Editor’s Note: Could our own dating habits be to blame for the economic mobility that, according to a recent study, has been stagnant for three decades?
In this adaptation of his recent book, “Everything I Ever Needed to Know About Economics I Learned from Online Dating,” Stanford economist Paul Oyer explains how our preferences to be around other people perpetuate our own socioeconomic positions.
Oyer’s first post for Making Sen$e – a compelling read for anyone reentering the dating market or trying to understand economics – is about how the separated economist gets discriminated against online.
—Simone Pathe, Making Sen$e Editor
I am a terrible dancer. I have no natural rhythm, I move in an ungraceful manner, and my lack of talent is compounded by pronounced self-consciousness when I’m on the dance floor.
If dancing were a primary issue in making couples happy together, it would be bad for me to date a woman who was a good dancer. She would be frustrated by my reluctance to dance, my inability to keep up with her while dancing, and with having to watch me try. So it would make much more sense for good dancers to date one another and for women who cannot dance to go out with the likes of me.
The top law firms recruit the cream of the crop from the top law schools. Many other firms employ the merely excellent. There are plenty of prestigious law firms that recruit new lawyers from good law schools and that are quite picky about whom they select. But there are a few firms that will, with rare exceptions, make offers only to the top students (those who make the law review) at the most elite law schools. These firms, such as Wachtell, Lipton, Rosen, and Katz and Cravath, Swaine and Moore, do some of the most important and complicated legal work in the business and financial world. They believe that they need the very best and brightest lawyers in order to do such work effectively. The fact that clients demand these firms’ services indicates that their business model is sensible. In the market for lawyers, then, the best and the brightest lawyers are matched to the best and most valued firms just as I would date women who cannot dance in a world where dancing was key to matchmaking.
Why do dancers and lawyers pair off this way? This is a pretty intuitive idea, but some economist somewhere — borrowing from the biologists who use it to explain the same phenomenon in the animal kingdom — gave it an unintuitive name: positive assortative mating. The basic idea is that people sort into partnerships or groups in a manner that is non-random and, in particular, can be “ordered.” So in the strictest version of positive assortative mating, the “best” woman would be mated to the “best” man, the second-best to the second-best, and so on.
When we see positive assortative mating, it will generally be the case that it leads to a “better” outcome than if people (or lawyers and firms) matched randomly. Consider the following simple example. Suppose that Addison is a very beautiful woman and Chris is homely. Now suppose that Bailey is an extremely handsome and well-built man while Devon is unattractive. Other things being equal, both men would probably like to date Addison and both women would choose Bailey. However, the best way to form pairs is probably to pair Addison with Bailey and Chris with Devon. Addison and Bailey would be able to enjoy one another’s company with less insecurity, for example.
Positive assortative mating is clearly an important force in the dating market. More attractive and wealthier people are always more in demand, but demand for attractive people on dating sites is especially high among other attractive people. More broadly, researchers have shown that there is strong positive assortative mating in terms of physical attractiveness, income, race, education and other traits on dating sites.
We will come back to Addison, Bailey, Chris and Devon regularly, when they will represent people looking for a spouse, employees looking for a job and others needing grouping. In each case, they can be ordered from highest to lowest and, to keep it intuitive, the ranking will always be alphabetical. Addison will be the “best” (or Addison and Bailey will be the top) and Devon the “worst” (or Chris and Devon will be the bottom). Their names were specifically chosen to be androgynous to allow for same-sex or opposite-sex pairings, depending on the context. We will look at several examples and ask: When do we see Addison pairing with Bailey? When do we see Addison pairing with Devon? And, why might these pairings make sense?
In the case of dancing (where the people are ordered by dance ability) or law co-workers (where they are ordered by lawyering skills), we have already seen that the Addisons of the world typically pair off with the Baileys. That pattern holds in many other contexts, but not all.
