In the excellent journal *Intelligence*, scholar Jonathan Wai attempted to estimate what percentage of American elites (billionaires, senators, Fortune 500 CEOs, federal judges) are intellectually gifted by counting how many of them attended elite schools (i.e. schools that recruit students with extremely high scores on tests like the SAT, LSAT, and GMAT). So for example, Wai found that 45% of U.S. billionaires attended elite schools and thus concluded that 45% are intellectually gifted.

At first glance that sounds incredibly simplistic. For many mediocre minds attend elite schools because they have rich and powerful parents or because they win athletic scholarships or are unusually driven. Conversely, many brilliant minds attend mediocre schools because they lack those advantages or just want to stay in their home towns. But Wai cleverly argued that both kinds of exceptions cancel each other out in the aggregate data. So while one can’t tell if any specific individual is gifted from the college they attended, one might be able to tell how many gifted people are in certain groups (i.e. billionaires) from how many elite college alumni are in those groups.

For the purpose of Wai’s study, gifted can be defined as the top 1% in intelligence (IQ 135+) because Wai argued that that’s the level of ability the average elite college alumni has. So since 45% of U.S. billionaires attended elite colleges, 45% should be IQ 135+. If nearly half of U.S. billionaires are 135+, the average U.S. billionaire should be nearly 135.

Is that plausible? As of 2013, there were 424 billionaires in America out of 242 million adults, which means billionaires are about one in half a million. Given that rarity, if there were a perfect correlation between IQ and money, the dumbest billionaire would have an IQ around 169, and the average billionaire would have an IQ of 172 (72 points above average!). But since the correlation between IQ and earnings is not perfect, but rather it’s 0.4, we multiply those 72 points above 100 by 0.4, which gives the average billionaire an IQ of 129 (similar to what I’ve previously estimated using a different method).

But not all billionaires got rich through earnings. About a third inherited a big chunk of it from their fathers or husbands. If the average billionaire has an IQ of 129, the heirs of billionaires would have IQ’s around 113 given the 0.42 to 0.45 IQ correlation between a man and his child or a man and his wife, respectively (I’m focusing on men because they’re the vast majority of self-made billionaires, especially in past generations). In other words, the heirs of billionaires regress to the mean, keeping only 42% to 45% of each IQ point above 100 that was needed to make that money, which is why in a few generations, the money is almost all gone.

So if roughly two thirds of billionaires are self-made (IQ 129) and the remaining third are legacies (IQ 113) the average IQ of all billionaires would be 124. Roughly 10 points less than Wai’s methodology would estimate using elite school attendance rates.

Why does Wai’s method seem to overestimate? In my humble opinion, it’s because Wai overestimates the IQ of elite college alumni. While it’s true that elite college students average the equivalent of IQ 135+ on tests like the SAT, where they average 1400+, those colleges explicitly select for high SAT scores. When you select for people who did extremely well on a particular test, you get a lot of people who over-performed on that one particularly test but will regress dramatically to the mean when given tests that weren’t used to select them. An excellent study by Meredith Frey and Douglas Detterman found that SAT scores have corrected correlations of about 0.8 with other measures of intelligence, so we would expect people who were selected to have an average IQ equivalents of 135+ on the SAT, do be only 80% as far above the mean on official IQ tests (IQ 128+). Indeed I’ve previously blogged about how Harvard students average an IQ of 128 when you move them from the SAT (where they score through the roof) to a more neutral test like the abbreviated Wechsler.

So I think Wai was correct in asserting (based on elite college attendance) that 45% of billionaires are intellectually gifted, but only if he redefines gifted from 135+ to 128+. If 45% are 128+, then the average billionaire would drop from the mid 130s to the mid 120s which makes sense given that many inherited their wealth.

This revised approach also makes sense when you look at U.S. presidents. An astonishing 2/3rds of recent presidents have attended schools Wai classifies as elite, but there’s no way 2/3rds of recent presidents were above IQ 135. That would imply their average IQ is an incredible 140 (10 points higher than they actually seem to score on average). On the other hand, it’s quite plausible that roughly 2/3rds are above IQ 128.

IC

said:In market economy, a very slight advantage can accumulate into huge profit/wealth difference. So relationship between IQ and wealth is not linear at all.

pumpkinperson

said:That’s certainly true if we look at actual dollars.

In one study, the distribution of annual income was found to have a mean of about $44,000 and a standard deviation (SD) of $48,000. That would make a typical self-made billionaire who might earn say $300 million a year, 6,249 SD above the mean! Now if we applied a linear 0.4 correlation between IQ and income, their IQ should be 0.4(6,249 SD) = +2,500 SD.

An IQ that is 2,500 SD above the mean is 2,500(15) + 100 = 37,600!

Obviously the average self-made billionaire does not have an IQ of 37,600, so what a Promethean once told me to do was to NORMALIZE the distribution of income. So if say one in a million U.S. adults makes $300 million a year, then $300 million a year is assigned a normalized Z score of about +4.73 because that’s the Z score a one in a million value has in a Gaussian curve.

This gives a much more reasonable number with which to apply linear regression. Actually what experts typically do to make relationships more linear is they take the natural logarithm of income.

And as you know, my really simple method was just to express income as how many figures a year you earn and correlate that with IQ:

https://brainsize.wordpress.com/2014/05/18/iq-and-income/comment-page-1/

Pingback: The IQ’s of Fortune 500 CEOs | Brain Size