Might it be that talk of intelligence tends to induce overly hierarchical thinking and ranking? After a long back-and-forth, we’ve concluded that intelligence usually is overrated, most of all by people who are smart. But research also suggests some very important cases where intelligence really matters. We’d like to give you a guided tour as to when smarts really matter, and what is known and also what is not known about this question. As we’ve noted before, context rules when it comes to talent search.
We’re going to start with the positive, because intelligence can help a person find new ideas and put the pieces together where others cannot, and extreme intelligence may be required to be credible when exhibiting the highest levels of leadership skills, especially when you are leading other very smart people. Let’s now dig into the details of that.
Let’s consider inventors in particular, because there are some very good data available and the results from them are noteworthy. These data are from Finland, and they cover the entire male workforce born between 1961 and 1984, matching their eventual professions with their measured IQs taken from their period of conscription (only men were conscripted during that period, and thus, women do not enter into this study). The researchers also have comprehensive information about those individuals throughout significant parts of their lives, including incomes and levels of education for them and their parents. As is often the case in the Nordic countries, the data are pretty comprehensive and considered to be reliable.1
The most striking results are for the job category of inventors. If you are looking for inventors, IQ is by far the most significant of all the measurable variables we have. Furthermore, at higher levels of measured IQ the probability of becoming an inventor rises all the more. The relevant IQ measure here, by the way, is a variant of the Ravens test, which focuses on visuospatial skills rather than verbal facility.
The relationship between IQ and the probability of becoming an inventor can be expressed in a number of different ways. For instance, being in either the 91st–95th percentile of measured IQ or the 96th–100th percentile increases the probability of being an inventor by about 2–3 percentage points. Alternatively, if we hold other factors constant, having all individuals in the highest IQ decile would, statistically speaking, involve a 183 percent rise in the number of inventors compared to the status quo.
It is noteworthy how much of the explained variation in the inventor profession is explained by IQ: a stunning 66 percent of that total (if you have statistical training, that is not the partial R-squared, as most of the choice of professions comes from unmeasured variables; instead, that 66 percent is how much of the available explanation, conditioned on measured variables, comes from IQ). That number is striking, in part, because it can be compared with other variables. For instance, the next most significant variable—parental education—accounts for only 1 percent of the variation in who becomes an inventor.2
The simplest way to put this is that most people do not become inventors, and who will become an inventor is very hard to predict no matter what. But still, of all the predictors we have in this data set, IQ performs by far the best.
You might be wondering how this all squares with our general view that IQ, or more generally, intelligence, is overrated as a source of professional achievement. Well, the same paper shows that IQ plays a much smaller role in determining who ends up in other notable professions. If we look at who in Finland becomes a doctor, of the explained variation IQ accounts for only 8 percent of the total. For a lawyer, the explained variation by IQ is lower yet, namely about 5 percent. In other words, IQ is not, in general, very important, but to the extent measured variables can explain the decision to become an inventor, IQ appears to be quite important, at least compared to other measured variables.
In that same data set, if you are considering what does explain who becomes a doctor or a lawyer, parental education (and not IQ) is the main explanatory variable, accounting for 39 and 52 percent of those career decisions, respectively. Furthermore, parental income plays a larger role in pushing you into being a doctor or a lawyer than it does in being an inventor.
One lesson is that inventors are, on the whole, pretty smart, at least in Finland but probably elsewhere too. As for doctors and lawyers, it is very important to “come from a good family.” Maybe that is the kind of socioeconomic background you really need to succeed at law and medicine, but another possible reading of the data is that entry into Finnish law and medicine is in some ways too exclusionary, and lots of smart potential candidates are being kept away merely because they do not have the right socioeconomic background and upbringing.
We also believe that super-talented individuals, near the top of the achievement distribution, are in some fundamental ways like inventors. These individuals arrive at the top of their craft because they have pioneered new ways of doing things, whether it is Picasso and Braque creating cubism, Henry Ford realizing he should pay his workers $5 a day, or Sergey Brin and Larry Page realizing that the search problem could be cracked by sufficiently clever mathematics. These individuals need to see around corners that others cannot, and that’s our best guess for where smarts will continue to be important.
