Being Rich Doesn’t Mean You’re More Hardworking: Economist Says Markets Amplify Luck

Why people with similar talents earn such dramatically different incomes

Share with your friends

More share buttons
Share on Pinterest

By Robert H. Frank

Why do hardworking people with similar talents and training often earn such dramatically different incomes? And why, too, have these earnings gaps grown so much larger in recent decades? Almost no other questions have proved more enduringly fascinating to economists.

The traditional approach to these questions views labor markets as perfectly competitive meritocracies in which people are paid in accordance with the value of what they produce. In this view, earnings differences result largely from individual differences in “human capital”—an amalgam of intelligence, training, experience, social skills, and other personal characteristics known to affect productivity. Human capital commands a rate of return in the marketplace, just like any other asset, suggesting that individual pay differences should be proportional to the corresponding differences in human capital. So, for example, if Sue has twice as much human capital as James, her earnings should be roughly twice as large.

Get Evonomics in your inbox

But not even the most sophisticated measures of human capital can explain more than a tiny fraction of individual earnings differences during any year. And since the distributions of intelligence, experience, and other traits across individuals don’t seem to have changed much during the past few decades, the human capital approach has little to say about growing pay disparities over time.

The human capital approach is also completely silent about the role of chance events in the labor market. It assumes that the more human capital you have, the more you get paid, which obviously isn’t always the case. Of course, most people in the top 1 percent didn’t get there just by being lucky. Almost all of them work extremely hard and are unusually good at what they do. They have lots of human capital. But what the human capital approach misses is that certain skills are far more valuable in some settings than in others. In our 1995 book, The Winner- Take-All Society, Philip Cook and I argued that a gifted salesperson, for example, will be far more productive if her assignment is to sell financial securities to sovereign wealth funds than if she’s selling children’s shoes.

If markets have been growing more competitive over time, why are the earnings gaps unaccounted for by the human capital approach larger than ever? Cook and I argued that what’s been changing is that new technologies and market institutions have been providing growing leverage for the talents of the ablest individuals. The best option available to patients suffering from a rare illness was once to consult with the most knowledgeable local practitioner. But now that medical records can be sent anywhere with a single mouse click, today’s patients can receive advice from the world’s leading authority on that illness.

Such changes didn’t begin yesterday. Alfred Marshall, the great nineteenth-century British economist, described how advances in transportation enabled the best producers in almost every domain to extend their reach. Piano manufacturing, for instance, was once widely dispersed, simply because pianos were so costly to transport. Unless they were produced close to where buyers lived, shipping costs quickly became prohibitive.

But with each extension of the highway, rail, and canal systems, shipping costs fell sharply, and at each step production became more concentrated. Worldwide, only a handful of the best piano producers now survive. It’s of course a good thing that their superior offerings are now available to more people. But an inevitable side effect has been that producers with even a slight edge over their rivals went on to capture most of the industry’s income.

Therein lies a hint about why chance events have grown more important even as markets have become more competitive. When shipping costs fell dramatically, producers who were once local monopolists serving geographically isolated markets found themselves battling one another for survival. In those battles, even a tiny cost advantage or quality edge could be decisive. Minor random events can easily tip the balance in such competitions— and in the process spell the difference between great wealth and economic failure. So luck is becoming more important in part because the stakes have increased sharply in contests whose outcomes have always hinged partly on chance events.

Cook and I argued that these changes help explain both the growing income differences between ostensibly similar individuals and the surge in income inequality that began in the late 1960s. In domain after domain, we wrote, technology has been enabling the most gifted performers to extend their reach.

Winner-take-all markets generally display two characteristic features.

One is that rewards depend less on absolute performance than on relative performance. Steffi Graf, one of the best female tennis players of all time, played at a consistently high level throughout the mid-1990s, yet she earned considerably more during the twelve months after April 1993 than during the preceding twelve months. One reason was the absence during the latter period of her rival Monica Seles, who had been forced to leave the tour after being stabbed in the back that April by a deranged fan at a tournament in Germany. Although the absolute quality of Graf’s play didn’t change much during Seles’s absence, her relative quality improved substantially.

