Monday, June 11, 2012

College is mostly about human capital, not signaling

Human capital formation, Stanford style

This is not an area I know a lot about, but I have some thoughts.

Some people think college is about signaling, not about building human capital. I'd say most economists believe this. But what exactly do they think is being signaled?

Intelligence? No, that doesn't make sense. It's way too easy to tell who's smart. To signal general intelligence, all you need to do is take some tests - AP tests and SAT 2's if you're trying to signal moderate intelligence, International Math Olympiad if you're trying to signal exceptional mathematical intelligence, and so on. You don't need 4 years at an elite school to show you can do some math problems or memorize some stuff.

In fact, in Japan, most employment decisions are based on exactly this sort of signal. High school students who want good careers spend all of high school studying for some really long college entrance exams, and employers basically pick the students who get the best scores on these exams. Yes, colleges also make their decisions based on those same exams. But Japanese college basically provides zero additional signal, because A) Japanese college kids do very little work, and B) Japanese employers don't even look at college grades. Whatever Japanese people's reason is for going to college, it isn't for signaling intelligence. No reason America should be any different.

So what else could people be signaling at college? The ability to do hard work in the face of massive leisure temptation? I actually think this is a reasonably big deal, especially in the U.S. If college performance is a signal, it's a signal of people's desire to study when they could be partying. (Update: Here's a signaling defender who agrees.)

But this doesn't explain Japan. In Japan, college students do mostly party. In fact, that is what you are supposed to do at college, especially at a top school like Tokyo University. It is encouraged. Fun fact: Many Japanese people call college "moratorium". As in, a moratorium on work.

Now, this could indicate that college is actually about consumption. Well, to some degree, it is; college is fun. But people tend to smooth consumption, and college is all concentrated at one time, so it doesn't make sense that college would be mostly about consumption.

This leaves human capital. Economists (including at least one in my PhD graduating class) have often tried to show that college doesn't produce useful skills. But I think that this is missing the point; useful skills, which you mostly learn on the job, are not the only valuable form of human capital. There are three extremely important forms of human capital that you can't acquire on the job:

1) Motivation,

2) Perspective, and

3) Human networks.

These, I believe, are the types of capital that college is designed to build, both in Japan and in the United States.

First, motivation. Motivation is widely recognized as the scarce quantity, or "limiting reagent" in an individual's human capital. This is the source of all those annoying motivational poster slogans. "Attitude is everything". "Your attitude determines your altitude". Or as Calvin Coolidge said:
Nothing in the world can take the place of Persistence. Talent will not; nothing is more common than unsuccessful men with talent. Genius will not; unrewarded genius is almost a proverb. Education will not; the world is full of educated derelicts.
These sayings are cheesy and annoying, but they are true; skills mean nothing if you don't have a reason to put them to work. Everyone knows that the biggest long-term threats to the career of any scientist or engineer are "burnout" and clinical depression. In poor countries, the threat of poverty provides ample incentive for hard work, but in rich countries - i.e., the countries where most smart people go to college - motivation is a more ephemeral thing.

What motivates smart people in rich countries? Here is where I start to conjecture, but my bet is: Human relationships. Friends and family. We work hard for other people. When we're young, we work because our parents want us to work. When we leave the nest, however, we need to find other relationships to motivate us. I can't tell you the number of people I've seen who excelled in high school under the watchful gaze of a Tiger Mom, only to lose much of their motivation once Mom was no longer looking over their shoulder (I can, however, tell you the name of one such person: Noah Smith).

To replace Mom, young adults need to form new relationships. Close friends. Romantic partners, and eventually a spouse (which in turn leads to kids, another motivator). But it is very difficult to form these relationships fast (which you need to do in order to start a career fast) without sacrificing quality; if you're just randomly searching, it takes a long time to find friends and a lover who really click with you, especially if you're a smart person who clicks best with other smart people.

This is where college comes in. College is an intense incubator where smart people meet other smart people. The large number of leisure activities and the close quarters in which people live facilitate the formation of friendships and romantic relationships, while the exclusiveness of college makes sure that the people you're meeting are pre-screened to be the type of people with whom you are most likely to click. In the U.S., the "college experience" includes parties, trips, clubs, athletic events, religious fellowships, communal drug use, study groups, endless late-night conversations, and more esoteric events like the one pictured above. In Japan, it includes "go-kon" (group blind date) parties, "nomikai" (pub nights), and clubs. American college works better, but it's much the same sort of thing.

The friendships and (especially) romantic relationships people form in college are a great motivator. That is an incredible boost to human capital.

Second, there is "perspective". This is about learning the set of possibilities for life. Before I went to college, I never knew people who went into the finance industry, or joined tech startups, or worked for the World Bank, or did sound engineering for movies, or taught English in foreign countries. In college I met people who did all of the above, and seeing them taught me a lot about the set of possibilities for human life. Simply knowing one's career choice set is a hugely important part of choosing the right career. And it's surprisingly hard to do. College is a great way to gain career and life perspective; if you go from high school straight to the workforce, you are basically assured of not meeting as diverse a group of high achievers.

That is part of human capital. It's much more important in America than in Japan (where careers are less differentiated and people switch careers only rarely). But it is very important. And it is more important for poor people, who grow up mostly seeing other poor people.

Finally, there are the human networks built by college. I won't talk a lot about this, because other people have done so quite a lot, and there are a bunch of scientific papers about it. And it's pretty obvious, just from looking at MBA programs, which cost more than college and are well known to be all about professional networking. (Note: As a commenter points out, this is actually called "social capital".)

All three of these types of human capital have to do with bringing people together. It is not something that can be built online. It is not something that can be built on one's own, or in high school. There is a reason why college is called "college", meaning "gathering".

So anyway, I don't want to drag this on too long, and I've made the basic points. College is useless as a mechanism for signaling intelligence. It's probably somewhat useful for signaling the ability to work hard and resist temptation, at least in the U.S. where many colleges require hard work (but not in Japan). It is about consumption, but it's too concentrated in time to be mostly about consumption. College is really about human capital, of the kind not conveyed in classes - motivation, perspective, and networking. Rather than a hideously, inefficiently expensive signaling mechanism, college is an ingenious technology for building the kinds of human capital that are scarce among smart people in rich countries.