One carefully done study considers the connection between IQ and lifetime wages at the very top of the IQ distribution, in this case in the top 0.5 percent. The data come from children initially chosen based on a survey in California schools in 1921–22, 856 men and 672 women, and the study traces how much those individuals earned over their lifetimes. In this study, one additional point of IQ correlates with earnings of about 5 percent higher, or in this data set about $184,100 more over the course of a lifetime. In other words, even within the category of people with very good test scores, being “smarter yet” correlates with a noticeable boost in pay. And here is the critical point: among this high-IQ group, that is a steeper pay/IQ gradient than we find for the population as a whole. That steeper gradient at the top is consistent with our view that intelligence probably matters most for the very high achievers.3
The claim that intelligence matters most at higher levels of achievement has been tested most extensively with a Swedish data set covering 12,570 workers and running from 1968 to 2007. The data for that population show that personality and conscientiousness matter most at the bottom of the distribution. For instance, in the bottom tenth of earners, non-cognitive skills—which include, for instance, features of personality—matter two and a half to four times more than do cognitive skills. However, for the population in general, a boost of one standard deviation in cognitive ability is associated with a larger wage gain than is a rise of one standard deviation in non-cognitive skills. (Standard deviation, by the way, is a statistical concept referring to spread; if a variable is “normally distributed” in a sample, about 68 percent of the individuals in that sample will be contained within a one-standard-deviation segment to each side of the mean or average value.) Furthermore, the relationship between cognitive ability and wages is convex, meaning that the higher we go in the wage distribution, the more potently cognitive ability predicts earnings. In other words, once again, the top of the distribution is where smarts really matter for performance.4
Yet even then we should not regard measured intelligence as any kind of guarantee of success, if only because most people with superior measured intelligence do not end up being extremely successful in their careers.
Arguably many top talents are well described by what is called the multiplicative model of success. In the multiplicative model, final success requires a fairly tight combination of several traits—variables expressing the strength of particular traits are in some manner multiplied together to achieve a powerful final effect. For instance, to be a top-tier classical music composer, you might need great work habits, musical genius, ability to play the piano, skill in orchestrating, persistence, and to have come from a major musical center in or near central Europe. If all of those traits come together, the result may be magic, as was the case with Mozart or Beethoven. But if you are missing just one of those traits, perhaps you fail altogether. Musical genius without great work habits, for instance, might mean you become a brilliant local improviser who never puts pen to paper to compose a major symphony.
Consider the words of Vladimir Akopian, the brilliant Armenian chess player who never worked very hard at the game and thus never rose to the top: “I believe there are many talented chess players. When I play, sometimes I see players who are very talented. And by talent, many players can be compared easily; it’s not something special. But hard work is very important. And not only hard work but also a player’s weakness in character or some psychological instability can make a difference. Chess is very complicated and all of this counts. Purely in terms of talent, I believe, not only me but many others even maybe surpass these top players. It’s possible. But when you consider all things together—not only talent but the willingness to work hard, to sacrifice everything else, to be psychologically strong—not many have it in them to make it to the very top … there are many factors that need to be in place for a player to reach the world’s elite.” And outside of the very top of the game of chess, where the champions are indeed very smart, the evidence does not support a very strong link between chess achievement and intelligence.5
It is easy to think of other examples of the nonlinear importance of intelligence at the very top tiers of achievement. Being a top CEO, baseball pitcher, or Nobel Prize–winning scientist may require a combination of numerous traits, where again, the total is greater than the sum of its parts. We call this the whole package.
As a result, if you have a limited data set, you may make all sorts of choices based on variables you think ought to matter, such as extreme intelligence, but find you are getting nowhere. The people you are choosing need to have the whole package, again, especially at the higher levels of achievement. Even if some relatively well-designed statistical studies show that intelligence has a marginal value of zero, many top performers still may need to be very smart at what they do, otherwise they have no real chance of offering the whole package. In that sense, looking for extreme smarts is one necessary part of the talent search at these levels.
