When a tenure committee evaluates the contributions of their colleague who is up for tenure at a top school, some publications count more than others. Specifically, those published in a handful of prestige journals count a whole lot more. In some disciplines, a single publication in the most prestigious journal is enough to get tenure practically anywhere. Eg, Annals of Mathematics. Prestige may not be so concentrated in every discipline, but even in more crowded fields, you can count the prestige journals with the digits of your hands. Why do disciplines and graduate departments do something so absurd? Are they in the grip of some ideology?
The economics of asymmetric information explains why. At the margin, merit is always hard to pin down. Moreover, the process must be seen as just and proper. So one must look for a relatively objective signal of merit. Furthermore, the signal must be treated as such by everyone concerned. In order to reproduce the legitimacy of the process over time, you must have general buy-in from all sides. But how can everyone concerned come to an agreement about signals? Is the choice free? Can arbitrary journals serve as the signal?
Underneath all the variation is the simple economics of asymmetric information. Whatever serves as the signal must be costly. It must be very hard to get published in the Annals. Only if it is harder to publish in the Annals than any other mathematics journal, can the Annals be considered the top journal in math. In general, prestige journals can only serve the function that they do, if they only accept top-notch articles and are very hard to get published in.
In general, signaling works precisely because it is costly. How costly does a signal have to be? Suppose that there are only two kinds of candidates: the high type and the low type. In such a situation, the signal has to be too hard for the low type for it to serve as the signal. Since you don’t know who is the high type and who is the low type, you must make the candidates jump through a hoop — one that is too hard for the low type to jump through. For in this manner, you can distinguish between the two types of players. In general, in order to agree on a signal to sort the population, we can pin down the solution at the margin, where we are again back to distinguishing between the high type and the low type. In other words, the result generalizes over the space of player types: the question boils down to who makes the cut and who does not.
Signals are a costly demonstration of private information. Not only are signals costly, if not in theory then at least in practice, they are also noisy. A perfect separating equilibrium is when you can cleanly distinguish between player types. Such clean separating equilibria are unlikely to obtain in the most important real world applications. When firms are choosing between candidates to hire, they face noisy signals like highest degree attained, GPA, SAT and other test scores.
When I studied game theory, I did not realize the significance of the theory. It was only later that I started to perceive the sorting mechanism at the top of the American social order. Like the elites of so many nations, particularly Britain, American elites sort themselves on school prestige.
Loaded elite firms do most of their hiring from prestige schools. Kids fight to the death to get into prestige schools granting undergraduate degrees. The most ambitious graduates compete further to get into graduate and professional degree-granting programs at the same handful of prestige schools. Firms pay top dollar to hire fresh graduates from prestige schools, and even more for the holders of professional and graduate degrees from the same prestige schools. School prestige matters more than your GPA, SAT and other thinly disguised tests of cognitive ability. In fact, most elite firms have a pipeline from elite schools for professional hires. They only incidentally hire anyone who does not sport a degree from a short list of prestige schools. To a good first approximation then, elite firms hire exclusively from prestige schools.
This is a classic, noisy, separating equilibrium: the brightest and most ambitious kids compete to get into the prestige schools, prestige schools only admit the brightest of the crop and firms hire them straight off the bat. The system works precisely because it is hard to get in.
There were 17 million college-aged kids in the United States in 2019. In the same year, the total undergraduate population of the Ivy League was 68 thousand. This means that 0.4 percent of these cohorts were enrolled at an Ivy League school. A better comparison can perhaps be made with the 1.9 million college degrees awarded in 2019. Of these, some 17,000 or 0.9 percent were awarded by the Ivies. There are, of course, top schools outside the Ivies: MIT, Chicago, Ann Arbor, Stanford and many others. Altogether, prestige schools admit no more than one percent of all American 18-year olds and no more than two percent of the ones admitted into college.
Daniel Markovits argues in The Meritocracy Trap that the elite school-elite firm system is meritocratic. The problem is not that the system is biased by race, gender, or even social class. Prestige schools are capitals of woke. They aggressively recruit minorities. Women are, if anything, over-represented at elite schools and colleges in general. It is true that college admissions are awfully biased by social class. But many top schools award lavish financial assistance on financially-disadvantaged students. The problem is not that school admissions are biased. The problem is that they are not. In other words, the problem is integral to the logic of meritocracy.