A second important feature of winner-take-all markets is that rewards tend to be highly concentrated in the hands of a few top performers. That can occur for many reasons, but most often it’s a consequence of production technologies that extend a given performer’s reach. That’s true, for example, in the music industry, which exhibits both features of winner-take-all markets. As the economist Sherwin Rosen wrote,

“The market for classical music has never been larger than it is now, yet the number of full-time soloists on any given instrument is on the order of only a few hundred (and much smaller for instruments other than voice, violin, and piano). Performers of the first rank comprise a limited handful out of these small totals and have very large in- comes. There are also known to be substantial differences between [their incomes and the incomes of] those in the second rank, even though most consumers would have difficulty detecting more than minor differences in a “blind” hearing.”

One hundred years ago, the only way to listen to music was to attend a live performance. Then as now, opera buffs wanted to hear the most renowned singers perform, but there were only so many live events those musicians could stage during any given year. And so there was a robust market for thousands of sopranos and tenors on the worldwide tour. The lesser-ranked performers earned less than their higher-ranked colleagues, but not that much less. Now, lifelike recording technologies enable fans to hear their favorite operas reproduced faithfully at home. And those who demand the entire stage spectacle can now watch HD broadcasts of performances of the New York Metropolitan Opera Company in theaters around the globe. All the while, local opera companies have been closing their doors.

Explosion in CEO pay

The forces driving recent trends in CEO pay shed additional light on how small differences in performance can translate into enormous differences in earnings. Consider a company with $10 billion in annual earnings that has narrowed its CEO search to two finalists, one slightly more talented than the other—by enough, say, to cause a 3 percent swing in the company’s bottom line. Even that minuscule talent difference would translate into an additional $300 million in earnings. Even if the better performer were paid $100 million, that person would still be a bargain.

CEO leverage has been growing quickly as firms have expanded in size. As the New York University economists Xavier Gabaix and Augustin Landier argued in a 2008 paper, executive pay in a competitive market should vary in direct proportion to the market capitalization of the company. They found that CEO compensation at large companies grew sixfold between 1980 and 2003, roughly the same as the market-cap growth of these businesses.But growth in executive leverage alone cannot explain the explosive increase in executive salaries.

A second factor necessary to explain explosive CEO pay growth—an open market for CEOs—simply didn’t exist in earlier decades. Until recently, most corporate boards shared an implicit belief that the only credible candidates for top executive positions were employees who had spent all or most of their careers with the company. There was usually a leading internal candidate to succeed a retiring CEO, and seldom more than a few others who were even credible. Under the circumstances, CEO pay was a matter of bilateral negotiation between the board and the anointed successor.

That focus on insiders has softened in recent decades, a change driven in no small part by one particularly visible outside hire. That would be Louis J. Gerstner, who was hired away from RJR Nabisco by IBM in 1993. At the time, outside observers were extremely skeptical that a former tobacco CEO would be able to turn the struggling computer giant around. But IBM’s board felt that Gerstner’s motivational and managerial talents were just what the company needed and that subordinates could compensate for any technical gaps in Gerstner’s knowledge. The company’s bet paid off spectacularly, of course, and in the ensuing years the trend toward outside CEO hires has accelerated across most industries.

Most companies still promote CEOs from within, but even in those cases, the more open market for executive talent has completely transformed the climate in which salary negotiations take place. Internal candidates can now threaten credibly to move if they’re not paid in accordance with the market’s estimate of their economic value.

The more open market conditions have affected executive salaries in much the same way that free agency affected the salaries of professional athletes. CEOs of the largest American corporations, who were paid forty-two times as much as the average worker as recently as 1980, are now paid more than four hundred times as much. So once more we see the growing importance of the seemingly minor random events that produce small differences in absolute performance.

