Extending the Race between Education and Technology

There is a new and updated take on this topic by Autor, Goldin, amd Katz:

The race between education and technology provides a canonical framework that does an excellent job of explaining U.S. wage structure changes across the twentieth century. The framework involves secular increases in the demand for more-educated workers from skill-biased technological change, combined with variations in the supply of skills from changes in educational access. We expand the analysis backwards and forwards. The framework helps explain rising skill differentials in the nineteenth and twenty-first centuries, but needs to be augmented to illuminate the recent convexification of education returns and implied slowdown in the growth of the relative demand for college workers. Increased educational wage differentials explain 75 percent of the rise of U.S. wage inequality from 1980 to 2000 as compared to 38 percent for 2000 to 2017.

Emphasis added, here is the full NBER working paper.  For background, here is my old NYT column on their earlier work on this topic.

Note that for the most recent rise in inequality across 2000-2017, most of it has happened within educational groups.  The less polite way of putting that — my words not those of the authors — is that the real marginal product of education is explaining less of the variation in earnings, or in other words the higher earners are drawing upon something they are not getting at school.

Students of the “education as signaling” debate also should note that, due to these results, now a) signaling is more relevant for your early wage offer, and b) signaling is less relevant for your eventual wage profile, which in fact is now more determined by your personal level of skill.


In community colleges for most courses there will be 5% or so that get jobs, the rest simply don't have the ability. Typically they don't vet applicants; the course itself does it, but not by graduation but by marks. It has gotten almost ridiculous; the goal isn't to prepare people for an occupation but to move bodies through the process. There are exceptions of course, typically because of external pressure or standards or the ability of a group of instructors to prevent administration from taking control of the specific program.

How much of this same dynamic has moved to degree granting educational institutions?

"now more determined by your personal level of skill."

Workers don't get raises because their skill is better. Your wage profile is determined by your negotiation level. Some of the most skilled individuals I know don't get paid their worth because they clam up when its time to negotiate. Some of the least skilled individuals I know are overpaid because they know how to BS

+1. You don't get what you deserve, you get what you negotiate, as a lawyer's self-help book once said on the dust jacket.

Can the OP findings also lend credence to the cliche that in the 1980s, "image is everything"? People realized that education was signaling, but it was already "too late" to turn back the clock after a brief growth spurt in the 90s. How then does one explain the breakdown in the 1960s and 70s with signaling? Could the height of the Vietnam war (which roughly coincided with wage stagnation in early 1970s) mark the turning point in education as signaling? People failed to embrace signaling since it was already on the way out, or, was it the other way around (the rest of the developed world was catching up with the USA so people realized education wasn't necessary). The 1980s as a last gasp? The grungy 90s, 00s, 10s, 20s as a return to hippie un-signaling trend?

Bonus trivia: both Ronald Reagan and Margaret Thatcher were concerned with image, which is telling, telling us something. Compare to the anti-positional, anti-image Trump, fitting for the times (Trump as good an actor as Reagan, keeping in mind he's a product of the 80s and knows well the power of signaling).

"Workers don't get raises because their skill is better. Your wage profile is determined by your negotiation level."

I always thought negotiation was a skill I learned from mom, dad and job experience. Now I know it's in the genes, you cannot learn it ;)

'in fact is now more determined by your personal level of skill': I agree 'skill' is the wrong word. It's probably just a label for "personal characteristics of an unidentified and unquantified kind".

I get the impression that Greg Cochran likes the American word "moxie" for this kind of thing. Maybe the British slang "nous" would do.

Not just negotiation, but a ton of path dependence too. A single good or bad boss or mentor or career choice early on can make or break your whole career.

It can even be important to graduate from university in the right year.

I'd argue that the only significant "time to negotiate" is when one is accepting a new job, and that to get to the negotiating table requires a "personal level of skill" (and that skill is not negotiation).

"now a) signaling is more relevant for your early wage offer, and b) signaling is less relevant for your eventual wage profile, which in fact is now more determined by your personal level of skill."

Interpretation: school gets you the job because of signaling and you learn the skills on the job. That seems like a long way of saying "signaling."

Why not “the higher earners are getting things from school that the others are not”? Why ‘education as signaling’ instead of ‘high variance in the effectiveness of education’?

