Merit-Based Egalitarianism for a Just Society

Well-functioning markets—characterized by mutually beneficial exchange among political equals—lead to egalitarian outcomes with respect to income and well-being.

This is possible because the skills and traits that lead to productive work are broadly distributed. The extreme inequality that exists in the contemporary United States and other countries is not the result of well-functioning markets or vast differences in natural ability. Rather, it results from political inequality and corrupted markets.

This is the opposite claim of most people on the political Left in the twenty-first century, who assume that free markets are the source of inequality and that government institutions— including state-supported worker institutions such as unions and the collective ownership of companies—are the only means available to counteract market injustice. My claim is also contrary to arguments on the political Right that income inequality is the result of a competitive but fair market process.

This manifesto sketches some of the conceptual issues that arise when thinking about how people could distinguish themselves in a just society and provides evidence that the traits people need for success are broadly available, even under current conditions in the United States, where social advantages in upbringing differ widely across groups.

COMPENSATION AND PRODUCTIVITY

In this context, productivity means the value created by someone’s work. Value, in this sense, is complex and highly dependent on the work being done, but humans have learned to define, recognize, and appreciate value when they see it in the course of their interactions with others.

“The extreme inequality that exists in the contemporary United States and other countries is not the result of well-functioning markets or vast differences in natural ability. Rather, it results from political inequality and corrupted markets.”

Consumers have proven to be very good at identifying the valuable traits of highly complex goods and services, ranging from artistic works, repair services, meals at restaurants, higher education, vacations, and automobiles. Companies such as Yelp, Rotten Tomatoes, Amazon, and Consumer Reports systematically collect and report ratings and reviews of a wide range of products, and in many cases, scholars have found that when assessed against other standards, these measures are valid predictors of product quality. Even for a service as intangible as higher education, recent evidence shows that consumers of higher education provide subjective evaluations of their experience that strongly predict better outcomes for other students, such as their probability of graduation, obtaining an advanced degree, or earning higher incomes after graduation.

Subjective evaluations of worker performance are also extremely common, particularly in large companies. But there is strong evidence to suggest that in this context there are often serious issues associated with reviewer bias. Employee evaluations that rely on objective performance metrics are associated with better management practices and firm performance. Firms are more likely to adopt performance-based pay where labor markets are relatively unregulated—and payment is based on market mechanisms.

Where employee pay and pay growth are tied to performance within firms, managers are not simply trading on political power, except insofar as their political advantages affect their productivity. These advantages might include greater learning opportunities, or, for many men, not facing the cultural expectations and burden of childcare and other domestic responsibilities, which so often derail the careers of many women.

Compensation based on performance will not lead to perfect equality. Rather, it will lead to greater fairness in compensation than currently exists because it diminishes the impact of political power, and links pay closer to productivity. In a politically egalitarian society, people will be given nearly equal opportunities to become productive.

Of course, there are limits to the extent to which compensation can be linked to productivity, especially when it comes to services that are partially or fully paid for by governments. Within a given public organization, whether a school district, hospital system, or military branch, it is straightforward to design a system linking relative pay to productivity, but the amount of revenue available to distribute to workers will depend on political factors, such as how the voters in a society and their political representatives value the services being performed and how much they can afford to allocate to those services.

Nonetheless, even in socially democratic countries, most goods and services are privately produced, and there is a great deal of room for labor markets to function well, if they are allowed to. Moreover, even when goods and services are publicly produced, political egalitarianism demands that compensation be fair, which necessitates that performance be considered. The current practice in the United States of paying teachers based on their education credentials and experience may seem like an egalitarian policy, but it is unfair in a fundamental sense: It grants power to older and more experienced teachers regardless of whether or not they are actually deserving. There is very strong evidence that subjective teacher ratings—based on research-based rubrics—predict a teacher’s ability to promote learning, and yet ratings of any kind are rarely used to influence compensation because of the bizarre politics of education.

