The social origins of inventors

That is a new paper by Aghion, Akcigit, Hyytinen, and Toivanen, here is one brief excerpt:

In particular we see that IQ is by far the main characteristic for the probability of becoming an inventor in terms of the share of variation it explains, followed by parental education.  These two groups of variables account for 66% and 16% of the overall variation captured by our model.  In contrast, IQ plays a relatively speaking much more minor role for becoming a medical doctor or a lawyer.  Parental education is the main explanatory variable for the probability of becoming a medical doctor or a lawyer (40% and 53%), with base controls and parental income also playing clearly more important roles than for inventors.

The paper offers many other points of value, and you will note the data are from Finland.


So this is the anti-Chetty "Who Becomes an Inventor in America? The Importance of Exposure to Innovation"?

Not really. According to the Chetty paper, by 8th grade test scores are explaining nearly half of the "parental income gap" in future patent rates, after a pretty steady trend upward. And there's no elementary regression table that says "here are the coefficients for test scores, gender, race, family income, etc", so my guess is the researchers made one, it turned out that test scores were very important, and so the regression table got suppressed in the paper.

Hah, to what extent can the prediction of innovation in Finland (small, homogenous, compressed financial returns to innovation) really be "anti" prediction of innovation in the US (large, diverse, higher potential returns)?

U can believe more Edisons r lost in the clumsy US vastness

That 16% is looking hot for such an important asset

Build more nuke electricity plants and midrises so more kids grow up near active engineers. Favor high-tech activities in the tax code

Ben Friedman says growth has many social benefits. Rational neglect of macro consequences is an adverse externality of market transactions, so the market will systematically undervalue growth

If Chetty and these guys and Friedman r right, seems important to stuff more and more engineering projects into pipeline

My aunt is a teacher, as is her daughter. My bff is a musician, so is his mom. His bro is a Harvard man, so is his dad. My other aunt is a nurse, so is my cousin. My mom is a nurse, so is my sis

"U can believe more Edisons r lost in the clumsy US vastness"

No. We carry a very wide net.

Find some. Prove me wrong. I'll wait.

Right you are Daws. Inventors get screwed by society, and only attracts nerds who find it an IQ challenge. Most people don't bother, for good reason.

Bonus trivia: add Diesel to the long list of inventors who committed suicide (reading about him now). He tried to implement the super-efficient Carnot cycle in practice, which is impossible, but in the process invented the constant pressure diesel engine, which powers all of big ICE engines and nearly 50% of EU cars by one estimate. Without Diesel you'd have Otto et al gasoline motors or possibly gas turbines powering ships, and globalization would be poorer. Due to high costs of manufacture (you need super high pressures to get diesel efficiency, which turn of the last century materials could not easily cope with), even with the help of prominent German industrialists the diesel engine was not initially a success. Another reason why we need, for true pioneer inventions, longer not shorter patent terms and government money / rewards. Diesel did have antecedents in the prior art, which some say he secretly knew about, despite his protests to the contrary, but still Diesel did pioneering work and should have become a millionaire. Diesel was a socialist who believed his motor would allow small artisans to live outside the city and away from electricity, which would be more healthy for them, and to power small motors with diesel motors. Ironically, it was the opposite (diesel motors became huge and electric motors became small).

Bonus bonus trivia: Tesla, inventor of the AC electric motor, also died in poverty. And you want your kid to become an inventor? That's child abuse if the child is not already wealthy.

+ forgot to add without diesel engines you'd have probably Parsons steam turbines powering ships. Parsons apparently died wealthy, as his engine was compatible with late 19th century materials technology and was quickly adopted.