Checking to see if they rely on the utterly meaningless "years of education" faux-proxy without regard to quality or field of education...

Affirmative. Well that's disappointing, file this under "not enough information to claim what is being claimed."

"Utterly meaningless"? As n grows the noise becomes negligible and the proxy asymptotically approaches intelligence. This in itself has been the subject of a lot of research.

Wow, China paid a Harvard researcher $50k a month! How come Harvard or other big American institutions with lots of money can't do the same? The sciences are so vastly underpaid in the US that nanoscientists and cancer researchers have to resort to crime while much poorer countries like China bid up the talent like its the NFL free agency. Something is very wrong here.


Eric Weinstein:

"Now would be an excellent time to ask how safe it is to underpay top scientists & engineers relative to say Investment Bankers, VCs or mgmt consultants.

Because you‘ve been habituated to bargain basement pricing for skills valued by our military rivals."


It’s sad how the government responds to this bidding up of scientist wages (a positive thing that should encourage more people to go into science) not by increasing funding for science but by making it illegal to take competing job offers.

I make $750,000 a year at ~40 yrs.

I was making 65,000 - 80,000 until I was 32. No additional schooling (4 year college + part time business school in my 20s) just got a better job and moved up the ranks.

I would say 90%+ of my wage increases has been luck, not skill. I’ve worked very hard, but so did everyone else I was working with in my 20s. I was unlucky to have a stagnant wage post financial crisis, and very lucky I was able to switch to a faster growing career after a grueling 18 month job search.

I don't see how you could make that much unless you run your own business. You really think everyone else was working just as hard and they're just as smart/skilled? Not been my experience.

@XVO - Wall Street types make several million a year so $750k a year is not unusual, if anything it's a bit low. Burnout rate however is high so that's a limiting factor.

Yes, it’s on Wall Street.

I ask below what we mean by "group." Maybe we also need to ask what we mean by "inequality."

There are plenty of college degrees that set you up for a solid middle class income, and position you for the long ball, if luck is right.

To me though, looking at the middle class, and higher, isn't looking at the critical or interesting parts of "inequality."

I'm more worried about "no college," "some college," or "bad degree" and how they relate to "poor" or "deaths of despair."

Also, I think hitting the long ball can be more about personality (competitiveness or even channeled aggression) than "signaling."

Nobody is out there examining everyone for signals. A subset just go for it. And if they are lucky, the long ball goes over the fence.

Same. Every single job I’ve had after my first job after college was not through a transparent application process but the result of the dumb luck of having the right connection at the right time. And there is a ton of luck within jobs too, such as getting assigned the bosses/mentors/projects that end up being important versus not. Smarts, hard work, and a commitment to a high-paying profession is probably enough to get people a better-than-even chance of making in the $100s (higher if you choose to be a doctor), but anything beyond that requires a fair degree of luck. The median mid-career salary for even a Harvard graduate is only around $80k.

And wealth is even more luck-dependent than income because so much of wealth is just market timing—if you had your peak earning years around 2008, you’d be much richer than someone who had the same income and savings habits as you but in 1999.

I suspect the big difference is returns to capital as compared to the returns to labor. A problem is that capital and labor are often difficult to distinguish. Thus, in an economy with a large banking sector, how does one distinguish returns to capital vs. returns to labor? In the tech sector, where compensation is largely attributable to rising stock prices (including gains from stock options), how does one distinguish returns to capital vs. returns to labor? Indeed, income inequality may well be overstated. While one can point to the data, how accurate is the data when the distinction between capital and labor is nebulous. My concern, expressed often, is wealth inequality not income inequality: dependence on capital appreciation becomes self-reinforcing, with tax policy highly favoring capital appreciation over wages. What's the Fed to do? Is Powell really the enemy of the people? https://www.nytimes.com/2020/01/28/business/economy/ordinary-or-enemy-how-jay-powell-is-positioning-the-fed-in-a-fraught-era.html

"Note that for the most recent rise in inequality across 2000-2017, most of it has happened within educational groups."

What are you calling a "group," graduates of UT, or UT Petroleum Engineers?