Every cohort is sorted on school prestige. And school pedigree continues to condition professional opportunities long after initial hiring. The distribution of salaries at law firms has been bimodal since 2000. Most lawyers do not make six figures. Elites from prestige schools working at elite law firms start at $160,000, and go from there. The setting of the TV show Suits is a law firm that only hires from Harvard Law. That may be hyperbole. After all, Yale Law is even more prestigious than Harvard Law. But fiction mimics the practice of elite sorting. And it’s not just elite law firms. The top investment banks on Wall Street, the top software firms in Silicon Valley, and the top consulting firms, all of them sport a hiring cycle whereby every year the bottom of their professional workforce is restocked by direct hires from prestige schools.
Pedigree never stops helping. Indeed, it gives you a leg up all the way to the very top. Gottesman and Morey (2006) examined CEO performance, compensation and schooling. CEOs with MBAs and law degrees from prestige schools performed no better, as measured by the firm’s stock returns, but they made more money than CEOs without prestige school pedigree.
Markovits argues that superstar labor income is a larger portion of the earnings of the one percent than capital income. More generally, he argues, one-percenters are overwhelmingly elite professionals rather than holders of capital. And they work much harder than the 99 percent. Is he right?
In order to examine the composition of the one-percent, we extract income, occupation and demographic microdata from IPUMS (ACS 5-year 2014-2018). Consisting of 15.84 million individuals, it is the largest survey of its kind. As such it is suitable for the question at hand. But note that as we go further and further up the cliff, say to the 0.01 percent, the data is more and more likely to be censored. That is, we are likely to undercount and underestimate extreme incomes because after a certain level the elite start investing heaving in income and wealth defense. But for our purposes, to understand the composition of the one-percent, this is the gold standard of all data.
Table 1 displays the income cutoffs and mean incomes of selected income classes. The median total personal income of Americans is $24,386. The bottom half of Americans who make less than that, on average make just $8,604. The middle half, between the 25th and 75th percentile, on average make only $26,000. By comparison, the cutoff for the top one percent is $329,031 or 13.5 times the median income. On average, one-percenters make $470,275 or 19.3 times the median income.
|Table 1. Income Classes.|
|Income class||Income cutoff||Mean total income|
|Bottom 50 percent||24,386||8,604|
|Middle 50 percent||8,169 – 52,000||26,014|
|Top 20 percent||60,970||121,281|
|Top 10 percent||90,986||170,540|
|Top 5 percent||128,079||237,530|
|Top 1 percent||329,031||470,275|
|Top 0.1 percent||629,470||737,867|
|Source: IPUMS, ACS 2014-2018; author’s computations.|
Table 2 displays usual hours worked in a week, the share of labor income in total income, and the percentage of respondents who report regularly working more than fifty hours a week. We can see immediately that Markovits is right: The share of labor income in total earnings of the one-percent is 75.4 percent, compared to 67.1 percent for Middle America. The share of labor income even in the top one-tenth of one-percent earners is 73.5 percent. The rich make the bulk of their income from labor not the ownership of assets.
|Table 2. Labor share and work hours.|
|Income class||Labor Share||Usual Hours Worked||Percent usually working more than 50 hr/wk|
|Bottom 50 percent||0.543||30.5||0.014|
|Middle 50 percent||0.671||37.9||0.041|
|Top 20 percent||0.789||44.7||0.145|
|Top 10 percent||0.780||45.8||0.176|
|Top 5 percent||0.750||46.8||0.209|
|Top 1 percent||0.754||48.9||0.299|
|Top 0.1 percent||0.735||50.3||0.381|
|Source: IPUMS, ACS 2014-2018; author’s computations.|
Markovits is also right about the working hours gradient. The rich systematically work longer hours than the working-class. Indeed, the higher up you go, the harder you work. The top 20 percent work 18 percent more than the middle half by income. The top 10 work 21 percent more; the top 5 percent, 24 percent more; the top 1 percent, 29 percent more; and the top 0.1 percent, 33 percent more hours than Middle America. The gradient is a testament to the workaholism of American professionals.