The value of pursuing intelligence also is relatively high when you are the first one on the scene and there is no general competition to hire that same talent. That means intelligence is a better indicator of promise for the very young, for individuals from remote or economically underexplored areas, and for individuals being brought into networks for the very first time. In contrast, intelligence is a worse indicator of hire quality if you are considering a sixty-year-old individual with an established track record.
One implication is that if you think you are especially good at spotting intelligence and other desirable qualities, you should spend more time working with the young and trying to locate and develop young talent. Arguably, you also might want to devote more attention to foreign markets or relatively untapped cultures within the United States, which perhaps have not been combed over so extensively as, say, the American coastal cities and suburbs. Alternatively, if you have decided that trying to spot intelligence is a hopeless activity, at least for you, maybe go into a sector where a lot of older people are working and you can judge them by their experience and vitas, as indeed is most appropriate.
The import of underexplored territory is another reason intelligence and other features of talent can be relatively important to find at the very, very top of the market. If you consider how the Mark Zuckerbergs, Paul McCartneys, and LeBron Jameses—the absolute top performers—developed, you can do very well by looking for stars early in their careers. Peter Thiel was the first venture capitalist to support Mark Zuckerberg, and manager Brian Epstein found and cultivated the talent of the Beatles, in both cases with handsome returns.
In what might seem like a paradox, it can be hard to spot intelligence, drive, and other positive qualities at the very, very top. Why? Well, the very, very top of the market usually is underexplored territory, virtually by definition. The most talented people usually are doing something extraordinary and fairly new, and often they are so unbelievably talented that most of us just don’t have the ability to appreciate their talents, at least not until their final achievements are on full display. If a young Gustav Mahler sat down in front of you and hummed one of his melodies, you probably wouldn’t have the talent required to see his potential to become one of the greatest Romantic composers of all time.
In other words, the super-talented are best at spotting other super-talented individuals, and there aren’t many of those super-talented talent spotters to go around. So if you are yourself a super-talented spotter of super-talented talent, you will find many instances of undervalued intelligence, undervalued positive work habits, undervalued drive, and so on. Those qualities will (correctly) appear to you undervalued because few other individuals will notice them. For example, circa 1961, who understood that the Beatles were going to shake up the world? (In fact, they were turned down by numerous record companies, and they offered their early American releases on the relatively obscure Vee-Jay label.)6
Do you know the story of John H. Hammond in music? He was the talent scout, enthusiast, and mentor who discovered and promoted Bob Dylan, Bruce Springsteen, Billie Holiday, Count Basie, Benny Goodman, Big Joe Turner, Pete Seeger, Aretha Franklin, George Benson, Leonard Cohen, and Stevie Ray Vaughan, among others. He also played a key role in reviving interest in Delta blues legend Robert Johnson. Now, that’s a pretty impressive record, but if you read about Hammond, it is not obvious what his magic formula was or how anyone else might replicate it. He did spend a lot of time around music and musicians, he was independently wealthy from childhood, and he worked in a non-racist way in a racist time, which meant he was focused on spotting black performers that others were overlooking. Perhaps George Benson offered one clue to Hammond’s success when he said, “John Hammond didn’t necessarily care about how many records I could sell: his approach was more along the lines of ‘look at what this talent can do, and I hope you enjoy him as much as I do.’” Still, at the end of the day, Hammond’s history as a scout shows that most of the top talent is not spotted by the establishment right away, precisely because those creators are doing something new and original.7
When the multiplicative model holds for potential top performers, the notion of underexplored talent is especially relevant. Usually it is much harder to spot the whole package than to look for a person who is smart, or who plays the guitar well, or who has a 98-mph fastball. Those particular traits are relatively easy to spot or measure. Yet seeing the whole package requires a much deeper synthetic ability, a good deal of luck, and what we are calling entrepreneurial alertness—that is, the ability to spot and perceive talents that others do not see.
