Day 3: Why does college cost so much?

A series following my new class: The Economics of Higher Education

Apr 26, 2025

Welcome to Day 3 of our 18 day series!

“But wait,” you notice, “Days 1 and 2 were adjacent in time but this Day 3 post is late. I thought she said the Block Plan meant class every weekday for 3ish hours a day. It is Saturday. This post was supposed to come out on Wednesday. Where has she been?”

It turns out I can’t keep up with class AND launching a new experiment AND managing page proofs for another article AND attending to service work AND posting this series “on time.”

So I’m just going to publish each day of this series when it’s ready. Onward!

On Wednesday of Week 1 we started talking about the thing everybody wants to know most about college: why the bleep does it cost so much?

Today’s readings

Archibald, R. B., & Feldman, D. H. (2011). Why does college cost so much?. Oxford University Press. Chapters 2, 4, and 5.

Chapter 2: Is Higher Education All That Unusual?

Prices have been increasing in higher education. This chapter asks: what about other industries? After all, institutes of higher education compete with other firms for workers and customers, they also purchase equipment that could be used in other businesses, and technical progress can affect higher ed just as it affects the markets for other goods and services. These authors ask two questions: “First, are there other industries whose costs and prices behave similarly to higher education? Second, if there are such industries, is their common behavior driven by the same factors?”

Archibald and Feldman investigate these questions by plotting the real (adjusted for inflation) average price of output in various industries, with 1970 as a base year (i.e. when the index equals 1). If prices increased at exactly the rate of overall inflation, the line would be flat. If prices rose faster than overall inflation, the line is upward sloping. If prices decreased relative to overall inflation, the line is downward sloping.

First, the authors show how prices in higher education compare to prices of durable goods in the US (new cars, furniture, and table settings). Prices in higher ed rose faster than inflation between 1940 and 2010, while prices for durable goods fell.

The data on these prices come from the NIPA accounts which are conveniently available in FRED!1 This means it’s fairly easy to recreate A&F’s plots using updated data. My students got started on these in class and then I cleaned them up a bit for you.

Check out what’s been happening recently. The updated plot makes clear that the trend A&F point out has persisted: the inflation adjusted price of durable goods has been falling, while the inflation adjusted price of higher education has been increasing.2

A&F create an analogous plot for gasoline, food, and women’s clothing to show that as with durables, the prices of non-durable goods like food and clothing have been decreasing. While the prices for some goods (gas) are volatile, the price of higher ed is steadily increasing.

These patterns also persist in the updated plot.

Finally A&F look at what’s been happening with the price of seeing a doctor, dentist, or lawyer. Ah ha! they say. Now this is interesting. The price of higher ed closely tracks the prices of physician, dental, and legal services.

A&F say that this plot reveals one of the primary points of their book: “the price of higher education will behave similarly to the prices of personal services offered by highly educated service providers.” Therefore, increasing prices in higher ed must have something to do with the price of employing highly educated workers.

What do we see in the updated plot? Well, the updated plot looks a bit different than what A&F observed. The updated plot shows that the price increases for all of these highly-educated service providers were tightly clustered through the early 90s. After that, higher ed broke away from the pack. The price increases of dental and professional services have remained tight together, while price of physician services started slowing and then falling relative to inflation in the early 2000s. (You can see the start of that trend in A&F’s original plot.) Prices in higher ed were increasing faster than these other services, but then even they leveled off starting in 2015 and have started falling since COVID.

A&F say that the price of higher ed is tied to the cost of employing highly educated workers. But that only leads to another question. Why has the price of highly educated workers increased, and why might the price of these workers be falling lately?

Chapter 4: The Costs of Employing Highly Educated Workers

Workers in higher education are much more likely to have college and graduate degrees than workers in other industries. So when the wages of college educated workers go up, A&F say, so do prices in higher education. Let’s look at some plots of the relative wages of college educated and high school educated workers. We in the biz call this the “returns to college education.”

This next figure from A&F comes from data in Goldin and Katz’s 2008 book The Race between Education and Technology. The returns to a year of college education were falling from 1914-1940, steady from 1940-1980, then increasing from 1980-2005. That middle period from 1940-1980 is what Goldin and Margo (1992) call the Great Compression. “In 1940, an American male at the 90th percentile of the income distribution (only 10 percent of working men earned more) earned five times as much as a man at the tenth percentile. By 1950, that gap has shrunk to a factor of three. … Goldin and Margo estimate that almost half of the compression was due to falling returns to schooling. … By 1999 the gap between the 90th percentile and the 10th percentile of earnings for male workers had risen to 5.4. … Over the last thirty years (up to 2011) the less educated have been getting relatively poorer, and the well educated have been getting relatively richer.”