Not everyone works long hours to the same degree in every income class. In Middle America, only 4 percent usually work more than 50 hours/week. By contrast, 30 percent of the one-percenters usually work more than 50 hours/week. For these elite professionals, the 5-day work-week is meaningless.
The occupational categories reported by the Census Bureau are too detailed. The latest taxonomy includes 427 taxa, making data presentation a challenge. Table 3 shows the top occupations, where I have merged some categories together (eg, physicians, surgeons and dentists have been folded into doctors). Nearly 1 in 7 one-percenter is a doctor, attesting to the importance of specific fortifications of the professional middle-class. CEOs make up a tenth of one-percenters; corporate executives another 14 percent. So the managerial elite make a quarter of one-percenters. Lawyers, largely partners in elite law firms, are 7.2 percent of one-percenters; slightly more than finance professionals who make up 6.4 percent. Elite software engineers, consultants and accountants make up roughly 2 percent of one-percenters each. Even elite professors make the cut. They account for 1 percent of one-percenters. Just these occupations together account for 60 percent of one-percenters in the United States.
|Table 3. Top occupations of one-percenters.|
|Occupation||Share of the one-percent|
|Source: IPUMS, ACS 2014-2018; author’s computations.|
We may reverse the question as follows: conditional on occupation what is the probability of becoming a one-percenter? That is, if all you cared about was joining the one percent, which occupation should you choose? The occupational taxa are not so easily combined to answer this question. So we use the original taxa and sort them on the conditional probability of making it to the one-percent.
Table 4 reports the probabilities. Doctors rule. Astonishingly, physicians and surgeons face one in three odds of making it to the one-percent — twice as high as CEOs. Even dentists and podiatrists face odds comparable to those for CEOs. Clearly these are not CEOs of large, publicly traded firms. If we were to condition of large firm size, the anomaly would almost certainly vanish. But still, the excellent odds faced by doctors is no coincidence. The odds for lawyers, finance professionals, consultants, actuaries and engineers are not so small either. In any professional field, it is unexceptional for the elite to earn enough to become a one-percenter.
|Table 4. Occupational hazard rates.|
|Occupation||Probability of joining the one-percent|
|Physicians and Surgeons||0.308|
|Chief executives and legislators/public administration||0.167|
|Lawyers, and judges, magistrates, and other judicial workers||0.133|
|Personal Financial Advisors||0.123|
|Securities, Commodities, and Financial Services Sales Agents||0.113|
|Petroleum, mining and geological engineers, including mining safety engineers||0.088|
|Aircraft Pilots and Flight Engineers||0.076|
|Economists and market researchers||0.068|
|Architectural and Engineering Managers||0.065|
|Financial Specialists, nec||0.062|
|Source: IPUMS, ACS 2014-2018; author’s computations.|
Markovits goes much much further. Not only are the rich extremely hard working and highly-skilled, he argues. The mass production of highly-skilled elite professionals has distorted the occupational structure of the United States. The hourglass economy is due, he argues, to the fact that extremely productive professional elites have cornered all the skilled work, in the process hollowing out the middle of labor market by skill.
He gives the example of financial intermediation in the housing market. The mortgage business used to be dominated by mid-skilled loan officers on the local beat. Originators are now entirely deskilled form-fillers. Mortgage credit allocation is done by sophisticated algorithms written by a small handful of highly-skilled finance professionals. Further up the pipeline, highly-skilled finance professionals securitize and trade sophisticated mortgage-backed instruments. The transformation of financial intermediation has relocated not only earnings but almost all of the work to highly-educated professionals in elite firms largely hired from a handful of prestige schools.
What has happened in mortgage finance is a script that has played out not only elsewhere in finance but across industries. It is not clear how to test this hypothesis. But consideration of dynamic firms and industries — from Amazon to Uber, from Goldman to Facebook, and what have you — suggests that Markovits may be close to the truth. Skill-biased technical change, that has given us the hourglass economy and wrecked the working-class family, has been driven by the supply of highly-skilled professional elites.