Largest diesel, gas turbine and electric motors in horsepower, below. Keep in mind for mobile applications, gas turbines are not practical. So diesel wins for marine transport and trucks, and on top of that is more maintenance free than gasoline engines, does not need special anti-knocking fuel, can run on smelly, unrefined fuel, and is so reliable, with no spark plugs, that marine engines, like aircraft turbine engines, are assumed to more or less never fail (aircraft engines of course have redundancy for safety reasons). -RL

world's largest diesel engine. The 109,000-horsepower Wärtsilä-Sulzer RTA96-C, which first set sail in the Emma Mærsk in 2006, weighs in at a rotund 2,300 tons, and it's 44-feet tall and 90-feet long.Jul 22, 2011

The world's largest gas turbine is a behemoth. Named Harriet, GE's 500,000 bhp gas turbine cost around $1 billion to develop and has the potential to run an entire 600 megawatt steam power plant all by its lonesome.Feb 24, 2015

"I know that NASA and Boeing aircraft both use 60,000 hp (ELECTRIC) motors on their wind tunnels."

This conclusion seems entirely predictable. At least in the United States, there's a wide range of medical and law schools, so being really smart isn't a prerequisite for being a low end attorney or doctor.

"At least in the United States, there’s a wide range of medical and law schools, so being really smart isn’t a prerequisite for being a low end attorney or doctor."

You probably don't need to be Nobel Prize material, but I think American doctors) not to mention Brazilian doctors) are disproportionally recruited in the very top of the xhievement distribution.

Low end lawyers are sometimes not very bright, but (AFAIK) there is no guild restricting the number of seats in law schools. But if you want to become a doctor, you'd better have an IQ above 120; then, the criterion is that you have to work very hard. It's hard to model these conditional criteria but I'll bet some mathematical economists are up to it.

Based on the premeds I went to school with, I'd say the main qualification to become an MD is doing hard work. Having family that pushes you toward that goal is also important. I suppose that's like having a career in the Marines -- nobody thinks IQ matters very much in that venue, as long as you aren't too deficient.

Hard work is essential of course, but, as far as I know, pre-Meds (in the United Stated and in Brazil), not to mention Med students, are expected to absorb an enormous body of knowledge. In Brazil, candidates to Medicine courses outscore their counterparts aiming for other courses by far. It is a bigger difference in intelligence than the difference (if any) in courage, discipline or stamina between Marines and the folks from the other services.

She said, “I’m getting tired of Timothy Geitner,” and she splashed her feet and Abraham held her feeble look, in his hands CLYDE. He was only two but a two year can old more weight in his hands than a tall man with a circus mustache can carve out with knife. Little Abraham saw that woman shovel out scoops of lomane from her pail he saw the worms crawling towards him.

From my (ex) marine neighbor







Well, it's true that there are low-end law schools (though some have been closing up shop in recent years). The chances of landing a decent legal job (or even passing the bar) as a grad of one of the bottom tier law schools aren't great, however. The MD guild, on the other hand, has retained a tight grip on medical school admissions and -- especially -- residency slots. So prospective MDs do have to be good students -- though not necessarily creative or innovative ones.

Comparing inventors to doctors and lawyers to inventors isn't a fair.

Inventors are the top parentage of their field.

Comparing people who graduate from engineering vs those who graduate from medical school would likely reveal the same underlying characteristics.

I would wager that looking (somehow) at top practitioners in law and medicine would likely show the same characteristics as inventors.

"Inventors are the top parentage of their field."

Assuming you meant percentage, what does this mean? What field are inventors in?

I assume this paper studies inventors by looking at people who file for patents. Filing for a patent does not make you automatically 'on top'. There are plenty of patents that are pretty stupid, useless and pointless. Holding a patent doesn't mean you're a genius.

It all about how exposed you are and your willingness to learn new things at all cost

Mean while visit

I skimmed the paper to see what exactly makes someone an "inventor". It looks like they just check to see who has filed patents.

My limited experience in the area suggests that people who file patents are not quite the same thing as people who make and develop useful ideas/things.

Any sufficiently motivated person (especially nutjobs and esoteric status-seekers) can obtain a worthless patent for a modest fee.


Someone with no understanding of computers or programming can file a few hundred dollars and a claim to "a processor configured to (some cool result, described entirely as a result)" and then let the skill belonging to one-of-ordinary-skill-in-the-art enable their "invention".