Why did the returns to education go up? Skill-biased technology change. The story goes something like this.

First technology changed to allow assembly lines, and high-skilled artisans were displaced by low-skilled factory workers. This depressed wages for high-skilled workers and increased wages for low-skilled workers. Later, technology changed again, away from assembly lines and into more complex machines that could do batch and continuous processing. These complex machines required more skilled labor to install and maintain; more demand for skilled labor pushed wages for skilled workers up. Later still, the assembly lines themselves were upgraded to incorporate robots, which again required more skilled labor to invent, install, and maintain them, which again pushed wages for skilled workers up.

Essentially, the argument is that the kinds of technology advancement we saw from 1980-ish to 2010-ish were complementary to the skills that college educated workers provide. So even though overall rates of college-going have increased dramatically (for men, and especially for women) pushing the supply of college-educated workers up, the demand for college-educated workers has been so strong due to skill-biased technology change that overall wages for college-educated workers went up.

The generation of workers in 1975 were “significantly more highly educated than the prior one, and this caused a general ratcheting up of living standards. In addition, surging surplus supplies of ever more highly educated labor had contributed significantly to a narrowing of income differentials between the highly educated and the less educated. In a sense, the United States had it all: rapid economic growth that benefited everyone and which supported a burgeoning middle class that could afford to send its kids to college at rate unimaginable a generation earlier.

“What has happened in the years since is that educational attainment has slipped, and skill-biased technical change has won the latest round. Skill-biased technical change clearly creates winners and losers. By pushing up the demand for highly educated labor, new technologies have changed the distribution of wages in ways that favor the well educated. The whole national income distribution gets stretched in ways that have been described as hallowing out the middle class. Today’s (2011) college wage premium stands at just about the same level in 1915.”

I find the end of this chapter super interesting to think about as we look around at politics in the US since 2010. Clearly there’s an educational divide in our politics that’s been brewing for a long time (see for example Polarized by Degrees). At the same time, people who have gotten college degrees are less and less satisfied with how things have turned out for them, many saying that they regret their major choice or don’t use their degree on the job. Hence all the talk about canceling student loan debt. People often frame this issue like college has gotten too expensive; another way to frame that conversation is that the wage premium college graduates were expecting hasn’t materialized.

And THAT, my friends, is our discussion for Day 4.

Chapter 5: Cost and Quality in Higher Education

So far this book has argued that the reason why higher education has gotten more expensive is primarily because the people employed in higher education are getting more and more expensive to employ. Readers of this Substack will know that I am skeptical of that argument. This last chapter of the day adds that other input expenses in higher education have gone up too, for two reasons.

First, the kinds of things we’ve always done now require more expensive equipment to do. Computers, projectors, first wired internet and then wifi, ever more advanced laboratory equipment, etc. The authors show a plot of the ratio of the real value of equipment compared to the real value of buildings, and show that it started increasing in higher ed the 1970s. This was true across the whole economy, but especially true in higher ed. (My students tried to find the underlying data series for this figure so we could make an updated version, but we couldn’t figure it out in the 30 minutes we allocated in class.) And of course, more technical and advanced equipment requires more technical and educated workers, which pushes the wage bill up in the non-instruction parts of a university or college too.

Second, institutes of higher education are now asked to do more things, both in the classroom and out of it. Some people might say that students are expected to graduate with more or different knowledge these days, and those kinds of knowledge may be more difficult to transmit and therefore require more and more expensive inputs. A wider variety of classes. More expert professors. More teaching assistants. More technology in the classroom. More field trips outside of the class room. Smaller class sizes. Student-faculty research collaborations. All of that leads to a higher instruction cost per student. (Note: I don’t see the share of expenditures spent on instruction at liberal arts colleges increasing since 2000 in the data. Nor do I see professors getting more expensive.)

Universities and colleges also now offer more support systems like career services, counseling centers, and accessibility offices. All of those additional people cost money. (This is what I think we actually see in the data since 2000.) This book doesn’t mention it, but I’d add to this list nice dorms, good food, quality gyms, more sports, and of course lazy rivers.

As Caroline Hoxby pointed out in our readings for Day 2, institutes of higher education exist in markets and since the 1940s they’ve compete on quality. If other colleges are offering updated science labs, well-connected career counselors, free therapy, and treadmills with AI coaches, well then by golly my school will too.

So what HAS been happening with that college wage premium? That is our topic for Day 4.

Tomorrow’s readings: The Returns to Higher Education

If you like this series so far, please consider subscribing or leaving a comment! I’m doing this because I like hearing from and contributing to my extended community. I’d love to know what you think about today’s topics!

The Liberal Arts Ledger

Discussion about this post

Ready for more?