Economics, Must Reads, recent

Rich Like Me: How Assortative Mating Is Driving Income Inequality

It may be useful to open this topic with an anecdote. Some ten years ago, I found myself in an after-dinner conversation, lubricated by wine, with an American who had been educated at an Ivy League college and was then teaching in Europe. As our conversation drifted toward matters of life, marriage and children, I was initially surprised by his statement that whoever he had married, the outcome in terms of where they lived, what type of house they owned, what kind of holidays and entertainment they would enjoy, and even what colleges their children would attend would be practically the same. His reasoning was as follows: “When I went to [Ivy League institution], I knew that I would marry a woman I met there. Women also knew the same thing. We all knew that our pool of desirable marriage candidates would never be as vast again. And then whomever I married would be a specimen of the same genre: They were all well-educated, smart women who came from the same social class, read the same novels and newspapers, dressed the same, had the same preferences about restaurants, hiking, places to live, cars to drive and people to see, as well about how to take care of the kids and what schools they should attend. It really made almost no difference socially whom among them I married.”

Trend Guessing Quiz: Assortative Mating Over Time

And then he added, “I was not aware of that at the time, but I can surely see it now.”

The story struck me then and stayed in my mind for a long time. It contradicted the cherished myths that we are all deeply different, unique individuals, and that personal decisions such as marriage, which have to do with love and preferences, matter a lot and have a big effect on the rest of our lives. What my friend was saying was precisely the opposite: He could have fallen in love with A, or B, or C, or D, and ultimately would have ended up in virtually the same house, in the same affluent neighborhood—whether in Washington, D.C., Chicago or Los Angeles—with a similar set of friends and interests, and with children going to similar schools and playing the same games. And his story made a lot of sense.

Of course, this scenario assumed that people who attended the same college would couple up. Had he dropped out of college, or not found anyone suitable to marry there, the outcome might have been different (say, a house in a less affluent neighborhood). His story dramatically illustrates the power of socialization: Almost everyone at the top schools comes from more or less equally affluent families, and almost everyone adopts more or less the same values and tastes. And such mutually indistinguishable people marry each other.

Recent research has documented a clear increase in the prevalence of homogamy, or assortative mating (people of the same or similar education status and income level marrying each other). A study based on a literature review combined with decennial data from the American Community Survey showed that the association between partners’ level of education was close to zero in 1970; in every other decade through 2010, the coefficient was positive, and it kept on rising. A different database provides another perspective on this trend; it looks at marriage statistics for American women and men who married when they were “young,” that is, between the ages of twenty and thirty-five. In 1970, only 13 percent of young American men who were in the top decile of male earners married young women who were in the top decile of female earners. By 2017, that figure had risen to almost 29 percent.

Percentage of U.S. men aged 20 to 35 in the top male decile of labor earnings who married women aged 20 to 35 in the top and bottom female deciles by labor earnings, 1970–2017

At the same time, the top decile of young male earners have been much less likely to marry young women who are in the bottom decile of female earners. The rate has declined steadily from 13.4 percent to under 11 percent. In other words, high-earning young American men who in the 1970s were just as likely to marry high-earning as low-earning young women now display an almost three-to- one preference in favor of high-earning women. An even more dramatic change happened for women: the percentage of young high-earning women marrying young high-earning men increased from just under 13 percent to 26.4 percent, while the percentage of rich young women marrying poor young men halved. From having no preference between rich and poor men in the 1970s, women currently prefer rich men by a ratio of almost five to one.

Percentage of women aged 20 to 35 in the top female decile by labor earnings who married men aged 20 to 35 in the top and bottom male deciles by labor earnings, 1970–2017.

In a very ambitious 2017 paper, Pierre-André Chiappori, Bernard Salanié and Yoram Weisstried tried to explain both the rise of assortative mating and the increasing level of education among women (which contrasts with a lack of increase in educational attainment for men). They argued that highly educated women have better marriage prospects, and thus, there is a “marriage education premium,” which is perhaps as important as the usual skill premium that education provides. While the skill premium is, in principle, gender neutral, the marriage education premium is, the authors argue, much higher for women. Underlying this must be greater “pure preference” for homogamy among men because if that did not exist, the rising education level of women might be as much of a deterrence in the marriage market as an attraction.