  1. Anonymous3:04 PM

    Is that how Americans spend 4 years of their life and a lot of money? I went to university to learn and get a degree (for some people was mostly about the second one, and for some about the first). Not to party, network, get a gf, be motivated (you do have a point about poor countries there), or whatever.
    There was a fair amount of signalling, because of course you wouldn't be tested again by a potential employer, they'd just assume that if you got a degree you know your stuff. But it is mostly about the most efficient way to learn what you need to get a job. Reading from a book at home is much more difficult and lacks the signalling power.

    1. In my experience, people who have a lot of motivation tend to take it for granted. Like being able to move your legs. Once you go through a period of severe de-motivation, you learn how absolutely essential it is. Partying, networking, and getting a girlfriend may sound like a waste of time and money to you, but if I'm right, it's an absolutely essential component of rich-country smart-person motivation.

      Also, in my experience, poor-country people lose a lot of motivation once they hit middle-class income levels. They get a job that pays a decent salary and they think "Yeah, I made it!!" But rich-country kids grow up with that. Their parents want them to do better than that, and society expects them to do better than that. Or at least not to do worse! At higher income levels, motivation becomes much trickier...

    2. Apparently there is no like button, but I think this is a fantastic point! Many believe their is a crisis occurring due to the loss of the financially inclined "immigrant mentality", but a collegial environment (pressure from your professor, peers and general campus) can definitely supplement the motivation previously derived from the immigrant mentality.

  2. I dunno, Noah . . .

    A College degree is a minimum requirement for a lot of entry level jobs now. That wasn't necessarily true when I graduated in 1968, and wasn't true at all a few years earlier.

    In that context, what is the degree, other than a signal?

    And isn't it true - most especially in Finance - that WHERE you got your degree is a big consideration? And isn't that a signal?

    As for networking - I have a BS, an MS, and an MBA, and literally almost never saw, heard from, or thought about any of those class mates again after any of my graduations.

    But maybe things have changed over the decades.


    1. Yeah sure, but all the signaling is done at the time of college admissions. Getting into college, not graduating or getting good grades, is what tells an employer you're smart. That signaling is done in high school. If that were all college was about, people wouldn't bother spending four years there; they'd just take the tests, do whatever currently gets you into Harvard, and send a list of their accomplishments to employers instead of to Harvard.

      Sorry to hear you don't keep in touch with the people you met in school...but with email and Facebook there's always a chance to go back and re-connect!

    2. Anonymous8:53 AM

      Good first effort, but needs further thought.

      "College is useless as a mechanism for signaling intelligence."

      I believe you have overstated your case here. The evidence you present suggests that college is redundant as a means of signaling intelligence. You haven't shown it to be useless. PhD programs, in fact, seem to take college performance very seriously, and faculty search committees take PhD performance very seriously. There are other intensely academic endeavors for which college performnace is a key qualifier. (DARPA, Google, etc)

      Jazz raises an institutional point. If hiring decisions require college, then something is being signaled. The dichotomy you present is between signaling and human capital. Uhhh, that may not be a good dichotomy. The point Jazz raises suggests we may be signaling the development of human capital. What you seem to argue is not that college is not a signaling mechanism, but that it is not a signaling mechanism for intelligence, but rather a signaling mechanism for human capital.

      To substitute tautology for jargon, educational success may be taken as a signal that one is educated. How much we focus on educational success depends on how much we care. Harvard is a stickler for educational achievement in its faculty hiring decisions. JP Morgan less so, but still pretty interested. Walmart wants to know its white collar workers have a degree - associate is good enough at some levels.

  3. Anonymous4:09 PM

    Employers requiring SAT scores would most likely face Title VII suits under disparate impact (if it could be shown a protected class was losing employment opportunities due to this and the employer was unable to specifically tie SAT scores to employment duties). Unless certain employers only asked for a certain section of the SAT that related to the employment.

    So, I don't think the signalling argument fails on that point alone.

    But, I think the networking argument is strong. In professions where networking skills are potentially important, degrees that augment networking skills are preferred.

    That's likely why the cashier or the manager at the grocery will likely not have a 4-year degree, but the grocery store director may.

    1. Employers requiring SAT scores would most likely face Title VII suits under disparate impact

      That was what I was going to say, and why I stopped reading so soon.

      Noah_Smith, if you're going to criticize the literature, please, read it first.

    2. Hey great sass there Silas_Barta, except for 426 U.S. 229 (1976), and the case law of the last thirty-five years. WELL DONE SIR

    3. Anonymous10:35 AM

      r money,

      That was a Due Process case, not Title VII. Completely different.

      --Orig Anony

    4. Anonymous10:54 AM

      Employers can't ask about SAT scores, but they can ask about which college you attended, /which is very nearly the same thing/. If you think about it, this supports Noah's point, albeit somewhat differently. The SAT test is still the signal sent, just muddied through college.

      Oddly, for attorneys, employers do care about LSAT scores, in addition to school attendance.

    5. No, it supports the signalling hypothesis: you go to college as the only legal way to send the signal (intelligence via SATs). If they could test for it directly, we wouldn't have to go through (as much of) wasteful processes like college.

    6. Anonymous3:02 AM

      Employers certainly do ask about test scores. I know every banking/consulting application I've filled out asks for SATs and GREs

    7. Anonymous10:11 PM

      actually, recruiters at Google ask for GPA, SAT, GRE, and any other quantifiable scores. They even verify them. At least they did when I was interviewing seven years ago.

  4. Anonymous4:19 PM

    Don't mean to be nit-picky but isn't #3 social capital, not human capital?

    1. Yes--and that's why a college degree is more like a guild membership or a social class marker.

  5. "Some people think college is about signaling, not about building human capital. I'd say most economists believe this."

    I remember this discussion well in undergrad econ.

    It really depends on your major. There's no way you can say an accounting major, who learns enough in college to get a CPA, didn't gain a ton of valuable human capital -- and human capital that's very time and effort consuming to obtain.

    Someone who majors in poly sci and then never uses it in his career, then much less human capital, much more signal, but still he learned how to learn, how to write better, some computer skills,...

    I'm not at all sure most economists believe this. I think most believe it to an extent, but may still believe that especially in some majors, there's still a great deal of human capital built.

    1. I've always found it humorous that econ professors so often believe it's mostly signaling. What does that say about their value as professors?

    2. And what does it say about the efficiency of the market?

      You can claim that markets are somewhat efficient.