Many couples meet in college, so it should be no great surprise that you are much more likely to marry a college graduate if you went to college than if you did not. For example, consider all American married couples in the year 2000 where the wife is between the ages of 18 and 40. A little more than one-quarter of the people in these marriages were college graduates. But two-thirds of the spouses of college graduates were college graduates. Similarly, about 11 percent of the people in these marriages had not completed high school, but more than half of those who had not completed high school were married to another high school dropout. So, if we were to rank our four friends by their education (that is, Addison stayed in school the longest, Devon the shortest), we would be very likely to find Addison married to Bailey and Chris to Devon.
This pattern of “like mates with like” is much broader than traditional marriages and education, however. Gay couples, lesbian couples and opposite-sex unmarried cohabitating couples are equally similar in terms of education as are married couples. Partners in all types of committed relationships are also more similar than a random match in terms of income. People who work relatively long hours and make relatively more money live with other people who work hard and earn high incomes. “Power couples” are fairly common because people with high earning power pair off, leaving the less-well-off to each other.
Positive assortative mating is very strong in couples of all kinds. The authors of one study conclude that, “We find evidence of positive assortative mating for all traits and across all types of couples.” So, no matter what characteristic is used to rank the four, we expect to see Addison married to or in a relationship with Bailey and Chris attached to Devon.
In the movie “Pretty Woman,” Richard Gere plays a wealthy businessman and Julia Roberts portrays a call girl who has had a rough life. They fall in love and live happily ever after, allowing Roberts’s character to climb the social hierarchy. Though a bit unconventional, it is the type of rags-to-riches story people love to hear and Hollywood loves to provide. But how realistic is it? Not very — and one reason is positive assortative mating: very few hookers marry rich men. We tend to pair off within the social class we grew up in, which reinforces our positions in the economic system.
Just as couples tend to have similar education, couples typically come from similar backgrounds. So not only is it the case that a person with a college education is more likely to have a partner with a college education, but the children of people with college educations are more likely to have partners whose parents have a college education. The similarity of parents’ education and other characteristics among married couples has grown over time, which is probably making the “Pretty Woman” scenario even less likely.
In most countries, it is very hard to move up the economic ladder. An important contributing factor to this quagmire is that we pair off with people who are more and more like we are, both in terms of family backgrounds and personal characteristics.
On dating sites, the people who are most like you are the ones most interested in your profile. You meet people similar to you in school and through your parents. So where are you going to meet people who are different from you? How about work? Lots of couples meet at work (for one piece of hard data, there were three workplace weddings in the nine-year run of “The Office”). Maybe work is a good place to broaden your horizons?
No such luck. Workplaces are another bastion of positive assortative mating. We work with people who are like us because different kinds of employers appeal to different types of workers.
There is positive assortative mating in the labor market, in that the most productive workers match with the firms that can use their skills most productively. Most notably, large firms pay their workers more than small firms do. This is not a new development — Henry Moore did a study of Italian women in the textile industry one hundred years ago and found that women at plants with more than 500 workers made about 40 percent more than women in plants with under 20 workers.
Moore went further than to just note the extra pay, stating, “…as the size of the establishment increases, the condition of the laborer improves in all directions — his wages rise, he is employed a greater number of days in a year, his employment varies less from month to month, and his hours of labor, per day, decrease.” (A hundred years ago, even if a person studied women, he felt compelled to masculinize anyone with a job.)
In the century since Moore went to Italy, many studies have found the same basic idea in many different work settings around the world. In the United States in 1993, the average man at a company with over 1,000 employees made 11 percent more than a man at a company with 100 to 500 workers. Some of this discrepancy can be explained by other factors — people at bigger firms tend to stay at the same company longer, have a bit more education, and are much more likely to belong to a union and work full-time. But these figures are also an understatement of the extra pay from working at a large firm, where fringe benefits are generally more generous.
It is incontrovertible that there is a strong and nearly universal relationship between firm size and wages, but why? There are surely a number of contributing factors, several of which have nothing to do with positive assortative mating. That is, there are several good reasons that a person would receive a higher wage from a large employer than that same person would from a smaller company. One possible explanation is that working at big companies might simply be less fun — think “Dilbert” — so people have to be paid a premium to be willing to work there instead of in a more enjoyable work environment.
