Marginalism and the Higher Ed Paradox

By now, you may be getting sick of reading articles and blog posts about the crisis in higher education. This post is different. It proposes an explanation of why students have been willing to pay more and more for undergraduate and professional degrees at the same time that these degrees are becoming both less scarce and more dumbed down. And that explanation rests on a simple and plausible economic hypothesis.

First, let me dispose of the idea that “college (and business school) is all about signaling.” The explanation I present allows signaling to represent a major part of the value of higher education, but it says that the historical increase in willingness to pay for education is not caused by an increase in its signaling value. (And the evidence for signaling or screening education premia, as opposed to human capital accumulation, is pretty thin anyway.) I’m certain signaling plays a role in creating value for certain degrees from certain institutions for certain people in certain situations. That it dominates the value proposition for college seems like a stretch.

My hypothesis is that it is precisely the dumbing down of U.S. education over the last decades that explains the increase in willingness to pay for education. The mechanism is diminishing marginal returns to education.

Typical graduate business school education has indeed become less rigorous over time, as has typical college education. But typical high school education has declined in quality just as much. As a result, the human capital difference between a college and high-school graduate has increased, because the first increments of education are more valuable on the job market than the later ones. It used to be that everybody could read and understand something like Orwell’s Animal Farm, but the typical college graduates could also understand Milton or Spencer. Now, nobody grasps Milton but only the college grads can process Animal Farm, and for employers the See Spot Run–>Animal Farm jump is more valuable than the Animal Farm–>Milton jump.

So the value of a college education has increased even as its rigor has declined, because willingness to pay for quality is really willingness to pay for incremental quality. This principle holds true in many markets. For example, a roof with mean time to failure of 5 years is a lot more valuable than one with a MTF of 2 years, but a 25-year MTF isn’t that much better than a 22-year MTF for most owners. A fuel economy increase from 12 to 15 miles per gallon is a bigger deal than an increase from 27 to 30 MPG.

Empirical points in favor of this diminishing marginal returns/reduced overall rigor hypothesis:

1. Rigor appears to be declining over time at all levels of American education.

2. Rate of return evidence classically suggests that the big marginal gains to education come from lower levels of education.

3. The median wages of college graduates have been flat, but the median wages of high-school-only graduates have gone down even more.

4. The MBA market has continued to support higher tuitions and enrollment despite the secular trend in rigor.

5. Employers increasingly favor those with more education even as they complain more about the quality of the graduates they hire.

Additional implications:

1. The incremental human capital gained from attending a (truly) better school rather than a typical school is increasing, since the additional learning is more basic (and hence more valuable) than it used to be.

2. Five and six-year undergraduate-to-masters programs should grow to accommodate those who would benefit from additional human capital.

3. More-rigorous high schools will attract larger premia (in either tuition, ability to be selective, or, for public schools, their impact on local property values), because at lower overall levels of rigor the increment of human capital is worth more.

Extensions of the logic to signaling considerations:

1. If you accept that the marginal ability and effort necessary to acquire education increases in the level of education (the flip side of the assumption about diminishing marginal payoff), then the signaling value of the typical degree is actually declining. The innate ability difference between the college and high-school-only graduate shrinks as both curricula are made less rigorous.

2. Signaling by the quality of the institution attended and the difficulty of the major subject studied is becoming more important; a very selective (or hard to complete) school or major adds back some of the lost signaling power of the typical degree.

3. We should see college degrees becoming more important in occupations that wouldn’t seem to “require” them under the old model of college, such as service staff in food service and hospitality jobs.

48 Comments on “Marginalism and the Higher Ed Paradox”

  1. Rich Makadok says:

    Steve, I like your hypothesis. What do you think of Taranto’s theory that it’s an artifact of the Griggs case? See:


    • stevepostrel says:

      The Griggs hypothesis is a signaling/screening theory rather than a human capital theory. It may hold some water, although it’s hard to believe that an expensive university degree is the most efficient legally feasible alternative to an IQ test. Many employers in finance supposedly look at SAT and GRE scores on resumes, for example, and couldn’t someone just get an acceptance letter from a college and show that to an employer?