Markovits goes further still. Indeed, he goes to the heart of what has concerned us for years. In his telling, the mode of reproduction of the professional middle class is not anchored in wealth transmission. Rather, professional elites pass on their advantages in two ways. First, by investing heavily in the human capital of their children. The game begins before conception. Professional elites increasingly marry each other and stay married. No stone is left unturned when it comes to providing the professional middle-class child with training, instruction and support. Already at 3, children of professionals sport vocabularies that are 70 percent larger than working-class kids. Professional middle-class parents move to districts with good schools or send their children to expensive private schools. Their kids are prepped from early childhood to eventually get into a prestige school. Markovits calculates that professional parents spend a cumulative total of $15 million on each child (the number looks absurdly high due to the magic of compounding).
Second, and much more importantly, professional parents ensure class reproduction by transmitting the secret sauce of their own success. If prestige schools are the principal sites of class reproduction for the professional middle-class, the kryptonite of the professional class kid is ‘a script of the self’ that involves an unlimited commitment to ruthless self-exploitation. The stark logic is illustrated by Amy Chua’s Battle Hymn of the Tiger Mother. Kids don’t just work like crazy to get into prestige schools. Even once they are in, they work relentlessly, in what amounts to a unrestrained arms race that ends only with retirement, if then. I discovered this when I arrived at Columbia a decade ago. In the middle of the semester, I could find nowhere to sit and work in Butler, the main library. At 3 o’clock in the morning, the place was jam-packed by 20-year-olds working their asses off. This professional-class work culture has virtually no counterpart in history. This is the centerpiece of professional class reproduction — a pathological obsession with sustained, disciplined and ruthless self-exploitation.
The connection to Clark’s theory of the industrious revolution is obvious. In Clark’s account, factory discipline was the decisive condition for survival in the competition for global industrial production. Only those nations whose working classes could be disciplined into doing sustained, hard factory work could afford to compete and hence industrialize. The anonymous economist who goes under the pseudonym Pseudoerasmus, has a similar story of Indo-Japanese divergence.
Markovits thinks that the logic of meritocracy traps American society into a situation that hurts both professional meritocrats and common people. The former are overworked, anxious, and virtually denied any leisure at all. The latter are underworked and denied access to industry, income and status. No one is happy.
Unfortunately, the last chapter problem is especially pronounced in this case. Doing away with the payroll tax, as Markovits suggests, may be a good idea. But it is unlikely to reverse the logic of the hourglass economy. Indeed, it is hard to see how the professional middle-class can break from the escalatory logic of hypercompetition that it has constructed. Nor is it clear how the skill bias of technological change can be arrested. As of writing, the hold of the logic is still intensifying. Indeed, over the horizon is next generation automation where practically the only work left for the less than highly skilled will be in personal service.
The Meritocracy Trap is an extremely important intervention. But it does have two lacunae. First, Markovits’s logic takes him awfully close to Murray and Herrnstein’s picture of the natural aristocracy. American society in both accounts is sorted on ability. For Markovits, parents ensure class reproduction by investing directly in their child’s human capital. In Murray and Herrnstein’s telling, they are fooling themselves — the bulk of the transmitted advantage is genetic. While it is understandable why Markovits didn’t want to engage with a veritable pariah, the logic of his argument demands that he rule out Murray and Herrnstein’s world.
Second, in effect, Markovits is looking exclusively at the top end of the professional middle-class that he labels the meritocratic elite. The rest is folded into a “middle class” that has no historical reality. Moreover, it plays virtually no role in his account, other than reacting to the world historical changes wrought by the professional elite, as when they elected Trump. Markovits is not alone. The tendency to look exclusively at one class and ignore the dynamic vertical interaction in elite-mass relations is near-universal. This is a significant caveat for Markovits’ analysis because it is not clear at all whether the working-class even bothers to compete in the world of the elites. And if the bulk of the American population does not even consider joining the hypercompetition, then in what sense can it be described as a meritocracy? It seems to me that one can describe the same pattern as one of a pathological elite caught up in a ruthless logic of its own making.