I saw this article going viral a few days ago, “If You’re So Smart, Why Are You So Poor?” A recurring theme in these “IQ does not predict much” or “IQ is overrated” articles, such as Taleb’s viral 2019 article “IQ is a pseudoscience swindle,” is the failure of the author to take into account individual preferences and conditional probabilities. Or an overreliance on outdated or irrelevant studies, often based on old data or from European countries, that are inapplicable to the competitiveness of modern American society.
He begins:
Here’s a number that should make you vomit: IQ explains about 21% of income variation. That’s it. Four fifths of why someone makes bank has nothing to do with their brain. Nobel economist James Heckman ran the numbers harder and found IQ accounts for 1-2% of income variance when you factor in everything else. One to two percent. Your SAT score is basically a rounding error on your paycheck.
But he’s ignoring individual preferences. As the follow-up quote in his article says, “Some get the version that notices every crack in the sidewalk and then writes poetry about how all sidewalks are moral failures. Both are ‘smart.’ Only one pays,” which seemingly refutes the aforementioned statistic that “IQ accounts for 1-2% of income variance”. The fact that people have agency to choose their careers or aspirations means that the above statistic cannot be taken at face value, because it lumps everyone together.
Longitudinal studies with lots of participants tend to have low effect sizes. This is commonly seen in nutrition research, where changing a diet slightly can produce a tiny differences in weight or life expectancy despite also being statistically significant. Or research about IQ vs life outcomes, where you have many people being tracked over many years. However, once you start adding conditionals or controlling for individual preferences, then effect sizes become much bigger.
In the context of a career setting, making a lot of money is conditional on passing the screening and interviews, which are indirectly IQ filters. Higher-paying jobs are much more selective, such as longer interviews and more credentialism. Thus, people who are smart and whose preferences are aligned towards wealth accumulation are at an advantage. You don’t get into Jane Street or Meta/Facebook unless you’re smart and can thus pass all the screening. From the 2023 article, “More harsh truths about IQ and wealth or job success: why IQ matters”:
1. The claimed low correlation between IQ and job performance or employer satisfaction is conditional on getting the job, which means clearing the necessary screening (e.g. interviews, Wonderlic test, background checks, etc.). All of these act as IQ filters. It’s not like Google picks employees blindly from the general population, unlike McDonald’s workers, who are more representative of the general population. If the carefully screened Google employees were replaced with McDonald’s employees, it goes without saying satisfaction would nosedive. Maybe this comes off as elitist, but such screening exists for a reason–all companies do it, even low-paying ones.
High-IQ people who prioritize making money will tend to apply to tech companies, investment banks, consulting firms, or hedge funds. They are not going to be attending art school.
But even for the humanities, IQ still matters. Similar to landing a high-paying tech or finance job, there is considerable competition. It’s reasonable to assume that given how verbal ability is a major part of full-scale IQ, that smarter people are at a major advantage for creative endeavors, such as writing. The top bloggers I follow, as well as successful authors in either fiction or non-fiction, all have by my estimation high verbal IQ. This is corroborated by self-reported scores on standardized tests (blogger Scott Alexander says he got a perfect score on the verbal section of the SATs). Or lawyers who later became writers and who scored well on the LSAT. A high verbal IQ is probably necessary for having a varied vocabulary and formulating a cogent story.
The above quote also addresses the common claim of “low correlation between employer satisfaction and IQ,” which although possibly true in the aggregate, overlooks how this is again conditional on getting the job and passing the necessary screening. Employers may not perceive much of a performance difference between a 120-IQ coder vs a 130-IQ coder, but coders in the aggregate are above average IQ. Someone with an IQ of, say, 120, can be trained to code, but someone with an IQ of 100 may struggle or take too long to be worthwhile from the perspective of the employer. If Google abandoned its screening and just hired people at random and tried to train them, performance metrics would tank.
Continuing:
The numbers get uglier. Swedish researchers tracked 59,000 people and found that past the national median income, about $50K, cognitive ability stops mattering. The top 1% of earners scored no better on intelligence tests than people earning half as much. Actually scored slightly worse.