      You can claim that college is mostly about signaling.

      You certainly can’t claim both (i.e. why wouldn’t someone hire cheaper labor without degrees).

  6. Anonymous5:02 PM

    "As for networking - I have a BS, an MS, and an MBA, and literally almost never saw, heard from, or thought about any of those class mates again after any of my graduations." - JzB

    Maybe so - but maybe part of the skills you learned during your MBA was how to find good contacts. While you were networking with classmates, you were learning how to network well...hopefully. So while you may have lost touch with those classmates, you have gained skills to find better contacts.

    1. Sure, Anon -

      But that is neither signaling nor human capital. It is developing a skill set, which (presumably) is the purpose of education.

      I got my job after college as a result of my chemistry degree. That was symbolic of certain technical and thinking skills I developed under the instruction of my professors.

      Doesn't technical ability and mastery of subject matter mean anything?

      This discussion simply ignores them.


    2. Developing a social skill set is human capital. And since you didn't seem to develop strong social bonds, it's clearly not social capital, though I would say it does signal future social capital. I'm never quite sure why economists attempt to dichotomize building capital and signaling. I'm someone that needs the capital to feel confident in the signaling. I know it's partially a personal failing, but my signaling is always an extension of my personal skills and experiences.

  7. "This leaves human capital. Economists (including at least one in my PhD graduating class) have dedicated untold numbers of papers to showing that college doesn't produce useful skills."

    Again, there are obvious exceptions to this, the ability to pass the CPA exam, from college, to pass the bar, to gain the basic technical skills to go on and be a scientist. If you're going to be a mathematician, how are you going to do it without the basics you learn as an undergrad? You think the calculus, linear algebra, statistics, etc. you learn for your B.A. are worthless and unnecessary for a mathematician? You think the computer languages and techniques a computer science major learns will never be used in his career, and are useless, or could be learned in a few hours on the job?

    It looks like there's too many big clear exceptions to this. It depends on the major and what career is gone into after. Some people get giant human capital from college.

    1. "You think the calculus, linear algebra, statistics, etc. you learn for your B.A. are worthless and unnecessary for a mathematician?"

      Well as someone who did a math undergrad I'll say "Yes" for the most part. Linear Algebra, Calc, Stats? Okay that's first year what do you do for the next 3 years? You take increasingly abstract courses consisting of endless "Theorem-Proof" based lectures covering a broad area of mathematics. If you're a professional mathematician you're only going to focus on one tiny area within mathematics and 95% of what you learn in undergrad will soon be forgotten or become fuzzy. The best you can say about a math degree is it helps to train you to think abstractly but, as in most degrees, the majority of the content is actually not important.

    2. Also as for Computer Science - what CS program actually teaches you specific programming languages again? None I've ever heard about. I'd say a CS major who goes on to be a Software developer gains the vast majority of actually useful skills out of 3-4 of their courses (the typical intro to programming course, a course on algorithms, a databases course and maybe something on automata theory) the rest is largely academic filler.

    3. As someone who got a PhD in math and was planning to be a mathematician for many years, I'll say no! This specifically applies to the US where you learn much less math as a math major than you would in Europe (where vast majority of your undegrad classes would be math).

      Maybe what you say is somewhat true about the qualifier-level PhD courses but that is one level up and not true for the top mathematicians.

    4. To go meta here ...

      Sometimes you learn things in college that are required to enter a certain field. You need to take a lot of accounting courses to pass the CPA exam.

      But how useful is the CPA exam? Does it really test what is necessary to be a good CPA? Or does it test what was covered in your accounting courses?

    5. Roger,

      This is another good issue. I have a few friends who became actuaries and took a lot of mathematical statistics courses in school. According to them these were really helpful for passing the Society of Actuaries exams but the content of the exams themselves are pretty irrelevant to the actual on-the-job work of an actuary you just need to pass them to move up in the profession.

    6. Anonymous10:19 AM

      I would like to point out that in order to take the CPA exam a certain number of accounting class hours are required. So, the fact that people who pass the CPA exam have taken a certain number of courses in accounting is not proof that the courses are required to pass the CPA exam. I believe that statisticians call this bias (i may be wrong about what it is called though).

  8. Even in MBA school, I attended Michigan, I learned a great deal that helped me in business and investing. It wasn't just networking and learning to work well with other smart people with big egos. If you really wanted to learn, you could become a much better businessperson. Lots of great lessons.

    You could have learned a lot of these lessons on the job with experience, but that could be extremely costly, learning by making mistakes and doing things sub-optimally when millions are at stake. If you could learn some of these things at college it would be a lot cheaper.

    1. And really you can see, in the language and actions, that a lot of managers are applying what they learned as MBAs, and at a top school from some really smart, experienced professors, and adjuncts who were CEOs.

    2. I think you're right, of course. Classes do build real skills. Being taught by top physicists taught me how to think. And many classes gave me important ideas for life. Also I learned Japanese which was key. My point in this post is that even apart from classes, college builds human capital that is incredibly valuable and cannot be replaced by distance education.

    3. I agree. But also there's the issue discussed above, how valuable is general learning, as opposed to specific, narrow, what you'll work on day to day, learning. For a computer programmer who's going to do very specialized work in a certain language, how worthwhile is it for him to learn other types of programming and languages in college?

      It's not an easy issue; if we only teach narrow specialties, who's going to have the important big picture, how things work together and interrelate. A lot of great ideas will be missed. On the other hand, many people will never use their general learning, and it will get largely forgotten. But having more general understanding, and having understanding of other things, can lead to great creative ideas. Not an easy issue. Not easy to see the value of general learning, when so many people then spend their careers working predominantly in a narrow specialty.

      And there's also the issue of citizenship. What is the value of the general education of college in making people better voters, with more of a national and global view, so they're less tribal, with some understanding of history and economics (although we could do way better by making sure things like externalites are taught in 101, not just how wonderful markets always are).

      And what is the value to people's happiness and quality of life of being able to broaden their horizons in college, instead of narrowly focusing on day to day specialized work skills.

      So, estimating the non-signaling value of college (or how that might be improved) is not easy.

