Two economists offer a historical perspective on rising college costs.
By Robert B. Archibald & David H. Feldman • Illustration by Tim Robinson
In 2010, we wrote Why Does College Cost So Much? to help demystify and depoliticize this hyper-charged topic. Rather than look for examples of waste by poring over the accounts of individual colleges, we chose to take a broader view of higher education to see if the behavior of our industry matches the predictable patterns of many other industries. This aerial view places higher education in the context of the entire U.S. economy and the economic history of the past century, offering a perspective often missing from discussions of the forces driving the tuition charged for higher education.
College Cost versus College Tuition
Any discussion of higher education's finances should begin by distinguishing between cost and price. Cost is what a college spends to provide education. Price, under the name of tuition and required fees, is the amount that students are charged.
For most businesses, the relationship between cost and price is simple. Price has to exceed average cost or the firm will not be able to stay in business. Not-for-profit colleges and universities are different. They are subsidized in the form of donor gifts, endowment earnings and, for public institutions, appropriations from state governments. These subsidies allow colleges and universities to stay in business despite charging a tuition that does not cover average cost.
Gifts, endowments and state appropriations are not the only subsidies involved. Colleges and universities have a published tuition, but they do not charge all students that price. They give scholarships and grants to some students.
All of these subsidies make it very difficult to interpret changes in tuition. When we see an increase in tuition and fees in the new college catalog, we cannot be sure what has happened. Has the cost of providing the education increased? Have gifts, endowment earnings or state support gone down? Or have the scholarships and grants to individual students gone up, so that only some students pay higher prices?
The Big Drivers of College Cost
Index of Real Higher Education Costs. (1970=1) 1948-2008.
Figure 1 lays out the costs of providing a higher education from 1948 to 2008. This is an average for all types of colleges and universities, public and private. The data in the figure are corrected for inflation, so when the line slopes upward it means that higher education costs are going up faster than the overall inflation rate. When it slopes downward, higher education costs are going up less rapidly than inflation.
We see three distinct periods. From 1948 to the mid 1960s, higher education costs rose more rapidly than the inflation rate. Then from the mid 1960s to the early 1980s the rise in higher education costs moderated and actually rose more slowly than overall inflation. Finally, after the early 1980s, the costs again began to rise faster than inflation.
The natural first question is whether this pattern is unusual. We discovered that it was not at all uncommon, that other industries whose cost behavior looks like higher education share many common features. These "like" industries include dentistry and medicine, legal services, life insurance and bank service charges, to name a few. The historical pattern of rising cost in these industries is not identical to higher education's pattern, but the similarity over long stretches of time is so striking it rules out simple coincidence. We uncovered three factors all such industries share.
First, all of these industries provide services instead of physical goods. Service prices almost always rise more rapidly than the prices of goods. Department of Commerce National Income and Product Accounts data show that the price index for services rose by a factor of 12 between 1947 and 2012, while the price index for goods rose only 4.57 times over the same time period.
The reason for this difference is well known. Technological progress over the last century has dramatically raised productivity in the manufacture of products like automobiles (more cars produced per hour worked) or in the production of agricultural commodities like wheat (more bushels per acre). This productivity growth holds down cost. Generally it is more difficult to increase labor productivity in the service sector, and this is especially true in personal services where quality is strongly determined by human interaction. Few people want to go to the fastest barber in town, and aspiring students tend not to seek out the college that crams the most students into each class. The problem with the usual measures of labor productivity (stuff produced per hour worked) is that they do not effectively capture changes in service quality over time.
The second similarity between higher education and the services of dentists, physicians and lawyers has to do with the education level of the service providers. Barbershops and colleges both offer services, and both have had fairly stagnant labor productivity, but the price of a haircut has not risen as rapidly as college tuition. The reason is that the labor force of a barbershop and the labor force of a university have very different average levels of educational attainment. The gap in earnings between someone with a college degree and someone whose education stopped with a high school degree has widened substantially since 1980. Labor market data show that the rate of return to a year of college, which had bounced around just under 10 percent from 1940 to 1980, climbed steadily after 1980 to over 15 percent.
The third similarity these education-intensive personnel services share is how they use technology. Most organizations choose to adopt new technologies if the new hardware, software or machinery allows them to cut costs without cutting quality. This is not how technology decisions are made in higher education. Colleges and universities have to meet an educational standard of care. We must prepare students to be productive citizens in today's workplace and in the world of the future as best we can see it.