      But the main advantage of my hypothesis is that it explains the simultaneous dumbing-down and pricing-up of higher ed. Signaling theories don’t do that very easily.

      • John Thacker says:

        I think you’re conflating rigor of what is actually taught (and thus the human capital gained) with the strength of the signal. But for the signaling hypothesis (without human capital gained playing a strong role), the strength of the signal depends largely on the rigor of *admissions* rather than that of the curriculum.

        But while the rigor of what is taught seems to have declined, I’m not certain that the rigor of the admission process has actually declined. Over the long term I think it has become *more* merit based at the same time as rigor has declined. (Including both in an academic sense and in simply being willing to perform other activities desired by admissions officers.) The extreme is something like the Japanese university system, which is widely believed to be a system where students study like mad for admission and then simply enjoy their university years, which are easy by contrast.

        There are reasons that employers would prefer something besides *just* SAT and GRE scores. SAT and GRE scores don’t measure being conscientious and a team player, things that are in many ways as important to many (most?) employers than simply intelligence. The value of a PhD to employers is often not what someone actually did, but merely that they could finish something and work to a deadline. As far as finance goes, stories abound about the most important thing being the candidate being a “good fit” and having the right sort of extracurriculars.

        Note that if being conscientious and a team player are important to most employers, then that creates a strong bias against students looking for a non-traditional way around the educational system. I think that the signaling model is easily consistent with both dumbing-down and pricing-up. I’m not sure that it’s really all that true, though.

  2. Steve Phelan says:

    I like your hypothesis, good job!

  3. I like this hypothesis — it sounds like empirical studies that don’t account for the change in HS content is a classic case of omitted variables bias, since the decline in HS content is unmeasured but correlated with the decline in higher-ed content.

    Even more powerfully, the decreasing-returns argument says that the returns to the “new” higher ed (and not just the students’ WTP) should be more than linearly higher (reversing a concave transformation). Given that many more people are seeking education, however, we might expect that more intense rivalry will transfer the quasi-rents from “lower higher ed” to employers — generating a higher social return to education but a lower private return.

  4. amerkhan says:

    Reblogged this on Class(ic) Stories and commented:
    Key implication: a business degree is still sought after, despite dumbing down of the curriculum, because the additional learning or human capital development that is perceived to be brought about by that degree, even dumbed down, is stilll highly valued by the employers. Very interesting! Read the original blog to get the full argument.

  5. John Thacker says:

    I believe that every one of your implications is equally satisfied by a signaling model that uses the fact that more people are attending college, and that admissions processes are getting more rigorous at the same time as curriculum is not.

    The high school graduation rate rose steadily until 1980:

    It was the high school graduation rate reaching that steady state around 85-88% that decreased the high school degree’s value as a signaling mechanism. Signaling theory predicts that, therefore, after 1980 we should expect a fall in the wages of high school graduates. Since only 30% or so of the population 25-29 has a bachelor’s degree, it still has signaling value. The rate is still increasing, which has caused people to look for more significant signals.

    It is thus inaccurate to say that the signaling gap of the high school degree and the college degree has decreased; rather, it has clearly increased. The high school degree pre 1980 had much more signaling power than now.

    • stevepostrel says:

      John: Given the indent limit above, I’ll try to address both your comments here.

      1. The education signal could have two components–getting admitted and getting through in good order.

      2. To the extent the signal is dominated by getting admitted, any equivalent sorting mechanism that used fewer resources would be preferred by all parties. Hence, the idea of a free-floating admissions office or just using test scores.

      3. To the extent the signal is dominated by getting through, difficulty and/or boredom would seem to be the keys to sorting.

      4. Since high school has no admission standard, any signaling a median degree supplies is entirely focused on getting through.

      5. We suspect that high school has become less rigorous over time, based on a smattering of available data. As you point out, that may make it easier to graduate, hence reducing the signaling impact of a high school degree.