It's rather critical to get this right, before deciding that it was something "not learned in School."

More than ever education is free, its proof in the form of degrees that cost the big dollars.
If it's education and not ability why are companies not doing more training?

Most people suck at training. The successful youtube people are sitting on a giant pyramid of the sculls of their vanquished competitors. Your HR lady is not that good.

This seems evidence that the education signal is weakening. Just having a degree tells you less but is a minimal cutoff. One guess is the unmeasured quality of the degree which is correlated with ability enhances underlying talent and smarts.

In my career, the education signal has basically gone to zero. I work in Software Development, and the new normal for interviewing (proving that "personal level of skill") is to have grueling 8 hour interviews, which are more like tests that are hard to prepare for.

Interviewers pose difficult, academic problems and the interviewee must code up a solution live, directly on a whiteboard or an interviewer-provided environment. These types of problems typically have little to do with the day-to-day work of the position, and are more of a hurdle to be jumped over.

I think some of this is due to big companies (FAANG: Facebook, Apple, Amazon, Netflix, Google) being able to implement an interviewing solution that effectively has zero false positives (but probably high false negatives) due to a huge supply of applicants.

An entire economy has built up around helping developers get ready for these types of tests. Unfortunately our sector has no official certification, so these tests have to be repeated for every job change.

Students of the “education as signaling” debate also should note that, due to these results, now a) signaling is more relevant for your early wage offer, and b) signaling is less relevant for your eventual wage profile, which in fact is now more determined by your personal level of skill.

So you might say that signaling gets you in the door, but once you're in, your true skill level or ability to develop skills keeps you there and levels you up.

Federal government gives students loans to pay for college with absolutely no underwriting requirements around ability to pay, degree, gpa, SAT. New, poor quality, colleges sprout up everywhere to take the government money. Existing colleges create new, poor quality programs to take the government money. College degrees no longer equates to skill or talent. They need to add school and degree to their data set.

Professor Percy Marks had the clearest insight into education a century ago. Of course, the "trained to learn" is far less these days. There was a time the college graduate could write well, but that started downhill in the 1970s when the English department started getting "woke". Similarly with developing arguments, and more importantly the ability "to entertain a thought without accepting it" it to quote Aristotle on the educated mind.

"The idea is, of course, that men are successful because they have gone to college. No idea was ever more absurd. No man is successful because he has managed to pass a certain number of courses and has received a sheepskin which tells the world in Latin, that neither the world nor the graduate can read, that he has successfully completed the work required. If the man is successful, it is because he has the qualities for success in him; the college "education" has merely, speaking in terms' of horticulture, forced those qualities and given him certain intellectual tools with which to work—tools which he could have got without going to college, but not nearly so quickly. So far as anything practical is concerned, a college is simply an intellectual hothouse. For four years the mind of the undergraduate is put "under glass," and a very warm and constant sunshine is poured down upon it. The result is, of course, that his mind blooms earlier than it would in the much cooler intellectual atmosphere of the business world.

"A man learns more about business in the first six months after his graduation than he does in his whole four years of college. But—and here is the "practical" result of his college work—he learns far more in those six months than if he had not gone to college. He has been trained to learn, and that, to all intents and purposes, is all the training he has received. To say that he has been trained to think is to say essentially that he has been trained to learn, but remember that it is impossible to teach a man to think. The power to think must be inherently his. All that the teacher can do is help him learn to order his thoughts—such as they are. "

Marks, Percy, "Under Glass", Scribner's Magazine Vol 73, 1923, p 47


This conforms to my 40 years of hiring experience: Where you went to school and how you did there are essential to getting your first job. After that, they are irrelevant – to retention, promotion, compensation or finding another job.

I was the hiring partner for my law firms. We hired exclusively from the top law schools (Harvard, Stanford, Columbia ...).

For initial hire, everyone cares exclusively about 1) your law school, 2) how you did there and 3) your personality in a 20-minute interview.

Once you're hired, nobody cares a whit about where you went to school or how you did. If you change firms, nobody who interviews you will care a whit.

My personal criteria for hiring? Military service (always hire a Marine; he/she will find a way to get the job done) and college major (STEM is positive, Poli Sci negative).

Comments for this post are closed