If we taught our chemistry and biology students using techniques and equipment used in the 1960s, this might indeed be cheaper, but we would be guilty of educational malpractice. These students would be unable to function in the laboratories of a modern biotech firm. This story is easy to tell about sciences and other STEM fields, but the same process is at work in every academic field. It's also at work in all of the other things a university does for its students, from academic support and career advising to psychological counseling and student health. Colleges and universities often have to invest in new technology, even if it drives up the cost of education. A similar process has been at work in many other industries. Just think of how a doctor's office functioned 30 years ago, and of how surgical procedures have evolved over the same time frame.
These three factors: (1) higher education's status as a service, (2) its comparative reliance on highly educated service providers, and (3) its need to meet a standard of care, explain why college costs tend to rise more rapidly than the inflation rate. These factors explain the time path of cost in higher education and many other service industries.
Alas, this isn't a particularly sexy story of rising cost. There aren't obnoxious villains making poor or selfish choices, and there isn't something obvious to cut.
Drivers of Price
These three factors are all cost drivers. But changing forces in the higher education funding model also affect price. As states have backed away from supporting colleges and universities, and during times when endowment earnings have been meager, tuition and fees — price — have had to rise more rapidly than costs. The fact that in most states appropriations per student have not kept up with higher education's costs explains why in the last 30 years tuition and fees have risen more rapidly at state-supported colleges and universities than they have at private colleges and universities.
The final driver of list-price tuition and fees is the increased use of tuition discounting. If other income — from donor gifts or endowment earnings, for example — does not increase, an institution has to increase the list-price tuition to non-scholarship students if it wants to give larger scholarships or aid to other students. If it did not, it could not meet its needs for revenue and the quality of the education provided to all would decline. Colleges and universities in both the public and private sectors have increased the use of grants and scholarships considerably in the last 30 years. Many private institutions have discount rates close to 50 percent. Widening income differences between families in the middle of the distribution and families at the top drives some of this discounting. As discount rates have risen, the difference between the price printed in the catalog and the average tuition and fees actually paid by students has increased.
Other Factors — the Dysfunctionality Argument
What then are we to make of the other explanations? What about supposed administrative bloat, wasteful country-club amenities, government subsidies that raise demand, and the low workload of tenured professors? We lump these factors together because they are all part of a claim that colleges and universities are dysfunctional institutions.
We would be the last people to claim that colleges and universities are efficiently run organizations. Some administrations are surely over-staffed, and some tenured professors don't pull their weight. Every college should continually re-evaluate its practices and procedures looking for ways to streamline operations. But it is hard to believe that the historical pattern of tuition increases in higher education is driven by progressively rising inefficiency. Even if all inefficiencies were cleaned up, college costs would still tend to rise more rapidly than the inflation rate. The three factors we highlighted above would still be at work.
Let's dig more deeply into the dysfunction claims. The fraction of the average college staff classified as administrators has indeed increased rapidly in recent years. But if you get behind the numbers, you will find that much of the increase in administrative staff comes from rising numbers of computer support technicians needed to keep a modern institution running. Information technology staffs have grown in higher education just as they have grown in the rest of the economy. In higher education, this portion of the rise in administrative staffing is part of the standard-of-care argument given above. The larger IT staff has little effect on the number of students taught, but a big effect on what goes on within the institution.
A second area of rapid growth in staff has been in university advancement, i.e. fundraising. If the advancement office is doing its job, added staff more than pays for itself, so these administrators help provide the subsidy that keeps tuition in check.
If tenured professors are a real difficulty, the higher education system should be well on the way to fixing the cost problem. Tenure is rapidly dying. In 1993–94, 62.6 percent of institutions of higher education had tenure systems. That percentage fell to 45.3 percent in 2011–12. Even at four-year private institutions only 43.7 percent of the full-time faculty had tenure in 2011–12. And many institutions have instituted continuing review for tenured faculty. Unfettered lifetime tenure is more myth than reality.
Anecdotes about pampered students dining on sushi and living in luxurious accommodations make excellent copy for those who think that colleges are dysfunctional. Yet there is an odd double standard and considerable ignorance at work behind these claims.
Room-and-board charges have indeed risen more rapidly than inflation. Corrected for inflation, from 1969–70 to 2011–12 room-and-board charges have risen by 171 percent at public institutions and 165 percent at private institutions. This should surprise no one. The rise in inflation-adjusted room and board tracks fairly closely the rise in the inflation-adjusted price of off-campus housing and dining options offered in the private sector. But yes, there are more choices on the menu, and the amenities available in dormitories — like the number of bathrooms per floor — are better. The average student today has also lived in a larger house and eaten out far more often than the average student in the 1970s. Expectations are different. Despite stagnant incomes over the last decade, notably for those not among the top wage earners, the average American family is far better off than its counterpart 40 years ago.