      6. But if college has also become less rigorous, as we have reason to believe from other data, then the incremental signaling impact it offers has also gone down, since as you point out, getting through and graduating is an important part of the signal. So the net signal of college over high school has not increased. That’s especially so if you believe in increasing marginal difficulty of advanced education–the cutbacks in college rigor have a bigger incremental impact than the dumbing down of high school because the marginal difficulty at college would have been higher at the old rigor level.

      7. If one tries to argue that the admissions office is making up for this by tightening the filter on who gets in, one runs into two problems. First, as argued in 2 above it doesn’t make sense that admissions offices could charge so much for the pure signaling effect. Second, it is empirically false, in that more students than ever are getting admitted to college despite no improvement in high-school graduates. (As I mentioned in the original post, the most selective schools and the harder majors should get a bigger signaling premium than they used to.)

      8. In any case, the diminishing-marginal-returns human-capital theory explains the facts fairly simply and cleanly.

  6. David Hoopes says:

    Nice post and a good explanation. I agree that signalling is present but a thin explanation.

  7. @mdryall says:

    Hi Steve,

    Nice post! You have definitely added an important idea into the discussion about the future of our industry: the fact that college may indeed be the new high school, the incremental value must be measured in incremental terms. Economics comes to the rescue again!

    That said, even adjusting for this important insight, my sense remains that at some point in the not-too-distant future, the markets will correct for grossly overpriced education services. When the marginal student’s absolute benefits from education are so low that he or she has no hope of paying off his or her student loans for services provided by suppliers facing a high-fixed-cost-almost-zero-marginal-cost supply technology that appears to have over-built capacity …, well, we know what happens to pricing in that situation. We are already seeing the signs as some marginal law business schools declare bankruptcy. If government pulls the plug on subsidizing student loans, my guess is things will go downhill very quickly.

    Questions raised but not addressed by your analysis are why academic standards have slid downhill — after a long period of improvement — and, presumably based upon the answer to that question, where the new equilibrium will wind up. For example, is parental pressure on high schools is partially to blame for the drop in standards at that stage? If so, by your reasoning, they are making themselves strictly worse off by taking actions that force them to substitute cheap high school education with expensive college education. Why would they do that? Perhaps there is a free-rider story at work (given current standards, I want my kid to get a degree with top grades regardless of effort … but, everyone doing this results in lower standards).

    Interesting stuff!


    PS: I’m not sick of reading about this. It’s important and, even if it weren’t, it’s interesting.

  8. Great piece, Steve! Sad I already handed in my book as I talk a fair amount about cost issues.

  9. Daniel says:

    It’s certainly true that more college going explains why people keep paying more and more (since it’s now so common it’s necessary but not sufficient for professional jobs), but this would be true no matter what the quality of the actual education looked like.

  10. […] Reihan, this is an excellent blog post.  Rather than excerpt, let me reproduce the whole thing: By now, you may be getting sick of […]

  11. […] Reihan, this is an excellent blog post.  Rather than excerpt, let me reproduce the whole […]

  12. Candide III says:

    Great idea! One thing I missed in your post, however, is a list of potential empirical points against your hypothesis. What does one look for in order to disprove it? Also, your Extension #1 assumes, along with most discussion of educational signaling, that what it is signaling is (largely) education and/or ability. But if much of the signaling is about indicating class (cf. admissions boards for the better colleges considering much more than academic merit, concentrating on extracurricular activities (the more refined the better), charity/voluntary work etc.) then the signaling value of a degree need not correlate very strongly with its academic content. In fact, hiring at big companies has began to focus on extracurriculars too. There was a paper about that recently, but unfortunately I have mislaid it.

  13. Mike M says:

    As a novice to this topic, I would have assumed that the decrease in education at all levels is a result of increased supply, both within the US economy and across the world. Not knowing the difference between signal theory or other structured investigations into this area, would someone be willing to explain why the simply supply/demand argument is invalid?