There is a further link between, on the one hand, assortative mating, and, on the other hand, increasing returns to investment in children, which only more educated couples are able to provide. They can, for example, expose their children to a learning-conducive atmosphere at home and introduce them to cultural experiences that less-educated parents may have little interest in (concerts, libraries, ballet), as well as to elite sports. The importance of linking these seemingly unrelated developments—women’s education, greater work participation by women, assortative marriage patterns, and the increasing importance of early childhood learning—is that it illuminates one of the key mechanisms of within-generation creation of inequality and its intergenerational transmission. If educated, highly skilled, and affluent people tend to marry each other, that by itself will tend to increase inequality. About one-third of the inequality increase in the United States between 1967 and 2007 can be explained by assortative mating, according to research by Koen Decancq, Andreas Peichl and Philippe Van Kerm. For countries in the OECD (Organization for Economic Cooperation and Development), assortative mating accounted for an average of 11 percent of increased inequality between the early 1980s and early 2000s.

But if, in addition, the returns to children’s early education and learning are sharply rising, and if these early advantages can be provided only by very educated parents, who, as the data show, spend much more time with their children than less educated parents, then the road to a strong intergenerational transmission of advantages and inequality is wide open. This is true even if—and it is important to underline this—there is high taxation of inheritance, because inheritance of financial resources is merely one of the advantages that the children of educated and rich parents enjoy. And in many cases, it may not even be the most important part. (Although, as I argue elsewhere, taxation of inheritance is a particularly good policy for leveling the playing field and increasing equality of opportunity, it is an illusion to believe that such taxation will by itself be sufficient to equalize the life chances of children born to rich and poor parents.)

* * *

High income and wealth inequality in the United States used to be justified by the claim that everyone had the opportunity to climb up the ladder of success, regardless of family background. This idea became known as the American Dream. The emphasis was on equality of opportunity rather than equality of outcome. It was a dynamic, future-oriented concept. Joseph Schumpeter used a nice metaphor to explain it when he discussed income inequality: We can see the distribution of incomes in any one year as being like the distribution of occupants who are staying on different floors of a hotel, where the higher the floor, the more luxurious the room. If the occupants move around between the floors, and if their children likewise do not stay on the floor where they were born, then a snapshot of which families are living on which floors will not tell us much about which floor those families will be inhabiting in the future, or their long-term position. Similarly, inequality of income or wealth measured at one point in time may give us a misleading or exaggerated idea of true levels of inequality and can fail to account for intergenerational mobility.

The American Dream has remained powerful both in the popular imagination and among economists. But it has begun to be seriously questioned during the past ten years or so, when relevant data have become available for the first time. Looking at twenty-two countries around the world, Miles Corak showed in 2013 that there was a positive correlation between high inequality in any one year and a strong correlation between parents’ and children’s incomes (i.e., low income mobility). This result makes sense, because high inequality today implies that the children of the rich will have, compared to the children of the poor, much greater opportunities. Not only can they count on greater inheritance, but they will also benefit from better education, better social capital obtained through their parents, and many other intangible advantages of wealth. None of those things are available to the children of the poor. But while the American Dream thus was somewhat deflated by the realization that income mobility is greater in more egalitarian countries than in the United States, these results did not imply that intergenerational mobility had actually gotten any worse over time.

Yet recent research shows that intergenerational mobility has in fact been declining. Using a sample of parent-son and parent-daughter pairs, and comparing a cohort born between 1949 and 1953 to one born between 1961 and 1964, Jonathan Davis and Bhashkar Mazumder found significantly lower intergenerational mobility for the latter cohort. They used two common indicators of relative intergenerational mobility: rank to rank (the correlation between the relative income positions of parents and children) and intergenerational income elasticity (the correlation between parents’ and children’s incomes). Both indicators showed an increase in correlation between parents’ and children’s incomes over time (rank to rank increased from 0.22 to 0.37 for daughters and from 0.17 to 0.36 for sons, and intergenerational income elasticity increased from 0.28 to 0.52 for daughters and from 0.13 to 0.43 for sons). For both indicators, the turning point occurred during the 1980s—the same period when U.S. income inequality began to rise. In fact, three changes happened simultaneously: increase in inequality, increase in the returns to education, and increase in the correlation between parents’ and children’s incomes. Thus, we see that not only across countries, but also across time, higher income inequality and lower intergenerational mobility tend go together.