  9. Anonymous5:34 PM

    Right now college is only semi-useful to people and is more about feeding the educational complex money than really benefiting people. College is no longer a necessity in the digital economy for all people because of the ease of entrepreneurship in the modern age. What might be considered accurate is that it is easier today to get a new business off of the ground than it was a few decades ago. Modern dynamic programming languages and frameworks, the ability to outsource the most monotonous of tasks from accounting to legal help to customer service, even social media (see BuyFacebookFansReviews for example to see the sheer number of companies that do nothing other than promote social media pages for startups), make it possible so that a lone person working part time can take on giants within important niches. That's the kind of thing that would not have been possible decades ago. So I think the continued thrust of technical growth is positive for entrepreneurship and it will be easier. But I think that it will always be much riskier to actually make the sacrifices necessary to launch a business and not everybody will choose to go that route. College is usually a distraction that takes away years of peoples most productive years. Yes, not everybody can be an entrepreneur, not everybody wants that kind of risk, but even without college, there can be things like apprenticeships. Right now kids that go into major debt to shove the real world away for 4-5 years are making a foolish choice IMO, but thats what the political system is designed to promote.

  10. This may be sufficient for "traditional" college students, i.e. those just out of high school who are just getting their feet wet in the job market, just starting to live on their own and build social networks, etc.

    But it also means that college, at least on its face, is an even bigger scam for "non-traditional" students that live off-campus, don't socialize much with their classmates, and possibly have already seen enough of the world to get plenty of perspective. For those guys who thought they were going to get an education that prepared them for higher-earning jobs, it turns out they really ARE just shelling out $50k minimum for a fancy piece of paper.

    1. See, I think the fact that employers require a college degree is just inefficiency in the system. Most employers don't care if you have perspective and networks, since those help you switch jobs just as much as they help you do your job well. As for motivation, if I'm right they can infer it from the fact that you're married.

    2. An interesting point. It used to be, in fact, that being married was an important signal for employers. Single men didn't advance, and a single executive was pretty much out of the question. Of course that was partly because the social capital of the COMPANY was mostly built by the wives. It was true at many universities, too. My mom felt totally caged in as a faculty wife in a southern town in the late 50's and early 60's. Fortunately we moved to Ann Arbor and she was able to go back to school. Got her Ph.D. in Economics just before the Friedman cult really got going.

    3. Anonymous3:04 PM

      Boss: Kirk, crackers are a family food. Happy families. Maybe single people eat crackers, we don't know. Frankly, we don't want to know. It's a market we can do without.

      Kirk: So that's it, after twenty years: "So long, good luck"?

      Boss: I don't recall saying "good luck."

    4. Agreed about the inefficiency in the system. It's repeated at higher levels in the private sector, when we go out and get various absurd little qualifications in nonsense things (Six Sigma Black Belt! Certified Project Coordinator!).

      I agree with your theory, by the way, I just lament that there's no alternative to college for those of us who go into it with what we'd be getting out of it already at hand.

  11. Anonymous5:44 PM

    In college you learn how to learn on your own. That is the most important skill.

  12. If it's your 1-2-3 (especially 2 and 3), these things could be done for a tenth or hundredth of the cost and perhaps in far less time. So at best the picture you paint is of an unbelievably inefficient college system.

    As an employer, I only care about the signalling, so I think you're wrong.

    1. Ha. Who said people go to college for your benefit? Employers don't capture all the value of an employee. The human capital people build at college will broaden their horizons and give them better options than working for you...

    2. In addition, what sort of business does Bret run which could be staffed by HS grads?

      (also, Bret should look for HS grads with high GPA's who aren't going to college, and hire them - they'd be cheaper than dirt).

  13. The comparison with Japan (and I've heard the same about China) seems particularly apt if, as Arum and Roska argue in Academically Adrift, college has become less difficult in recent decades. It may be more about partying, less about studying. If true, that would suggest that college is moving towards a signaling function, as in Japan.

    But you ask, why do employers need the signal of a college degree? Why not just look at scores? The selection process for prestigious colleges is fairly involved. It involves looking at more than scores (since many applicants have very strong scores) and is probably more than your typical employer would want to deal with. Think of how many applicants there are to Ivy League schools (for example), vs. annual graduates of Ivy League schools. Perhaps businesses have essentially "outsourced" the process of selection to college. Colleges perform what is, for employers, a free service--namely sifting through the multitude of good students out there. And students are willing to pay for the service because the payoff, if you succeed i.e. get into a prestigious college, can be large.

    As colleges become less prestigious, the value of the signal drops. But I think that's what recent grads are discovering. They've overpaid for something which they thought was more valuable than it actually is.

    1. It's not signaling intelligence or motivation, it's signaling social class.

  14. One problem I have with what seems like an over-emphasis on the value from signaling is that it treats degrees rather homogeneously. The employment market values a chemical engineering degree from the University of Houston differently than an electrical engineering degree from Cornell. Which of these two does Exxon hire as a chemical engineer? I'm willing to bet it's the Univ of Houston chemical engineering major. Yet who doubts the intelligence of the EE grad from Cornell?

    And I've seen no one make the case as to where the scales are tipped from human capital to signaling. Is it high school? Is it once you learn how to read? Is it once you learn how to use a computer?

    From the rhetoric, it seems as if some people think the value of college is almost exclusively signaling. As online education becomes more prevalent, I'd be interested in seeing if employers discriminate for or against those who get a large amount of their credits via online learning. Is an online learner exceptionally motivated? I know it's harder for me to get my work done in online classes, because the time management is left exclusively up to me.

    Or are online students cutting corners? Would an online education flag you as someone who sees the value in a degree as mostly a signal, and therefore didn't put much effort into developing themselves?

    I think it is also incumbent on the signaling extremists - to borrow from Bryan Caplan's "human capital extremists" phrase - to state what exactly it is that's being signaled. General intelligence? Intelligence in a specific sphere (mathematical, problem solving, etc.)? Familiarity with the concepts relevant to chemical engineering, to use my earlier example? It seems to me that at some point you're signaling - wait for it - the presence of human capital!

    1. Darren11:59 AM

      "I think it is also incumbent on the signaling extremists ... to state what exactly it is that's being signaled. "

      You missed an obvious one: Obedience.

    2. The Author, that is a major weak point here - I don't see that Noah has really proven, or even given evidence, that the signalling function is dominant. (personally, I believe that it's clear for Ivy League grads/Wall St, where the major seems to be irrelevant)

  15. I've thought about this a lot, and I have the added perspective of having taught in Japan, including teaching English to highly motivated Japanese HS and college students. I know that Japanese kids learn more in HS than the average American college kid, and it's a kind of trial by fire. If you get into Todai, as Tokyo U. is known, you've proven your metal. If you get into Waseda (maybe like Yale) and I forget the others for now, you've done what employers need to know, and yes, that's a clear signal.