Meanwhile, unlike airlines, we don't really have first-class and second-class options for room-and-board plans. If a school offered a bare-bones Spartan alternative for poorer or more price-conscious families, it could probably cut the average cost of room and board by a small amount. The fact that schools don't do this suggests that their approach may be a practical response to student and family expectations.
Finally, many discussions of rising tuition place the blame squarely on federally sponsored financial aid, particularly federal loan programs. This analysis is based on a simplistic reading of supply and demand taught in any Econ 101 class. The argument is that financial aid increases demand, and when demand goes up prices tend to rise because of the extra costs of creating places for more students. Using economic jargon, the supply curve slopes upward.
There is a certain irony in two economists questioning basic supply and demand, but the simple textbook model really does not work for higher education. For smaller schools, adding students might actually lower the cost per student served. Many schools, for instance, could probably add a hundred students without needing to build a new library or new classrooms. Once a school attains a larger size, the costs of expansion tend to correlate directly to the size of the expansion, moving demand and costs more in a straight line than a curve.
In plain language, these facts mean that federal subsidies tend to raise the number of students attending colleges and universities, but they do not drive up the average cost of educating those students. So it is actually no surprise that there were no changes in the trend of growth in tuition at the critical times when federal aid programs were initiated or expanded. Federal financial aid started in 1965, and was ramped up dramatically in 1972. The effect was a burst of enrollment, not a sudden rise in the growth rate of college cost.
The Beneﬁts of a College Education
Perhaps the most devastating attack on higher education today is the claim that a college degree does not pay off anymore. We also hear that earning a college degree is "no longer a guarantee of success."
First, there is actually a growing body of evidence that a college education pays off in significantly better health outcomes, more community involvement and better job and life satisfaction. We might expect people with higher incomes to be healthier and happier, but this conclusion holds true even when adjusting for the higher income that college graduates earn on average.
Second, frightening claims about high student debt are often based on individual stories about people whose behavior is extreme, while low payoff rates on loans happen typically at schools that are not at all representative. In 2013, the College Board reported that for private colleges the average debt of students who borrow is $29,900. This is lower than for the previous year, and roughly the same as it was in 2006, before the Great Recession. The "asset" acquired with that borrowing pays off: on average, bachelor's degree recipients earn $400,000 to $800,000 more in lifetime earnings (measured in today's dollars) than students who achieve only a high school diploma.
It is true that new graduates' earnings have stagnated in recent years, along with median family income. And we have seen significant increases in student borrowing and repayment struggles. These problems largely reflect fallout from the financial crisis and slow recovery in the U.S. over the past six years. Despite the fact that a degree isn't quite as golden as it was a decade ago, as a financial proposition earning that degree remains one of the best investments a person can make.
But is it a guarantee of success? In fact, it isn't. But going to college never came with an economic guarantee. The real question is about economic life chances. It turns out that earning that degree has become an ever better bet over the last 30 years.
Percentage of College Graduates Earning More Income than the Median High School Graduate, 1964-2010.
We can actually measure this using Census Department data. For young people between the ages of 25 and 40, there is a wide range of earnings. Some high school graduates get high-paying jobs while some college graduates wind up in low-paying jobs. For each group, however, the distribution of earnings looks very different because the median earnings of a college graduate are so much higher. In Figure 2, we estimate the percentage of college graduates who earn more than the median earnings of high school graduates. We chose this measure because it shows the overlap in the earnings distributions. The evidence is clear. The chances that a young college graduate will do better than the median income of high school graduates has been rising since the mid-1970s.
The Last Word
To conclude, much of the discussion about higher education is overhyped and alarmist. College cost tends to rise more rapidly than the inflation rate because higher education is a personal service industry whose primary employees are highly educated. It is also an industry in which technological progress — and we have experienced a lot of that in the past 40 years — must be embraced, even if it increases cost.
This said, college leaders are quite concerned about the growing gap between median family income and net tuition and fees. But for that there is no magic technological bullet or simple policy remedy. For small private universities like Bucknell, the challenge is to find ways to use new digital technologies to raise the quality of the learning environment, while slowing the growth in cost. This will not be easy. The supposed online revolution is not a panacea, and much of the value created at colleges and universities relies on close human interaction among students, and between students and their teacher–scholar professors. Nonetheless, it's fair to say that in years past passing cost increases on to families whose incomes were rising rapidly allowed schools the luxury of not making hard cost choices. Those days are over.
Robert B. Archibald and David H. Feldman are professors of economics at the College of William and Mary. Last December, Bill Gates named their book, Why Does College Cost So Much? (Oxford University Press) as one of the "best seven books I read in 2013."