These Swedish studies of claimed low IQ thresholds vs. success, are also unconvincing. The typical US ‘FAMNG+’ company, AI company, consulting firm, investment bank, or hedge fund is way more competitive in terms of being hired and pay compared to the typical equivalent European firm. The IQ bar is higher. This is the problem when trying to extrapolate European data to the US or make inferences about the later from the former.
Economically and culturally, Europe and America are quite different even if both are under the umbrella of ‘Western civilization’. Many can relate that America is much more competitive. The US has a much fatter ‘right tail’ in terms of income and wealth, where IQ becomes way more important. Reading the post histories (like on Reddit) of young people who land these lucrative, high-status tech or finance jobs in the US, they tend to all be really smart, objectively speaking going by proxies like standardized test scores, math competition placings, and grade-acceleration/enrichment, than just merely ‘above average’. Same for admittees for any top 20 college. Hard work alone will not cut it.
The author continues:
The average billionaire IQ is around 133. Smart, sure, but your average Nobel Prize scientist clocks in at 145. It takes more raw brainpower to discover a new particle than to own a yacht manufacturer. The particle discoverer dies with a pension. The yacht guy dies with a archipelago.
But percentile IQ tests max-out around 130-145, depending on the type of test for practical reasons, such as test design limitations and time constraints (trying to design a proctored test with enough fidelity to assess 1-in-10,000 rarity (IQ >150) would require so many questions as to be impractical). This is what is called the test ceiling, above which trying to ascertain a higher IQ becomes much more difficult. This does not mean their IQs are capped at 145; it’s reasonable to assume some top physicists are smarter.
Also, from the post, “IQ, Modesty, Hard Work, and Luck,” acclaim is much rarer than having a nominally high IQ, as Lewis Terman and Leta Hollingworth’s studies showed, in which in the aggregate the participants were more successful compared to the average-IQ control group, but few achieved any acclaim. There are many brilliant physicists whose discoveries will never merit a Nobel Prize, because their theories are theoretical and cannot be tested given the limitations of existing technology.
Continuing, he argues that luck is more important than IQ:
Georgetown found that affluent mediocrities have better life outcomes than brilliant poor kids. “It’s better to be born rich than smart,” they concluded. Parent’s income correlates with your income at 0.5. Your IQ? Maybe 0.2 on a good day. Italian scientists ran simulations: moderate talent plus luck beats extreme talent plus average luck. Every. Single. Time. In their models, the most successful were never the most talented. Randomness is the house, and the house always wins.
Few would deny that luck plays an important role. However, luck is an independent variable from IQ. There is no reason to expect that a high IQ is offset by worse luck, or the converse. People with average IQs or who do poorly in school can still become wealthy, an example being Richard Branson as he mentions, who has dyslexia (although dyslexia does not preclude having a high IQ). But as discussed in “The Myth of the Super-Successful ‘C’ Student,” these tend to be outliers who get a lot of media coverage, in which survivorship bias plays a huge role.
Something is newsworthy not because it’s necessarily attainable or feasible–but just the opposite–because it’s so rare that it’s remarkable and newsworthy in its own right. People get this entirety backwards. No one is writing human interest pieces about millionaire hedge fund employees or millionaire Meta employees, compared to non-college-educated nonagenarians who bequeath millions of dollars. By comparison, Kenneth C. Griffin, a hedge fund manager, had to donate $300 million for it to be deemed sufficiently newsworthy, versus $8.2 million for the 96-year-old secretary. In other words, even after having a lifetime to accumulate significant wealth, it’s still remarkable when a 90-year-old without a college degree does it. What good is having all that money when you’re too old to enjoy it, that the only option is to give it away?
The way I see it, being smart is no assurance of making a lot of money, but it sure helps, too. Just look around: who is landing these high-paying jobs, or retiring at 30-40 in tech? By in large, its smart people. Trying to come up with unhelpful counterexamples or edge cases (such as pretending that America is like Sweden) just comes off as another way of coping (or as I call “cope culture“), than accepting the reality that IQ matters more than we would like to admit. The social sciences are based on the implicit optimism that policy is a solution to social problems, but when it comes to something like IQ, policy runs into significant limitations, as IQ is innate. But acceptance, I think, is preferable to constructing a narrative or inventing a fictional reality of IQ not mattering.