    Also well established is that the tenure system in Japan (at least a decade or two ago) meant that professors weren't at the top of the game, compared to MIT, Yale, Harvard, Stanford, etc. So what could companies expect from that experience? Hence, just getting there and out of there was: the signal.

    As for the American experience, I like the hanging out with smart people theory. I once went to see an Indian guru, named, Swami Muktananda, speak. He was asked in the Q&A, "Why do we need a guru?" He answered (or his translator did), "It's easy to get sick. Hang out with sick people. It's also easy to get well. Hang out with well people."

    I made the connection: going to college was hanging out with well people. It worked for me.

  16. I personally think that college had a different effect on me because I became more sensitive about my social class. In regards to its usefulness, college is for people to get together and be enlightened. If people think that getting a degree will automatically get them jobs, then they're incorrect because degrees aren't cutting it anymore. Some employers won't hire you if you don't report your GPA on your resume. As silly as it sounds, between a 3.4 from Harvard and a 3.8 from Berkeley, the Harvard guy gets hired because of the brand name even when a GPA implies that the Berkeley girl is more qualified. Sociologists would say that degrees are important because we live in a credential society. However, much like the American Dollar, I have no confidence in degrees or certificates because there's been people who successfully fabricate credentials.

  17. sumsum1:36 AM

    I agree with you that what motivates people is human relationships. Almost all that we do, we view it in reference to other people. A long time ago I was watching an obscure movie on late night television. Susan Sarandon was in it and at one point says something to the effect that "we get married because we need a witness to our lives." I dont know if its profound or if its because I think SS is the most beautiful woman who ever lived, but it has stuck in my mind ever since.

  18. Anonymous2:02 AM

    "College is mostly about human capital, not signaling"

    Does it have to be just one or the other? Seems like most colleges do a lot of both.

    Also, I've long been of the opinion that most of the value from a liberal arts degree is external to society, and doesn't necessarily show up in the paycheck of the graduate but is real nonetheless.

    1. Does it have to be just one or the other?

      Of course. Check out this post, where I hypothesize that while human capital is the most important factor, signaling is probably happening too:

  19. Great post, Noah. I'm sold!

    1. This comment has just made my week. :D

  20. David C4:45 AM

    I am finding it interesting to compare this discussion with the educational system in Switzerland, where I now live. I earned a Ph.D. in science in the US, and managed scientists and technicians here.

    Only about 25% of Swiss pursue a university education. You have to work really hard as a young teen to get in (to pass the test). At least in the sciences, nearly everyone who starts university goes on to a Ph.D. Tuition is basically free, and your parents are required to support students until they are 30.

    What about the rest? The system is highly vocational. Nearly everyone has a diploma in some career area, from waitress to accountant to aromatherapist to engineer. The schools are highly specialized (no wasting time on Psych 101), and very practically focused on a specific job. Usually there is a apprenticeship involved that mixes classroom time with on-the-job experience. Less technical jobs require relatively short stints of post-secondary education (maybe 18 months). Highly technical fields could require 3 or even 4 years.

    As a manager, I loved this system. I could get highly trained employees right out of school, and I could be confident in the skills they had.

    It works pretty well for the young people, as well. I think Switzerland is better than Germany at supporting people who change their minds about careers mid-stream. There are lots of cross-over points between vocational and university tracks, the low cost makes it feasible to strike off in a new direction.

    I am not sure how this system plays into the "human capital" story. Certainly the things Noah mentions apply to the minority who attend university. But the majority, who have specialized educations, seem to do OK in motivation, perspective and networking.

  21. Anonymous6:44 AM

    As a student at a let's say top-10 MBA program I see the signaling function of my program to be the most powerful explanation for the value of my experience - recruiters from top consulting companies and investment banks come to campus and get the best students, while the rest go in to marketing or corporate finance, and all these recruiters know that they will find people who are at least adequate to the task. Yet interestingly maybe 20% at my institution strike it out on their own and do something different, requiring extensive use of a social network and giving weight to the human capital explanation.

    1. Anonymous7:03 AM

  22. Anonymous7:11 AM

    it seems to me the lots of people here got cheated on their education. Being a liberal arts graduate and lawyer, it is very easy for me to credit the schools I attended with: (a) giving me a deeper and more solid moral foundation; (b) giving me exposure to the skills deployed by greater thinkers, on many different subjects; (c) appreciating what can really be learned from great literature and art; (d) developing judgment, critical thinking, and empathy; and (e) starting to employ mental checklists to think about situations, problems, and people.

    to give a hard example. 50% of CEO are sociopaths. It takes a lot of education and insight to be forewarned and forearmed, just for starters

    And this is just a start

  23. From what I gather, signaling is another viewpoint from which you can interpret economic behavior. Is there any way of quantitatively measuring the signaling component of economic behavior? Or is it simply another Austrian-like, innumerate, qualitative argument to support a favorite bias?

  24. Anonymous8:19 AM

    Noah, love the blog but this post is weak.

    Focusing on the signaling and employment aspects, this is not a new area for economists.

    In particular, look at Richard Vedder, good summation here:

    Given the lack of a national curriculum and extreme quality variability across schools, a standardized degree from a credentialed university is a far more tractable indicator of ability than high school, SATs, etc. Testing applicants for ability is legally problematic.

    Elite university degrees are less about ability signalling than about attitudes. An ivy league graduate can be reliably expected to have social views acceptable to corporate leadership, uncritical acceptance of existing hierarchies, and obedience to authority.

    1. Yeah_No9:15 AM


      "But what exactly do they think is being signaled?

      Intelligence? No, that doesn't make sense. It's way too easy to tell who's smart. To signal general intelligence, all you need to do is take some tests - AP tests and SAT 2's if you're trying to signal moderate intelligence, International Math Olympiad if you're trying to signal exceptional mathematical intelligence, and so on. You don't need 4 years at an elite school to show you can do some math problems or memorize some stuff."