WHEN MARKETS GENERATE LOW WAGES

One important concern that may be raised against linking pay to markets relates to workers who provide what are typically poorly compensated but socially valuable services, such as early childcare, elementary education, elder care, home healthcare, and social services. These services are recognized as socially valuable because they embody widely shared social and ethical obligations to children, the poor, the weak, and other groups who need the support of their fellow citizens. Other services that are essential to the smooth running of society but are not deemed socially valuable in the same sense, such as cleaning services, food preparation, delivery services, and retail sales jobs, also pay consistently low wages. In addition to their social value, the provision of these services frees others to do more specialized work of greater market value.

Yet, consistently low pay suggests that markets bestow little value on these basic services. The providers face extreme competition, which surely holds compensation down. Most of the population could perform these tasks if they had to, and the providers of the services are competing against the buyers of the services themselves.

“Compensation based on performance will not lead to perfect equality. Rather, it will lead to greater fairness in compensation … ”

For instance, the parents of young children and the adult children of elderly parents can provide educational or caretaking services to their family members, but may decide that it is more efficient or satisfying for them to do something else with at least a large portion of their time—like hold a full-time job—than devote most of their waking life to taking care of others for no pay. As a result, early childhood caretakers and educators have to compete against parents. If the price becomes too high, parents will find it impractical or unaffordable and elect to quit their job and provide their own childcare. Similar competition holds down prices for cleaning services and restaurants. Most people are capable of preparing their own meals or cleaning their own living space, but if they are affluent enough, they may decide to hire other people to provide these services. As a result, there will always be a subset of occupations that pay a small percentage of the average worker’s wage. The economist Alan Manning has developed a formal analysis of these dynamics showing that demand for low-wage work will grow alongside the local income growth of the affluent.

Other labor markets do not face this kind of pressure. Most people would find it impractical if not impossible to manufacture things they buy—like a pencil, clothing, home, or new technology. Unless they live on a farm, it is highly unlikely that they could produce even a modicum of the food they need to consume. Likewise, for professional services: In law, finance, medicine, engineering, architecture, computers, and higher education the product is complex and specialized enough such that competition between providers sets prices, rather than competition between buyers and sellers. The same would be true for entertainment services. People don’t pay to watch amateurs play backyard football games or sing karaoke at bars, but they do pay to see elite athletes and musicians perform.

Within low-paying occupations, there is considerable variation in pay, and more productive workers may be rewarded, but even highly effective early childhood educators, nannies, and home healthcare workers often get saddled with pay rates that would fall below the levels of the worstperforming professional workers. They are also underpaid relative to their cognitive and noncognitive skill set, as I discuss later. Redistributive policies—such as income supplementation and the provision of benefits from the public— are an appropriate response to this problem for several reasons. Such forms of redistribution are inherently fair relative to market distribution, because they come closer to accurately compensating people for the application of their skills and abilities as well as their real contribution to society.

PRODUCTIVE TRAITS, INCOME, AND HEALTH

In practice, performance is based on a set of skills and traits that are widely distributed now and would be even more broadly available under egalitarian political regimes.

For example, in 1979, the U.S. Bureau of Labor Statistics (BLS) started collecting data from a random sample of 12,682 Americans born between 1957 and 1964 and continued to do so until 2012. In 1981, it measured cognitive ability with the Armed Forces Qualification Test (AFQT), an achievement test similar to college admission exams. Researchers at the BLS also collected results of traditional IQ measures, such as the Stanford-Binet, as well as a simple battery of five questions meant to capture the “Big 5” personality traits: extroversion, openness to new experiences, anxiety, self-discipline, and agreeableness. Additionally, the survey measured life outcomes such as family income and health status.

An important research question is: To what extent do cognitive ability and personality traits predict income and health? The answer is quite a lot. But the answers would surprise those who believe that IQ is the only thing that matters in earnings or life success.