The winner-take-all account of rising inequality has not persuaded everyone. Some critics complain, for example, that the explosive growth of CEO pay proves that executive labor markets are not really competitive—that CEOs appoint cronies to their boards who approve unjustifiably large pay packages. We’re also told that industrial behemoths conspire to drive out their rivals, thereby extorting higher prices from captive customers. To be sure, such abuses occur. But they’re no worse now than they’ve always been. As Adam Smith wrote in The Wealth of Nations, “People of the same trade seldom meet together, even for merriment and diversion, but the conversation ends in a conspiracy against the public, or in some contrivance to raise prices.” CEOs have always appointed people they know to their boards, so that’s not enough to explain recent trends.

Events of the past two decades have provided little reason to doubt that runaway growth in top incomes has resulted in large part from increasing leverage in the “winners” positions, in tandem with growing competition to fill those positions. By every measure, markets have grown more competitive, and the most productive players have gained additional leverage since The Winner-Take-All Society’s publication in 1995.

What’s also clear is that the economic forces that have been causing the spread and intensification of winner-take-all markets have by no means run their course. We can expect continued growth in the intensity of competition on the buyers’ side for the best talent, and on the sellers’ side for the top positions.

In his widely discussed 2013 book, Capital in the Twenty-First Century, Thomas Piketty suggested yet another reason for rising inequality, which is the historical tendency for the rate of return on invested capital to exceed the overall growth rate for the economy. When that happens, he argues, wealth continues to concentrate in the hands of those who own the most capital. All things considered, then, it appears prudent to envision a future characterized by continued growth in income and wealth inequality—which is to say, a future in which chance events will become still more important.

Because the enormous prizes at stake in many arenas attract so many contestants, the winners will almost without exception be enormously talented and hardworking. But they will rarely be the most talented and hardworking people in the contestant pool. Even in contests in which luck plays only a minuscule role, winners will almost always be among the luckiest of all contestants.

Excerpted from Success and Luck: Good Fortune and the Myth of Meritocracy by Robert H. Frank. © 2016 Robert H. Frank. Published by Princeton University Press. Reprinted by permission.

2016 May 7

Donating = Changing Economics. And Changing the World.

Evonomics is free, it’s a labor of love, and it's an expense. We spend hundreds of hours and lots of dollars each month creating, curating, and promoting content that drives the next evolution of economics. If you're like us — if you think there’s a key leverage point here for making the world a better place — please consider donating. We’ll use your donation to deliver even more game-changing content, and to spread the word about that content to influential thinkers far and wide.

 $3 / month
 $7 / month
 $10 / month
 $25 / month

You can also become a one-time patron with a single donation in any amount.

If you liked this article, you'll also like these other Evonomics articles...


We welcome you to take part in the next evolution of economics. Sign up now to be kept in the loop!

  • Duncan Cairncross

    What a lot of cobblers
    “one slightly more talented than the
    other—by enough, say, to cause a 3 percent swing in the company’s bottom
    I simply do not believe that any individual in the CEO’s position has anywhere near as much influence as that

    Wait – I do believe a CEO can destroy that much value and more – but contribute it?
    No way!!
    So your ” minuscule talent difference”
    Is actually a massive huge talent several orders of magnitude higher than is sensible – like a 20ft tall basketball player

    • David Whitlock

      Why is it that there are examples of CEOs destroying that much value? Why are people in charge of picking CEOs so inept as to be unable to notice the ongoing destruction of value? There are not a small number of examples; Sears, Enron, HP, Home Depot, and this list.

      If a CEO can destroy a lot of value, then the construction/destruction of value is being controlled by what the CEO does.

      There may be external forces that produce unique circumstances/opportunities that can either make-or-break a company.

      I am reminded of something that GWB said. something to the effect that the only great presidents are wartime presidents. So he instituted a war so he could be a “great president”. That didn’t turn out so well, now did it.

      Many of those terrible CEOs were paid a lot of money, far more than they were “worth” (which was actually negative). The value they destroyed didn’t belong to them. Their destruction of value also increased income inequality.

      • Duncan Cairncross

        “I should be in charge,” said the brain , “Because I run all the body’s systems, so without me nothing would happen.”

        “I should be in charge,” said the blood , “Because I circulate oxygen all over so without me you’d waste away.”

        “I should be in charge,” said the stomach,” Because I process food and give all of you energy.”

        “I should be in charge,” said the legs, “because I carry the body wherever it needs to go.”