      It depends on how rigidly you define the kind of "intelligence" employers are looking for. College can help a little in weeding out people who test well but don't obey well. College degrees signal the college's prestige, which is always nice to advertise to your clients. Moreover, employers aren't the ones paying for their employees' educations, as student debt protesters demonstrate, which explains the bloat in higher ed.

    2. Hey Anonymous, love your comments but your reasoning is weak.

      If you're right, why do people spend 4 years at college rather than 2?

  25. Anonymous9:04 AM

    I think you overstate your case. In my domain (Wall Street), firms hire almost exclusively based on 1) selectivity of university and 2) GPA. It is possible to make up for lack of one or the other, but never both. I also think it's accurate to say that the interview process is designed to ensure selected new hires understand the social norms of the industry as well.

  26. It's way too easy to tell who's smart. To signal general intelligence, all you need to do is take some tests - AP tests and SAT 2's...

    In the United States it is illegal to ask a prospective (or present) employee to take an IQ or general intelligence test. It is also illegal to ask a present or prospective employee about performance on such a test. It is illegal for the employer to consider the employee's score if the employee volunteers it. Since finding out opens an employer to potential liability (an employer would have to argue that she had the information but didn't use it), HR departments do their best to keep employees from volunteering this or other prohibited information.

    1. Has this changed in the past decade?

      Half of Wall St firms requested or expected SAT scores on resumes

  27. Doesn't the signaling start at a much younger age than college? Whether church or scouts, team sports or the choice of an elite preschool, don't parents impress their young into "tribes" that send the signals that they want? In the B2B world, more managers will seek employees from the pool of graduates from their Alma-mater because they have the pedigree from their chosen tribe. They naturally like them because they are like them.

    And, if college really has descended into a parking place and party place, how much screening for motivation can there be? Wouldn't a non-grad that used their four years to get involved in numerous community service projects, building a Linked-In network of 1000 active contacts, and honed their learning through reading, blogging and publishing articles make a more interesting candidate? How many of our tribally entitled grads read and ponder blogs on economics?

  28. This post is all interesting, but it has a strong "view from 50,000 feet" quality to it. The strongest issue I have with it is that it fails to distinguish between the question of whether college is used for signaling, and the question of whether it is a good signal. These are entirely different issues.

  29. Larry Headlund11:24 AM

    Whatever Japanese people's reason is for going to college, it isn't for signaling intelligence. No reason America should be any different.

    And yet America is different. Companies in America do not hire based on scores of a high school exam. Other posters have pointed out the legal reasons why this is would be impossible now but it was never the practice in the US to do so, even before the era of discrimination law suits.

    Nor is the college experience in the US the same as that in Japan. For that matter, as you brought out in your previous post about private equity and Japan, the work environment is different in the two countries as is corporate finance.

    Everything is different yet there is no reason why the purpose of an institution to be different?

    PS: In my entire career the only potential employer who took the slightest interest in my GPA, GRE. etc. was the NSA. And they had their own test in addition.

  30. Anonymous12:36 PM

    "All three of these types of human capital have to do with bringing people together. It is not something that can be built online. It is not something that can be built on one's own...Rather than a hideously, inefficiently expensive signaling mechanism, college is an ingenious technology for building the kinds of human capital that are scarce among smart people in rich countries."

    I like the point about college being a "moratorium" on work - a time to build human capital - and I agree that people need some transition between high school and starting a career, but I think you extrapolate too much. I agree that meeting people who inspire you, meeting people who are different from you, and building social skills are important. But couldn't college just be a hideously, inefficient, expensive and outdated way of doing this?

    Why do people assume that online learning implies being on your own or not interacting with others? Think of all of the diverse and interesting people you could meet if you weren't confined to a college campus. You would certainly have access to better teachers, and you might be able to afford travelling around the country or world while you study rather than playing beer pong in dorm rooms.

    Lastly, I think it's interesting to consider the difference between signaling and screening. For students, college is a less effective signaling method than it used to be. You're paying more to compete with a far greater number of college graduates for a limited number of attractive jobs than you would have thirty years ago. For employers, college might still be a pretty efficient screening method, a way of cutting down the applicant pool.

    1. Larry Headlund12:47 PM

      ... meeting people who inspire you, meeting people who are different from you, and building social skills

      Sounds like military service (recruiters version).

  31. "Whatever Japanese people's reason is for going to college, it isn't for signaling intelligence. No reason America should be any different."

    You are apparently unfamiliar with Griggs vs. Duke Power. That's a huge reason why America is different. It's why employers can feel safe requiring a college diploma, even though they also know that blacks and Hispanics from any college will be likely to have lower skills. At the very least, it saves them from lawsuits. (SAT and AP scores are both considered proxy intelligence tests).

    The top 10-15% of our high school students have had a more demanding curriculum than about half of all college grads. Many employers of moderate intelligence jobs (bank tellers, sales managers, car salesfolk) would be pleased to hire straight out of high school if they could knowingly draw from that population. But there's no way to ask for it without getting smacked with a disparate impact lawsuit.

    This requirement has, of course, devalued a college diploma, because an entire industry exists to pull unqualified, often illiterate students through the farce of a college education (see the article The Education of Damian Cathay, for one example).

    As someone pointed out, it's rather astonishingly ill-informed to post on employer preferences and the signaling value of college without reference to disparate impact or Griggs. It makes you look a tad silly.

    (and whoever posted on Wall Street firms requiring SAT scores--I'd be surprised. It may have been something that applicants did, but if they used them in any way, they'd have been sued if discovered.)

  32. Noah,

    Just as it is illegal to ask about or use test scores in determining employment or advancement, it is illegal to use marital status. Obviously, a long-time employee has probably done things to make that information clear. However, if a job interviewer asks that question, he has violated the law, and may get his employer in trouble (so HR departments train their people to be sure not to ask or allow it to be brought up). No job application has a "marital status" box anymore.

    1. Though it comes up in interviews a lot: when employers are probing you to determine how "permanent" you'll be to the area/company, the best answer is something along the lines of "I've got a wife/kids and a house"

  33. To all the folks trying to say that disparate impact laws make it impossible to signal intelligence through testing, I say:

    Ha. Nice try.

    1. No law prevents people from putting their test scores on a resume.

    2. In fact, tech companies explicitly hire people based on their scores on the International Math Olympiad and Putnam tests. A 40 on the Putnam = instant job anywhere.

    3. College was popular for smart and rich people well before the existence of disparate impact laws.

    4. No disparate impact laws in Japan.

    Nice try, guys!!

