfered with their ability to earn four-year college degrees. The article cites research by Chingos showing that even students from low-income families who earn higher test scores than their wealthier classmates still complete college at lower rates, as well as Reardon’s finding that the difference between high- and low-income students’ scores has grown by 40 percent in the last quarter century. (Jason DeParle, “For Poor, Leap to College Often Ends in Hard Fall,” New York Times, December 23, 2012.)

As Jeff Selingo has warned, enrollment caps in states such as California may be at least as serious a problem as reductions in appropriations, since some students, especially those from modest backgrounds, may be deprived of any in-state enrollment option in the public university sector. (Jeff Selingo, “For Have-Nots, the Rockier Road to a College Degree Increases the Appeal of Alternatives,” Chronicle of Higher Education, March 23, 2012.) Selingo worries that enrollment caps and other large disruptions in the higher education system “could worsen the divide between the haves and have-nots.” He cites two examples of such disruptions, including Western Governors University—a nonprofit institution at which students work through course materials, many of which are available online, at their own pace, and earn credit by demonstrating their mastery of course content on standardized examinations—and MITx—an organization offering not-for-credit online courses taught by MIT professors, currently free of charge, to the general public. Selingo reports that he has never heard any of the critics of the value of traditional higher education say that “they’d surely send their kids to Western Governors University or choose a certificate from MITx over a degree from nearly any four-year college.”

71. See William E. Kirwan, “The Research University of the Future,” speech at AAU Public Affairs Network Meeting, Washington, DC, March 22, 2010. Kirwan went on to note: “We have, of course, experienced periods of fiscal decline in the past, one as recent as the early part of this decade. But, this decline has a different character. In the past, economic downturns were followed by periods of economic boom and losses were recovered relatively quickly. I know no one who predicts that will be the case with our current fiscal decline.” Kirwan then commented on “the disconnect between the aspirational rhetoric at the national level and the reality on the ground” by observing that, in a single week, “President Obama announced his laudable goal for leadership in higher education completion rates and Charlie Reed, Chancellor of the California State University System, announced that Cal State was turning away 30,000 students this spring because of inadequate funding.” This under-appreciated talk is well worth reading in its entirety. Gary Fethke and Andrew J. Policano agree with Kirwan that “the diminished role of state government funding is permanent” (Fethke and Policano, Public No More: A New Path to Excellence for America’s Public Universities [Stanford, CA: Stanford University Press, 2012], 218). For another sobering assessment of what might happen to colleges and universities, see Jeff Selingo, “The Fiscal Cliff for Higher Education,” Chronicle of Higher Education, August 12, 2012. Selingo describes a possible “death spiral” for some institutions, and he seems to be referring especially to lower-rated private colleges—which are definitely threatened by increased competition from lower-priced educational options. I suspect that my colleagues and I, in our focus on the large public university systems, have paid inadequate attention to the problems facing the regional private institutions.

72. See William J. Baumol, The Cost Disease (New Haven, CT: Yale University Press, 2012), especially chapter 4.

73. Kevin Kiley, “Californians Approve Measure That Will Avert Major Education Cuts,” Inside Higher Ed, November 7, 2012.

74. Jonathan Medina, “California Cuts Threaten the Status of Universities,” New York Times, June 1, 2012.

75. See Baum, Kurose, and McPherson, “Overview.”

76. See Dee J. Hall and Samara Kalk Derby, “Gov. Scott Walker Unveils Agenda for Wisconsin During Speech in California,” Wisconsin State Journal, November 19, 2012. Other states that, with encouragement from the Obama administration, are using performance measures rather than enrollment as the primary determinant of higher education funding include Indiana, Ohio, and, most prominently, Tennessee. (See Joanne Jacobs, “More States Utilize Performance Funding for Higher Education,” U.S. News, February 24, 2012.)