        “I should be in charge,” said the eyes, “Because I allow the body to see where it goes.”

        “I should be in charge,” said the rectum, “Because Im responsible for waste removal.”

        All the other body parts laughed at the rectum And insulted him, so in a huff, he shut down tight. Within a few days, the brain had a terrible headache, the stomach was bloated, the legs got wobbly, the eyes got watery, and the blood Was toxic. They all decided that the rectum should be the boss

        The Moral of the story? Even though the others do all the work…. The ass hole is usually in charge

      • David Burns

        Why should we think it is symmetric? A bad bus driver can crash the bus and create a disaster. But a good bus driver who is as far above average as the bad one is below it, will not be able to do much better than the average bus driver, assuming the average bus driver doesn’t crash. CEOs do have a bit more potential upside than bus drivers, but it is way easier to destroy value than create it.

        • David Whitlock

          It doesn’t need to be symmetric, but if the CEO has the capacity to destroy value, then the CEO has the capacity to destroy less value and so the creation/destruction of value is (at least in some way) controlled by the CEO.

          I am thinking about this from a viewpoint of control theory. If the operating line of the CEO controlling a company can have different value generation operating points, then the CEO is controlling value generation to some extent.

          • David Burns

            “If a CEO can destroy a lot of value, then the construction/destruction of value is being controlled by what the CEO does.”
            This doesn’t follow. Non-destruction is not the same as construction. Clearly the CEO can easily destroy value. Just as clearly, creation of value is much more challenging. It could still be the case that a CEO is nearly irrelevant to the positive consequences, or no more relevant than the fact that the HQ building didn’t collapse.

  • David Whitlock

    Humans at birth are completely helpless. They require care by adults 24/7 for the first few years to even survive. To become a competent adult, human infants require food, shelter, nurture, education, health care and opportunities, and differential access to these necessities results in differential effects on adult competence.

    When parental income is the source of those requirements; then adult competence depends on parental income, much more than on anything else.

    Heritability of legacy capital makes adult competence “heritable” solely through parental access to resources for infant development.

    I am pretty sure that this is the major mechanism for the “heritability” of things like “intelligence” observed in twin studies and genome wide association studies. To the extent that there are genetic differences that associate with different ethnicity (SNPs due to genetic drift), those SNP differences will associate with everything that associates with ethnicity; as in discrimination and poverty due to skin color.

    • Angry Pedant

      Hi David,

      I’m not sure I know what you mean. The Twin studies you’re referring to find genetically identical twins who were raised apart (i.e. adopted at birth) to prevent environmental contamination. If the heritable elements were cultural, you would find low correlation between the twins on a given trait, and high correlation between co-habitants (i.e. adopted siblings). In fact you find ~40-50% correlation between genetically related but environmentally separate (i.e. twins separated at birth) and as little as 5% correlation between unrelated but cohabiting (i.e. non-related adoptees) in traits from intelligence to conscientiousness. The ‘missing’ 45% is down to other social and environmental factors – probably things like schooling, friend groups etc.

      • David Whitlock

        Twins raised apart also share an in utero environment where they grew from a single cell to 10^12 cells. To the extent that there are gene-environment interactions that vary with development time, they are exactly “in sync” in MZ twins (because they started out as the same single cell).

        The models that are used to show 40-50% correlation are the ACE models, which force all variance into linear and additive environmental, genetic and noise compartments. Variance due to everything else, including interactions; as in gene-environment interaction, is imputed to be exactly zero.

        What happens then, is that when there is gene-environment interaction that affects phenotype, but the gene-environment variance is set equal to zero, all of the phenotype variance has to be “taken up” by the gene component. This is a major source of “phantom heritability”.

        • Angry Pedant

          Hi David,

          None of this reinforces your point that “Heritability of legacy capital makes adult competence “heritable” solely through parental access to resources for infant development.”

          I fail to see a connection between ‘in utero environment’ and ‘parental income’, unless you’re implying that these effects are caused by malnutrition or some other biological factor caused by extreme poverty?