    1. Probably 2/3 of the resumes we get have SAT scores on them, and absolutely all of them have college (or, failing that, high school) GPAs on them. We don't ask for it, but I've never not seen it included.

    2. Larry Headlund1:10 PM

      To all the folks trying to say that disparate impact laws make it impossible to signal intelligence through testing, I say:

      Ha. Nice try.

      4. No disparate impact laws in Japan.

      But in your article you said:
      In fact, in Japan, most employment decisions are based on exactly this sort of signal. High school students who want good careers spend all of high school studying for some really long college entrance exams, and employers basically pick the students who get the best scores on these exams.

      So Japan, where you say employers use standardized tests for choosing employees and which has no disparate impact laws, shows that in the US, which has disparate impact laws, employers can do the same thing?

  34. Anonymous1:25 PM

    I thought college was about baiting naive high school kids and their parents into trading their hopes, dreams and wads of cash for a life of indentured servitude.


  35. Anonymous2:12 PM

    I think this is really useful -- in other words, I find it a stimulating way to frame the issue, regardless of whether the empirical hypothesis can be validated.

    One quibble, though, and that has to do with general claims about the "types of capital that college is designed to build" or (not your words) what college is "for." What colleges are *designed* for is surely beside the point. Who knows, or cares, what the intentions of the founders of any particular institution were? The question is what the institution's value is, and that's where I think your hypothesis is a stimulating one.


  36. Zlati Petrov2:40 PM

    What does this all imply about the supposedly inevitable virtualization of college?

    It seems that at least some people expect the college experience to migrate online.

    But you argue that physical proximity matters for the cost-efficient formation of human capital. If so, then virtualization destroys some of the value-added of college.

    This means that the digital revolution will create a bit of a cleavage: those who lean heavily on getting their education online will mostly do it for the signalling component of college education, whereas those that choose to attend physically will get the full benefits.

    Does it follow that universities need to establish a two-tier pricing system?

    If so, how do we measure the value of physical proximity/human capital formation so that universities can accurately price the two tiers of the education they offer?

    1. Michael Harris6:14 PM

      It struck me thinking about Noah's post that it made clear (to me at least) why the "virtual university" that I hear trumpeted as the next step in higher education is as unlikely as the "paperless office" -- yes, it's the "logical extension" of what technology enables, but that's far from assuming it's bound to happen.

      Distance education existed pre-digital revolution, and (in Australia anyway) seemed to be the option for those unable for one reason or another to do on-campus degrees. Which meant, either mature age students (working, or caring for families), or students who didn't get accepted into a "proper university".

      I don't know that being able to access lectures online from Yale or wherever is going to change things all that much. I would presume that "going to college" is about GOING to college, and Noah's articulated a way to think about what that means quite sensibly.

    2. Michael - Exactly. ;)

    3. Zlati Petrov8:34 PM

      But as you said, some people do take online courses (or entire degrees) and I think the example you gave illuminates Noah's point: mature age students, aka students that have already formed the networks and choice sets that college helps students develop.

      It sounds like they take the courses for the signalling component of the college education and/or to develop specific skills.

      Hence we need some way to approximate the value of the signalling component alone, so we can accurately value online education.

      The question is, are the various components of college education separable? I.e. is it the case that physical attendance affects the strength or quality of the signal?

    4. There's a few things about college:

      1) Asocial skill development (e.g., solve a math problem, figure out what the logical structure of a document is). These still are best learned socially, to get support from one's peers.

      2) Social skill development - both academic and non-. Not doable online.

      3) Signalling - currently, not doable online, but this might change (with extreme difficulty, please note).

  37. All University taught me was that there is no way I'll ever know what is true and what is not, although I am not sure it did that either. There is so much heterogeneity and self selection and all that I am really not sure why it's useful to generalise.

  38. I think many of your ideas are half-baked. The one that bugged me enough to make me stop reading:

    "Now, this could indicate that college is actually about consumption. Well, to some degree, it is; college is fun. But people tend to smooth consumption, and college is all concentrated at one time, so it doesn't make sense that college would be mostly about consumption."

    Consumption is smoothed over a lifetime. If I continue with your logic, it wouldn't make sense that people go on a weekend vacation, because people smooth consumption.

    I know I'm honing in on a pretty inconsequential part of your post, but I think you should think through your ideas more.

    1. I know I'm honing in on a pretty inconsequential part of your post, but I think you should think through your ideas more.

      Perhaps you should think through my ideas more... ;)

      Vacations are smoothed over a lifetime. You take vacations when you're young. You take them when you're old.

      College only happens once in a lifetime.

      The very fact that, as you say, people smooth consumption over a lifetime means that your vacation example does not apply.

      Whaddya have to say to that, smart guy?

    2. "College only happens once in a lifetime."

      Noah, ask a sociologist about how society has organized in this way. Imagine trying to 'take' college one year per decade.

    3. Noah, ask a sociologist about how society has organized in this way. Imagine trying to 'take' college one year per decade.

      Yeah but you're still spending your lifetime consumption all at one time. Consumption of individual goods is not smoothed over the lifetime; total consumption is.

    4. Anthony10:53 PM

      Interesting. Thanks for the thoughtful post, Noah. I went to medical school in South Africa. Straight out of school in the British system of medical training. Six years later, I could deliver babies, take care of the local maladies, and was ready for an internship. We mostly worked, and my networking was minimal. Some fields just don't have time for the capital accrual you describe: we learned something else entirely - productivity.

  39. Anonymous9:44 PM

    Currently in a graduate program at an Ivy League university for public administration (MPA). Looking to get into consulting post graduation. This is my second masters from the same university.

    Thought your article was really interesting.

    Was especially intrigued by your marriage comments. Is it true that employers would value a married individual more over a single one simply because marriage "signals" a need to bring paychecks home?

    1. The classical view (back in the days of families with a male breadwinner and a housewife) was that a married man had commitments, where an unmarried man did not.

      Employers like people with commitments.

    2. Anonymous11:14 PM

      doesn't a married-with-children-worker (gender neutral) mean that one is also inflexible, unavailable for travel assignments, more expensive health-copay/insurance premiums over an unmarried-upwardly mobile worker?