Talented people make each other better, often in dynamic, nonlinear fashion, and this holds for intelligence as a measure of talent as well. One of Tyler’s colleagues at George Mason, Garett Jones, has written an entire book on this topic, called Hive Mind: How Your Nation’s IQ Matters So Much More than Your Own, and in that book he stresses how intelligence can have a nonlinear positive effect. That is, smart people can feed off each other and make each other better, within companies and even within nations. If you put Bill Gates in the jungle alone, he might not do better than a less talented manager, but Bill Gates with a team of one hundred hand-picked helpers may enjoy a pronounced advantage.8
Or consider the intuition from another direction. Let’s say you put one super-smart person into a dysfunctional company. That person probably cannot improve things very much on their own because bad cultures represent a series of dug-in bad practices, norms, and expectations, and such situations are very difficult to reform. Yet a company full of effective, cooperative people can make a big difference. In similar fashion, putting a good defender on a basketball team that doesn’t play solid, attentive team defense probably doesn’t help much. Someone is going to mess up their assignment and leave a shooter open, and a single good defender just can’t cover all of the other team’s shooters. Nonetheless, a basketball team full of first-rate defenders will demonstrate the power of team defense to great effect.
Silicon Valley has had much more success in writing software than any other comparable region, and that is because of its human (rather than natural) resources and how well those human resources cooperate with each other, which in this context is the smart thing to do. The productivity gap between Silicon Valley in software and, say, Chicago, Illinois, is much larger than the (relatively small) wage gap between higher-IQ and lower-IQ individuals. Silicon Valley has many very smart people working together trying to write software and also trying to teach each other how to do start-ups. Chicago does not have the same focus, and so Silicon Valley has many unicorns (start-ups that reach a valuation of $1 billion) and Chicago does not. That difference is a testament to the power of well-organized groups.
There is, furthermore, direct evidence that higher-intelligence people are better at cooperating. Researchers Eugenio Proto, Aldo Rustichini, and Andis Sofianos paid individuals to play varying games of cooperation for real money rewards. The researchers had data on the personality characteristics and IQs of the individuals playing the games, so it was possible to measure the strategies and successes of different types of people. The results were clear: high-IQ individuals in general cooperated more in these games, and IQ mattered the most in games where there were trade-offs between short-run goals and longer-run considerations. The researchers put it this way: in this situation, “intelligence matters substantially more in the long run than other factors and personality traits.”9
You also need to get just the right kind of cooperation in place, and that means looking for complementarity across your smart and otherwise meritorious hires. There is a difference between individual intelligence and how an individual contributes to the productivity of a broader group. Stephen Curry fit into his Golden State Warriors team perfectly, as he was paired with other strong shooters who made it harder to guard Curry exclusively and to whom Curry could pass the ball; Curry probably would have been much less effective on a team of slower, bigger men.
For you as an employer or talent scout, what does all this mean, practically speaking? If you are hiring one person into a relatively mature institution, intelligence and other features of talent will matter much less, while ability to fit in will matter more. If you are creating a start-up, or otherwise building an institution from scratch, and hiring a whole team, various markers of talent—including intelligence and cooperativeness—will matter much more. Hiring a whole batch of very smart people has the potential to create strongly positive, dynamic, nonlinear benefits. Look more for markers of talent when you are hiring groups of people in a relatively short period of time.
So if you have a buddy in the start-up world (Tyler’s friend here is Daniel, and Tyler wrote this particular paragraph) and he or she is working with very, very smart people … well, you don’t have to feel jealous. Instead, you can go to bed at night knowing that something has gone right with our world.
That all said, smarts have truly significant limitations, and now let’s look at those. The more you are talking about the population in general, the less intelligence is likely to matter for achievement and success.
Let’s start by considering the views of one of the smartest people we have met: Marc Andreessen, who is a general partner and co-founder at Andreessen Horowitz and who invented the Web browser. As a venture capitalist, Marc helped fund companies such as Facebook, Twitter, Groupon, Lyft, Airbnb, and Stripe, among many others. If you have any doubt about Marc’s smarts, imagine inventing the Web browser when such a thing never had been done before! But it’s not just that. If you ask Marc about political philosophy, ancient Roman history, or how a Hollywood contract might work—virtually any topic—he will stream a torrent of brilliant observations that will just blow your mind.