  14. […] Further Material: Bryan Caplan on signaling. David Autor on Signaling Tyler Cowen on Signaling Hanson on Signaling A good intro article Survey Paper on Returns to Education Stylized facts about the labor market Agricultural yields as an instrumental variable Marginal returns to college might be going up, because high school is so bad these days. […]

  15. […] Cowen sends us to this very interesting post by Steve Postrel on higher education: My hypothesis is that it is precisely the dumbing down of […]

  16. MikeMcK says:

    Fascinating: Three-fourths of the examples meant to illustrate an overall decline in academic rigor are anecdotal (Maddox painfully so), so one might think the initial premise of this post is unsupported, HOWEVER their use by a university professor does show – at least in some cases – there is a clear lack of rigor in higher ed.

    And by the way, there are some “rigor” issues with Academically Adrift:

    • stevepostrel says:

      This post has blog-level rigor, not academic publishing rigor. It’s pretty difficult to compare levels of academic difficulty over time, and of course one would need to be much more precise about exactly what is meant by rigor if a formal study were to be performed. In particular, one would need to distinguish the aspiration level of the curriculum from the achieved population distribution of learning and from the achieved population distribution of retention of that learning.

      As I am proposing a hypothesis, I am only trying to render it plausible, not claiming to have proved it. I can tell you that Mike Ryall’s concerns about the declining rigor of MBA education are shared by everyone I’ve spoken to about it, from a range of instructor age cohorts. If someone has evidence that today’s college students are just as well-educated on comparable writing and thinking skills as those 10, 20, 30, or 40 years ago I’d certainly be interested.

      • MikeMcK says:

        I’m not sure I disagree with the premise, but I haven’t seen consistent evidence that overall academic rigor has declined at the K-12 level. Let me give one (admittedly anecdotal) example:

        “Some days I don’t much enjoy Jeopardy because on those days I don’t know many of the answers. For example, teenagers snap out answers on math, geometry, astronomy, and chemistry, while I sit there without a clue.”

        The author here is saying that today’s teenagers learn more STEM content than she did when she went to school 60 years ago. I realize students smart enough to make it into the Teen Tournament aren’t a representative sample, but this source is just as valid as the ones you cite in the original post. In fact, it’s one of the original sources. The above quote is from another Maddox post ( In this post she again laments the lack of Shakespeare in HS curriculum, but now says that literature has been replaced with science and math. If accurate, the original article does not describe a decrease in overall rigor – only a transfer in what is being studied. She is upset that _her_ subject is understudied, when compared to the past, not saying current students are doing less hard work.

        Again, I’m not sure I disagree with you and I like your hypothesis as an explanation IF rigor is declining. I’m just not sure that it is (or isn’t). The point is, we really don’t know.

  17. […] From Steve Postrel: My hypothesis is that it is precisely the dumbing down of U.S. education over the last decades that explains the increase in willingness to pay for education. The mechanism is diminishing marginal returns to education. […]

  18. […] many students are making a seemingly irrational choice? In a post at strategyprofs, Steve Postrel explains here that while it may be true that college degrees may be becoming more common and watered down in the […]

  19. […] Read it. My hypothesis is that it is precisely the dumbing down of U.S. education over the last decades that explains the increase in willingness to pay for education. The mechanism is diminishing marginal returns to education. […]

  20. SporkHero says:

    Intriguing hypothesis.

    The 80’s saw a big premium jump. 1979-1986 college wage premium increase for graduates with 1-5 years’ experience was 45 percentage points for white men, 30 for white women, 27 for black men, and decreased by 3 points for black women (Eide 97 Southern Economic Journal)

    So did disproportionately black public schools deteriorate less over the period (perhaps they’d deteriorated pre-79?)

  21. bob sykes says:

    Your thesis is flawed. Early in the 20th Century, very few people attended high school. They were an intellectual elite, and they got elite educations. The quality of high school educations has declined because every goes to high school. Today, many high school students are simply too stupid to understand any written text and are functionally illiterate. It’s all about IQs.

    The same process is underway in our colleges and graduate schools. The utterly lunatic idea that everyone should get a college degree will eventually destroy these degrees. Today, even some PhDs are if not actually illiterate so mal-educated as to be superstitious and delusional. Some of them even achieve high political office.