  40. People go to college because firms hire people who went to college.
    And firms hire people who went to college because those people send the most interesting signal of all to the firms, they are cost-effective to hire: easy to spot (because they hold a degree) and easy to manage (because they can be expected to be reasonably intelligent, motivated and conformist).
    Firms possibly miss great opportunities by not hiring non-conformist, creative people, and there can be two explanations for that:
    - either those "weird" people are on average too expensive to find and manage
    - or there are not enough of those people in firms in the first place, and firms are not smart enough for hiring "weird" people. If this is the reason, then it creates a vicious conformism circle.

  41. Anonymous7:20 AM

    Absolute smackdown. Toga! Toga!

  42. I think attending college in the US can, in fact, be a signal of intelligence/ability. The college admission process in the US is fundamentally different. The academic statistics (GPA & SAT) used are not particularly meaningful. GPAs can't be compared between schools, and even within the same school may be nearly meaningless without weighting. And the SAT I is quite elementary compared to high-stakes testing in Asian countries. So, it falls to the colleges to sort that out. A student who earns a degree from, say, Columbia, Stanford, Yale, etc. is almost certainly quite bright simply because four years earlier they were admitted to their school after passing a holistic filter process that looked at both academic achievement as well as accomplishments outside the classroom. This admissions process is a lot more sophisticated than the process in most other countries, where some combination of test scores and family connections are the main factors. And, in the U.S., students who achieve high scores mainly by single-minded devotion to long hours of study are less likely to be admitted than students with similar scores who had significant non-academic accomplishments. Arguably, the latter group are brighter, amplifying the "signal" for U.S. college attendance.

    This is one reason why some firms historically recruit heavily from the most selective schools. The initial screening (for basic smarts) has already been done for them.

    Once outside the realm of the elite schools, the signal is a lot weaker. A degree from a less selective school doesn't say, "this is a really bright young person," even if the grad is as smart as any Ivy Leaguer.

    Having said all that, I like the human capital argument, too. I think it will be interesting to see how that works out as the college experience changes. With the ever-rising cost of a traditional 4-year residential undergrad education, I see that model crumbling as an option open to most kids.

  43. 1. Comments upthread about how little of one's formal education actually was useful/relevant are strikingly naive. All standardized forms of training or education can have only a small overlap with any one of the highly diverse jobs people fill in an advanced economy. In addition, wasted time (both recreational and misdirected vocational) is endemic in everyday work life of all kinds. Using made-up numbers purely to communicate, Wal-Mart seems like an engine of efficiency not because it productively mobilizes 90% of employees' bandwidth but because it harnesses 30% instead of the usual 10%. (As for quality of education, remember Sturgeon's Law.)

    2. Signaling is an appealing model that doesn't fit much of the data, e.g.

    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.

    3. It is true that colleges in the US teach less than they used to. But so do high schools. The value difference between college and high school learning is therefore likely to be more important today than it used to be, if you believe in diminishing marginal returns. It used to be that everybody could read and understand something like Animal Farm, but the college graduates could also understand Milton. Now, nobody gets 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.

  44. Anonymous11:22 PM

    [Sorry if somebody else said this, but I didn't feel like reading the comments]:

    So college is useless at signalling intelligence yet it's a place where you go to meet other people who are smart...and you know they're smart because...they're at college...sowhobadawhat?...

  45. Anonymous8:15 PM

    You've missed the point. The PRIMARY societal function of college is to keep millions of college-age individuals out of the job market (college students are not counted among the unemployed). If they all went immediately into job-search mode, the unemployment rosters would skyrocket. In addition, hopefully they will mature a bit, learn some things, and then enter the job market later. Even better, they will go to graduate school and stay out of the job market even longer. OK, you say, what about when they all graduate 4 years later? Nope--Stanford's 98% graduation rate is nowhere near typical. The average 6-year graduation rate is between 30 and 50%, so there is a gradual attrition of drop-outs entering the labor market, rather than a sudden influx. Colleges are to be congratulated for this contribution, and not much else.

  46. Matt Strictland1:56 AM

    8:15 has a point about the job market.

    However IMO college is about programming, not in the computer sense but signaling that you are a compliant worker who meets social expectations. Both the US and Japan being very conformist societies crave this sort of thing.

    This ties in well with what you said about marriage, kids, parties and yes consumption. Someone with little interest in those things is often perceived as a threat to the established system. They are harder to tie down and control, may be power hungry or detached and both the US (with the worlds highest incarceration per capita) and Japan are highly controlled societies.

    As for "hijinks", they are also tolerated so long as they are the right kind. Generally this is liberal with few exceptions

    Get busted protesting for say some Green cause or for trivial stuff underage drinking or in some places pot , well whatever. Kids will be Kids

    Get busted at an anti abortion rally or an anti immigration one or some Hard Right cause and assuming you aren't expelled, you'll be a pariah.

  47. HVAC technicians are among those professionals whose skills are indispensible and always needed, thus they enjoy stability and longevity even if they don’t always show up in a suit to work. If you want a more detailed information about this, click here. HVAC Schools in Washington

  48. This is very interesting to me because I didn't do the 4-year university route.

    I picked my own school when I was 19, so I ended up going to a school that was not very well researched. So, halfway through decided I was taking the school more seriously than several of my instructors. After deciding that the school was merely tuition-mining, I finished early, got an associates degree, and planned on getting my experience mostly on the job instead. It's been a lot of footwork, but I've got a relatively stable career in a field I am happy with.

    I have friends who've accomplished more with even less college exposure.

    And ever since then I've been asking myself "Why do people spend so much at college? What ARE they learning?" I've read a few opinion blogs and articles on the subject, and my conclusion is that most authors have never justified the cost of school, but are attempting to justify their own 'moratorium' from education by partying for 4 years. I enjoyed the article, and you raise some very interesting points about perspective and motivation, but I feel you are ultimately in the same camp as others justifying their own college experience. If college is merely a social pit-stop on the way to a career, shouldn't it be MUCH more affordable?

    I guess my question is really: If I can figure out all of the skills (including the 3 you've listed) associated with my field of work (graphic design) without going to a 4-year school, why should anyone?

  49. This is a great article. I 100% think that education is power. Going to college was the best decision of my life and I have benefited from it every since. I received a bachelor of science in management degree and it has helped me establish myself in the business industry. The most important thing my degree did for me is create opportunities. And with the knowledge and education I learned, I was able to apply myself and take advantage of those opportunities and become successful.