Medical doctors and Lawyers require licences. Engineering does not. Wonder if that plays a role in why entry into the progression teens to be biased towards children of doctors and Lawyers

On engineering licensure:

Very few working engineers actually have licenses as the vast majority of engineers work for firms. You don't need a license unless you're essentially a sole practitioner with a private engineering practice, and there's not much of a market for that.

Incidentally I have actually taken the EIT. I just never bothered to proceed with all the requirements to be a "practicing engineer" as it's totally unnecessary to work in my industry.

My very good friend is from a small town in western Massachusetts that was the location of a very large GE plant, which employed dozens and dozens of engineers. Then the plant closed. The town went from high IQ to middling at best almost overnight. Today I scanned the web site that keeps track of IPOs. I'm pleased to report that the IPO market has finally recovered. But looking at the company names concerns me, as its mostly banking/investing, pharmaceuticals, and "tech". While that's not bad, the old days of innovation by industrial companies appear gone forever. We are even losing the race to build self-driving cars, as China and Europe are taking the lead (including a Ford facility!). Computer engineers and quants are innovative, but not in the way of industrial tinkerers (innovators). I had high hopes for GE (, but the new Chairman is scaling back the move to bring tech to industry (

All of our industry leaders have turned into financial firms: GE, GM, ...Harvard.

GE and GM are no longer in finance to a significant degree. It's not 2006 anymore.

See what happens when you go that route? (thanks for the correction).

Agreed, they made a big mistake trying to be banks.

I find this loss of manufacturing alarming. It is much easier to steal or violate a patent than it is to ramp up a manufacturing infrastructure.

I'll repeat what I said regarding the American study: patents are not proxies for innovation, except at the societal level, over many years with a stable patent regime.

If you want to estimate X by measuring Y, you want Y to be cheap enough and easy enough to obtain that X is almost always accompanied by Y.

Patents are quite expensive and take several years to issue. They can be useful to a business, but not always. The value of patents has been changing in recent years with the adoption of international treaties and some odd and controversial rulings by the Supreme Court (in the US, at least). They're a terrible proxy for innovation! At best, they're a proxy for innovation plus business plan.

Worse. Decilized uncorrected single-test IQ scores aren't good proxies for IQ either. They make no allowance for measurement error in their model that I can see (and surely Finland has good numbers on how low the correlation between the decilized score and true latent intelligence is, probably r=.7 or something not good like that). So their interpretations are wrong: they do not actually control for IQ in any of their analyses. In particular, the supposed 'interaction' may just be using paternal SES as an indirect measure of offspring IQ which reduces error... (Westfall & Yarkoni 2016: "Statistically Controlling for Confounding Constructs Is Harder than You Think" ) The higher paternal SES, the higher paternal IQ, and higher offspring IQ, through the indirect arrows. Anywhere in this paper they say "IQ", replace it mentally by "noisy decilized recorded score" and see if the claim holds up.

For most of their results, this implies IQ probably matters more than they conclude.

Maybe economists should talk to psychometricians once in a while and ask how one goes about modeling hypotheses about latent factors like, say, intelligence? (Economists overmathematicizing economic analyses - *you had one job*!)

IQ just predicts how good you are at taking an IQ test.

There are multiple intelligences.

IQ is merely a western social construct whose scores are meaningless for non-whites.

IQ is just about expectation and stereotype effects.

Flynn effect proves IQ is absurd.

Yada yada yada

fyi students go to med school and law school straight out of high school in Finland. Med school lasts six years, law studies’ last 3-8 years depending on one’s study pace. Law school admissions are more competitive than in the US, med school is about the same. Student demographics and status implications are about the same. Not sure if this makes any difference in how the findings should be interpreted (maybe the authors discuss this)

Something I found very interesting from p. 32 (Fig. 10): people with only secondary degree education have much higher likelihood of being inventors if their education is non-scientific. For people with college education the situation is almost exactly mirrored. Any speculations on this?

Comments for this post are closed