Yet even Marc realizes that, in general terms, smarts are overrated. In 2007 he wrote an essay called “How to Hire the Best People You’ve Ever Worked With.”10 He argued that intelligence is overrated in making hires. Intelligence is context-dependent, and it matters most when a company already is favorably situated with respect to the market, Microsoft and Google being two examples of that phenomenon. No company has succeeded simply by putting out its shingle for intelligent individuals or by asking hires to solve difficult logic puzzles. While intelligence is, of course, a good thing, Marc argues that, all other factors equal, the more important qualities in a hire are drive, self-motivation, curiosity, and ethics. He also suggests that drive and curiosity coincide to a pretty high degree, especially in an era when the internet allows you, in your spare time, to keep up on your field for free.
We think Marc’s points are well taken. Let’s also consider some more systematic and data-driven reasons why maxing out on intelligence is not, in general, the right way to go.
One way of judging talent is to look at how much a person eventually earns. It is not that the market always gets it right, but if a trait does not predict earnings at all, that trait probably is not so important for productivity.
One classic study of IQ and earnings, by economists Jeffrey S. Zax and Daniel I. Rees and based on data from Wisconsin, finds that, on average, one IQ point predicts less than a 1 percent increase in lifetime earnings. Overall, associating one IQ point with about a 1 percent increase in lifetime earnings is a slightly generous estimate of the correlation, as in some studies the point higher in IQ correlates only with a 0.5 percent increase in lifetime earnings. In sum, more IQ just doesn’t convert into that much more money.11
Or for another result, a study co-authored by Nobel laureate in economics James Heckman found that moving from the 25th to 75th percentile for intelligence was correlated with a 10 to 16 percent boost in earnings. That 10 to 16 percent pay gap is a pretty modest difference in living standards (pre-tax, by the way!), and you would readily expect to find much bigger pay differences among those living on a typical American block, or between a starting salary and what an individual might earn after a year or two of raises in a typical job.12
Alternatively, you might consider a recent study on Canadian earnings. The math here is a little more complex, but the main result is that a one-standard-deviation increase in cognitive ability corresponds to an earnings boost of 13 to 16 percent. Again, that’s a fairly large difference in smarts corresponding to only a modest difference in wages. In other words, we are looking at individuals who are pretty different but not extremely different in terms of intelligence, and finding they earn only a small amount more or less.13
Yet another approach is to look at top achievers and see how smart they are, at least as we typically measure that concept through intelligence tests. That is hard to do because the data are difficult to find, yet there is one study of Swedish CEOs, based on very good data, that does exactly this. The main results are that the median or “most typical” small-company CEO is above 66 percent of the Swedish population in cognitive ability, and the median large-company CEO is above 83 percent of the Swedish population in cognitive ability. In both cases those individuals are smarter than average, but they are not in the top 5 percent, much less the top 1 percent. So at least when it comes to CEOs, even very high achievers, at least as measured by their intelligence scores, are not as smart as you might think.14
There is also a sizable academic literature on intelligence and job performance. The most measured and thoughtful study is by Ken Richardson and Sarah H. Norgate and is titled “Does IQ Really Predict Job Performance?” IQ need not be the same thing as general intelligence, but the researchers’ conclusions are pretty sobering: “In primary studies such correlations [between IQ and job performance] have generally left over 95% of the variance unexplained.”15
So to put it another way, chasing high-IQ hires with few other considerations is not a good way to find talent. The general danger is that smart people—and maybe you are one of them—overrate the importance of smarts. Perhaps that is not a huge surprise.
Another reason not to obsess over intelligence per se is that often it is priced into the market. Most people value some kind of smarts, so if you just go running after the people who are obviously smart, you may find you are paying full price for them. The obviously smart people are not always the obvious bargains.
Consider an analogy from finance. What if someone told you to “buy up the stock of companies full of really smart people”? That is not good advice. You can believe in the importance of smarts all you want, but quality companies also tend to have expensive share prices, and that will be true of these companies full of smart people too, at least insofar as those smarts really matter. Economists have known for a long time there are no extra gains to be had from investing by running after positive qualities but neglecting price. The key instead is to find undervalued companies, and that means companies with hidden virtues. The importance of hidden virtues holds for quality hires as well, whether the dimension in question is smarts or something else.