Furthermore, it’s possible that big firms simply have deeper pockets and either cannot or do not find it worth the trouble to bargain as hard on salaries. We might expect this to be true simply because of who does the bargaining. If I work at a big company and go in to ask for a raise, it does not come out of my boss’s own pocket. At smaller firms, however, the wage rates are likely to be set by the owner, or at least someone with a bigger stake in the company, who has strong incentives to guard expenses closely.
But while those explanations are important and surely a part of the firm size/wage relationship, positive assortative mating is also a major driver of the firm-size premium. The types of people who work at big firms are the types of people who would make more money wherever they work. In addition to the data above showing that people at large companies are more educated and have more work experience, many studies have also shown that people who work at larger firms are simply “better” (more productive) workers. That is, big employers are able to attract employees who would be more productive than the typical worker, no matter where they worked.
For example, one detailed and careful analysis of half a million French firms found that most of the difference in wages between big and small firms can be explained by the fact that “better” workers work at larger firms, even holding constant such characteristics as education and labor market experience. Other researchers looking at other countries have reached similar conclusions.
We can safely conclude that there is positive assortative mating between workers and firms in the sense that more skilled workers go to bigger firms. Put another way, suppose Addison is the most educated, the hardest working, or simply the smartest one of our four friends and we rank the others the usual way. Our best guess would be that we would find Addison working for a Fortune 500 company, Bailey at a medium-sized regional business with maybe a thousand employees, Chris at a small company with 20 people, and Devon as a sole proprietor operating out of his house. Remember, however, that these are only averages, and that firm size is not the only factor determining productivity. There are, of course, plenty of very productive people at small companies and many unproductive people working at big companies.
People work with people like themselves. Before that, they go to undergraduate schools with other people of similar intelligence and accomplishments. In fact, we “track” people in academics from an even younger age.
My children’s school puts the best math students together in a group and separates them out from the other students. So, if Addison and her friends are high school students and we rank them by innate math ability, my kids’ high school (and most of its peer institutions) believes in positive assortative mating. They set up classes where Addison and Bailey are in one class while Chris and Devon are in another. It’s not just math, of course. By senior year, many of the students are taking only Advanced Placement classes, while others are not taking any. At this point, Addison and Bailey see each other regularly in a variety of classes (and compare notes on how their applications to Princeton are going), while Chris and Devon also have plenty of classes together but rarely see the other two.
The main reason for this sorting is fairly obvious and completely different from why we see positive assortative mating on dating sites and in marriages. In those settings, we pair off on the basis of who we want to spend time with. But in choosing classrooms, we group kids so that they will learn the most as a collective. If we put Devon and Addison in the same math class, the teacher would either have to go so slowly for Devon’s sake that Addison’s learning would be impeded, or move the material along quickly enough to keep Addison engaged and lose Devon in the process. The total amount of learning is greatest when Devon and Chris are in the same class and Addison and Bailey are in a different class.
But this sorting comes at a cost to Chris and Devon. The magnitude varies substantially depending on where or when you look, but economists and other social scientists have generally shown there are important “peer effects” in schools. If Addison is added to the group of students in a class, everyone learns a little more; if Devon is added, everyone learns a little less. The overall level of achievement across the whole class is highest when Addison is in a class with other high achievers, but Devon pays at least a small price for being excluded. Positive assortative mating in the form of classroom tracking may be optimal, but it is yet another factor that makes moving up the socioeconomic ladder increasingly difficult because people in Devon’s position are denied an opportunity to move up.
For better or worse, fair or unfair, the world is full of positive assortative mating. People like to live with, go to school with, and work with people like themselves, and they are generally more productive when they do (at least across certain dimensions).
We may wish our inclinations were otherwise so that economic mobility could be greater, but we are fighting nature. Our colleagues in the physical sciences have shown that Red Crossbills (birds), Icelandic Threespine Sticklebacks (fish), and Water Striders (bugs) show positive assortative mating by bill and body size. Why should we humans be any different?