          Even if the whole of the 40-50% is in fact due to ‘gene-environent interaction’, surely the genetic component is still relevant and variance cannot be explained by parental income “much more than on anything else.”?

          • David Whitlock

            Some of the variance is due to gene-environment interactions.

            your model deliberately imputes that gene-environment interactions are
            zero, then all effects that have a genetic component, even if that
            component is mediated through the environment, that effect shows up as

            For example; we know that discrimination is
            heritable because people of certain ethnicity are discriminated
            against. To the extent that discrimination has adverse effects (as in
            poverty, low SES, trauma, poor schools, abuse), those adverse effects
            will show up as “genetic” because any gene-environment effect has been
            imputed to be zero.

            In other words, the discrimination that PoC
            experience is mediated through the environment. That discrimination is
            inflicted upon PoC because they are PoC. To the extent that there are
            genetic markers for being a PoC, the adverse effects of discrimination
            will show up as “genetic” in any study that imputes gene-environment
            interactions to be zero.

            To the extent that white people have a
            higher SES, people with genes for being white will have the advantages
            of a higher SES, but those advantages will be attributed to having the
            genes that associate with high SES, not due to high SES itself (because
            the model disallows any gene-environment interaction).

          • Angry Pedant

            Hi David,

            That’s really interesting, I have to admit I haven’t thought about it like that before. The example of discrimination was particularly helpful. Thanks for taking the time to explain.

            I would love to see a study which had tried to quantify the extent of gene-culture interaction, but I can only imagine it would be very difficult to do.

          • David Whitlock

            Probably impossible. There are too many degrees of freedom. The only reason that the ACE model is used is because it is mathematically tractable because it is linear and additive. Nothing in physiology is linear. Everything is coupled and non-linear with tens or hundreds of thousands of parameters, per cell.

            My own hypothesis is that human intelligence has essentially no “genetic” component; that it is all dependent on neuroanatomy and neuroanatomy is only a product of development. For nerve cells to work together “in sync”, they require auto-tuning of the parameters of signaling, or the large ensembles of cells that instantiate thoughts could never function together.

            I suspect that when human tribes were evolving, the most important aspect of human intelligence was diversity in cognitive styles. A tribe only needed a single expert in every field, so long as they could communicate and share with the other experts. The easiest way to evolve cognitive diversity in a small gene pool is to make neurodevelopment sensitive to the environment at the level of noise. Then every brain will be unique and will have unique cognitive skills which can be differentially developed through learning across the life span.

          • Angry Pedant

            I’m not sure I understand what you mean. Intelligence obviously has a ‘genetic component’ in that our genes make intelligence possible. If several key genes (like FOXP2) were missing, we would be incapable of what anyone would reasonably call intelligence. So do you mean there is no genetic component in intelligence variance? And by that are you implying that other genes which affect cognitive function are the same in every individual?

            Equally, neuroanatomy development is also coded for in the genes, are you implying that environment has a larger impact on how development plays out than the genes do?

            Lastly, on diversity in cognitive styles, how do you explain the evidence for general intelligence and the positive manifold? The correlated areas of achievement are impressively wide.

  • John M Legge

    It wasn’t Gerstner’s motivational skills that turned IBM around: it was simply that he had not developed the overwhelming corporate arrogance that infected IBM at every level. There was no internal candidate capable of taking a dispassionate look at IBM’s problems (and chasing sacred cows with fire and sword).
    More generally, external CEOs are preferred because they don’t understand the business and don’t care how much long term damage they do as long as they can demonstrate earnings growth and justify a rising share price. CEOs don’t manage, they gamble, using the operating divisions of their company as chips.
    Rio Tinto plc sacked its chief executive a couple of years ago for destroying $15 billion in value by making bad takeover decisions. In the four years it took him to do this he was paid $13.5 million. If he had been paid more he might have done even more damage.
    Exploding CEO pay is a symptom of financialisation, not superior talent. Subject to that proviso, luck certainly plays a role in their selection. The lucky ones are those whose previous disasters don’t get exposed before they are considered for the CEO slot.

  • Mankind Global Media
  • Finally, this voice is growing up. For anyone who believe skill is linear form leading to success I suggest try to trade.