To study this I calculated average income in middle adulthood and compared it to cognitive and personality traits measures in late adolescence. First of all, in terms of earnings, the achievement test of the AFQT is a notably better predictor of income than classic measures of IQ. If both are in the same model predicting income or health using the 1979 National Longitudinal Survey of Youth (NLSY), only the AFQT is significant. This is not surprising. IQ tests are more abstract and less relevant to real-world problems than the AFQT and other measures of cognitive achievement. A one standard deviation in cognitive ability, as measured by the AFQT, predicts 22 percent higher family income, after controlling for other demographic factors, education, and personality, whereas more traditional IQ measures predict only a 9 percent increase. The AFQT is also more predictive of educational attainment than traditional IQ measures. To put it bluntly, achievement tests are a better measure of cognitive ability than conventional IQ tests, even though they are highly correlated with one another.

“Forms of redistribution are inherently fair relative to market distribution because they come closer to accurately compensating people for their real contribution to society.”

Second, educational attainment is a stronger predictor of family income and health status than cognitive ability. In other words, when both factors are used as independent variables, educational attainment has stronger statistical power (a higher t-statistic). If you take two people with the same cognitive ability, the first person earns 10 percent higher family income for every additional year of study and reports significantly higher levels of general health.

Third, personality traits—as measured by responses to just three questions—are as important or almost as important as cognitive ability in explaining income and more important than cognitive ability in explaining health status. The questions that have the most predictive power are those that ask how people would rate themselves on three characteristic traits: extroverted/enthusiastic, dependable/self-disciplined, and anxious/easily upset, with the last trait negatively predicting income and health and the other two being positively related. Likewise, when people report feeling that fate and luck control their lives, they tend to earn less income and be less healthy than those who feel a greater sense of agency. Personality is far more important than even parental education.

Fourth, the main analysis presented here assumes that income and cognitive ability move up in a linear relationship, such that a little bit higher IQ matters as much at the top as it does at the bottom. The more complex reality is that changes in IQ matter less at the top than at the bottom. Cognitive performance maximizes income at about the 88th percentile. After that, there is no income benefit of having a higher IQ for the average person. The super-rich are no more intelligent than the merely affluent.

In sum, cognitive ability is not as important as educational attainment to income and health, and very rough personality measures are equally important for income and even more so for health. In other words, self-discipline, enthusiasm, and the ability to avoid getting anxious and upset in the face of stress matter about as much as intelligence in terms of predicting success in life.

BEYOND SKILLS

The chief problem with labor market inequality is not the market per se. Rather, it is the way political power operates to control and affect markets. This power is sometimes acute, such as when Wall Street banks and shareholders are “bailed out” by taxpayers, saving them from bad decisions and irresponsible behavior. Yet, more often, the power that shapes markets works gradually and cumulatively through decades of lobbying, cultural persuasion, and minor changes here and there to state licensing laws, tax laws, or other means of affecting how much people are paid.

The notion that skills—both cognitive and noncognitive—determine pay is the basis for the “human capital” theory. It has a lot of theoretical strength, as the above discussion illustrates, but it is also limited. The problem, for the theory, is that social rules are a key determinant of pay for many occupations. This explains why there is such variation in pay across industries and occupations, even after taking into account cognitive and noncognitive skills.

The importance of nonmarket factors is fairly obvious when considering compensation for public employees, which depends on the willingness of state and federal legislatures— and their constituents—to pay for the services provided, including policing, fire prevention, education, and, in many countries, healthcare. The vagaries of politics, not productivity, explain why teachers are paid different salaries in different cities, after adjusting for the local costs of living. Likewise, there is no reason to expect the “market” to determine prices in which the government is one of the largest buyers. That includes healthcare. It is political power—not the market—that leads to nurse practitioners being paid only half as much as family physicians, even though they are qualified to perform the same services.

The bottom line is that if political equality prevailed, markets would produce far more fair and egalitarian outcomes. The traits that predict performance and performance itself vary far less than actual income, and under conditions approximating equal opportunity, these differences would shrink further.