    • MikeMcK says:


    • Candide III says:

      Yep. That’s what happened in the Cat City by Lao She. Some 80 years ago he observed contemporary China and saw the endgame:

      When modern education was being first introduced, schools had grades and students were evaluated based on their knowledge, but gradually examinations were abolished (they were seen as primitive) and students graduated from school even if they never went to classes.

      However, high school graduates and college graduates had unequal privileges. High schoolers became unhappy with this situation, arguing that they studied no less than college students. Then a drastic reform was carried out, and all students graduated from college [Note: all ‘students’ graduate at the top of their class] on the day of their enrollment into primary school. Then… sorry, there was no “then”. What “then” can we hope for?

      There are … squabbling “young scholars”, who pose and talk in foreign languages “so that nobody understands them. They don’t understand what they’re saying themselves, but they enjoy the lively atmosphere that all those foreign sounds create.”

  22. Great post!
    We discussed the post at our law-and-econ lunch last Thursday. A diagram with a concave production function of education output and quality-units input would help.

    A lot of our discussion was on how education quality has declined. The time frame does matter a lot. We focussed on since the 1960s. Some impressions:

    1. High school science and math has gotten better at the top high schools. But those kids all go to college.
    2. Everything has gotten worse at the bottom high schools.
    3. For both reasons the gap between suburban and inner-city Chicago has widened.
    4. There’s a lot more homework than when we were kids, but mostly busywork.
    5. There’s less work at school— more films, games, crafts, and such.
    6. There’s less discipline in schools.
    7. It isn’t at all clear that college quant subjects have gone down in quality or rigor in the older colleges, tho with increasing college education there are many low-quality new colleges for the low-quality new students.

  23. […] sure a Yale economist could formalize this with set theory, but the idea is pretty straightforward. This is from Steve Postrel: My hypothesis is that it is precisely the dumbing down of U.S. education over […]

  24. stan says:


    Regarding your point about understanding Animal House in the context of changes in education, you seem to be saying that the more we try to make sure that all humans are equal, the more it seems that some humans are ‘more equal’ than others.

    • stevepostrel says:

      I was thinking more about Manor Farm than Faber College, but I’m sure that Snowball and Blutto would agree that “Knowledge is Good.”

      I don’t have a crisp explanation for the hypothesized dumbing down, but misguided egalitarianism could be a factor.

  25. […] education, in the midst of at least an impression that its rigor has been reduced? Here’s a new theory. It has a feel of plausibility, at least. Worth thinking about, anyway. Like this:LikeBe the first […]

  26. richard40 says:

    So basically, it used to be that employers could assume that if somebody graduated from HS, they could read, write, and do simple math. But now they cant assume that anymore, and thus basic required education, even for simple jobs, is inflated to a 4 yr college degree, and even there you cant really be sure. Really encouraging assessment on the effectiveness of our education system.

  27. […] 12 Reasons College Costs Keep Rising, and related “Marginalism and the Higher Ed Paradox” i.e. as the quality of  higher education declines, it’s necessity and cost rise.  […]

  28. […] a Yale  economist could formalize this with set theory, but the idea is pretty  straightforward. This is from Steve […]

  29. […] a Yale economist could formalize this with set theory, but the idea is pretty straightforward. This is from Steve […]

  30. […] for secondary education. VN:F [1.9.13_1145]please wait…Rating: 0.0/10 (0 votes cast)ReferencesMarginalism and the Higher Ed Paradox […]

  31. […] a novel explantion of the tremendous increase in the willingness to pay for higher education: My hypothesis is that it is precisely the dumbing […]

  32. says:

    I actually was initially searching for suggestions
    for my own blog and uncovered your own blog, “Marginalism and the Higher
    Ed Paradox”, do you really care in cases where I actually utilize some
    of your tips? With thanks ,Cory

  33. How much time did it take you to publish “Marginalism and
    the Higher Ed Paradox”? It offers a
    large amount of wonderful info. Thanks a
    lot ,Silke

  34. Haeloxspace says:

    […] Postrel, ”Marginalism and the Higher Ed Paradox,”, June 19, […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s