In some cases, a phenomenon known as “winner’s curse” may mean that you end up overpaying when you get into bidding wars. If a number of companies are bidding for the same worker, the company most likely to win is the one that overvalues that worker and ends up paying too much. Even when the winner does not overbid relative to quality, the resulting gains from winning a bidding war may not be what you expect either.
How much you should chase after a particular talent marker depends on the nature of your business and the source of your profit. Let’s say you have a powerful brand name, or a special location that your competitors cannot easily copy, and thus high returns. In that case, you do not need to profit from your particular hiring decisions; you just need workers who can keep the franchise up and running. It works perfectly fine to pay the going market rate for talented individuals—and that can include smarts—and you don’t need to work hard to beat the market assessment. Your business has other virtues going for it, and you need to keep those virtues intact; going for the safe hires and paying full price for that talent is fine as a strategy.
Alternatively, your business model might be based on the superior picking of talent, as in the case of venture capital firms or sports teams. Then it is not good enough to pay the prevailing market price for ordinary measures of talent, and you need to think about going beyond readily available information to find the undiscovered gems. For instance, the three-time NBA champion Stephen Curry was picked no higher than seventh in the 2009 draft, right after the now-unknown Jonny Flynn, but Curry is headed for the Hall of Fame. At the time of the draft, Curry seemed too short, and when he arrived in the pros, he was not an obvious star and was injury-prone at first. So you can’t just look at a person’s underlying statistics and expect a revelation; you have to think carefully about how to select for the particular kinds of intelligence you really care about. It was not obvious in 2009, but Stephen Curry had not only an amazing athleticism but also the off-the-charts basketball intelligence to practice and master the deep three-point shot in a way that no one had before.
If you need to be talked down out of obsessing over intelligence and talked back into the importance of understanding context, look at many other jobs. How about U.S. president? To get even close to being president, you probably have to be pretty smart, at least along some dimensions, such as knowing how to appeal to enough of the public to get elected. But within that range of individuals, how much do smarts really matter for the job? Well, the data do not obviously support the view that the smarter presidents were the better presidents. Historians, for instance, typically have regarded Woodrow Wilson, Richard M. Nixon, and Jimmy Carter as three of the intellectually smartest presidents of the twentieth century. We don’t intend to start political arguments here, but the records of these individuals are at best mixed by most reasonable standards. Wilson brought more racist and segregationist practices to government, he cartelized parts of the economy (temporarily), and his World War I legacy paved the way for World War II. Nixon had some significant achievements in foreign policy and with the Clean Air Act, but he was an inveterate liar and he was forced to resign, mired in scandal and disgrace; nonetheless, the level of intellect and literacy displayed in his books was remarkable. Jimmy Carter is one of the harder cases to debate, but many view him as a failed one-term president, weak in foreign affairs, who left America with a legacy of high inflation, high interest rates, and a troubled economy; maybe that wasn’t his fault, but he is not an obvious example of a successful American president.
Partisans will disagree about which American presidents were successful, but Franklin Delano Roosevelt and Ronald Reagan are two common picks (obviously the choice depends on the party and ideology of the person being asked). Neither was intellectual in the traditional sense, and while each certainly had plenty of shrewdness, they are not typically considered the smartest of U.S. presidents either, at least not as might be measured by standardized testing.
We’re not saying we should elect presidents with no consideration of their intelligence levels. Rather, we’re pointing out that once we restrict our attention to people who have any chance at all of winning the presidency, other factors besides intelligence are very often more important.
You might look at the National Association of Colleges and Employers 2020 job outlook survey, where you will find more emphasis on intelligence, which is not surprising since the focus is on college graduates (a minority of Americans). The number one desired quality in hires is “problem-solving skills,” which seems pretty clearly directly linked to intelligence. Still, next are “ability to work in a team” and “strong work ethic,” and overall, seven of the top ten entries do not refer to intelligence. In this survey, intelligence clearly matters; it simply does not dominate in relevance.16
Okay, so that is our take on smarts. What about personality?