So far, we have only looked at relative mobility. We should also consider absolute intergenerational mobility, that is, the change in real income between generations. Here, too, we see a decline: absolute mobility in the United States declined significantly between 1940 and the 2000s, as a result of a slowdown in economic growth combined with increased inequality. We should keep in mind that absolute mobility is very different from relative mobility, since it depends largely on what happens to the growth rate. For example, absolute mobility can be positive for everyone if the income of every child exceeds the income of their parents, even if the parents’ and children’s positions in the income distribution are exactly the same. In this example, complete intergenerational absolute mobility would coincide with a complete lack of intergenerational relative mobility.



Excerpted from Capitalism, Alone: The Future Of The System That Rules The World, by Branko Milanovic, published by Harvard University Press. Copyright © 2019 by the President and Fellows of Harvard College.



  1. Charles Murray already wrote this book.

    Our increasing freedom and prosperity are allowing us to associate with the like-minded more than ever before. This includes “assortative mating”. The result is, as the author says, widening “inequality”.

    But this article, like so many on the Left, fails to establish that that’s a problem. Rather, circular reasoning is used to claim that inequality is bad because it causes inequality.

    It resembles the Left’s use of the word “capitalism”, which has become formulaic: “There’s something about the world I don’t like. Isn’t capitalism awful!” No attempt at linking the disliked thing to “capitalism”, as opposed to innumerable other candidate causes, is even made.

    We do have a problem with people clustering into bubbles, but “inequality” isn’t it. We also do have an “inequality” problem, but the problem isn’t the inequality itself; it’s that “inequality” is the statistical symptom of the casualties of the welfare state: a permanent underclass of drug addicts, dependents, criminals, and homeless vagrants.

  2. What’s really weird is that Americans tend to marry Americans, Brits other Brits, Russians other Russians, Chinese other Chinese…it must be a global cabal of evil capitalists victimizing the most well off population of humans history has ever known.

    The other bizarre thing is that people tend to marry people around the same age. Catastrophe in the making! I propose that all us old white guys marry young Asians so as to fix the inequalities.

  3. I’m pretty sure it’s the patriarchy’s fault.

  4. I’m sorry but I cannot take any more of this rubbish form American thinkers who are bloody stupid that they think that social mobility can be determined by looking at income. It beggars belief.
    Inequality of income is a measure of inequality of income and not much else. To look at social mobility you have to be able to judge evidence of people’s rise and fall in social class and status. There are measures for social class that have been in use for years. After all Maggie Thatcher won the 1979 election by famously winning over the majority of the C2s, i.e. the skilled workers.

  5. Any examination of the tendency of children to resemble their parents in terms of income that doesn’t mention the very high level of genetic heritability of intelligence reveals the author to be either ignorant or, much more likely, deliberately neglecting it as it doesn’t fit an a priori position.

  6. @AndyJ This is ignored by most people. Intelligence is massively heritable so it’s no surprise that childrens income will be similar to their parents irrespective of their upbringing.

    This is also the reason why I believe historically there was much greater movement between classes as you had intelligent parents with low incomes because they never had educational opportunities. They had the raw intelligence but they could not use it. A good example would be the men who blossomed in working mens clubs in Britain when they finally got a chance to use their brains.

    It’s no surprise then that the children of the parents who had the educational opportunities that their intelligent parents did not have would end up with higher incomes. The law of diminishing returns then starts to apply.

    It follows that it is unreasonable to expect the children of parents who in their day were not intelligent and therefore voluntarily did not pursue further education to be really bright , pursue higher education and end up in a higher income group than their parents.

    As an aside I believe that higher education is massively overrated but unless you have practical talents it can be difficult to go another route.

  7. What you say:

    Compared to what you’re reacting to:

    Do you see how you strawmanned @jdfree49’s position in an obvious and rude way? Why even bother posting such a pointless, immature, raging comment? And why did you stop using capitalisation a couple of weeks ago?

  8. @DSzczesniak, You imply that @jdfree49 supports rampant, exploitative, unregulated capitalism and insult him/her on that basis. If this is the case, please provide links to such statements. Until you do, I will agree with @Stephanie who states, in other words, that are misrepresenting @jdfree49’s views. It seems you two have some history: Don't Feed the Troll .

    Do you have any critique of the observations and arguments in Charles Murray’s book which @jdfree49 cited? What about @AndyJ’s contention that success is highly dependent on intelligence, which is largely inherited? I don’t know any commenters here who are arguing for exploitative, union-destroying capitalism - though a few would argue against the subset of unions who are arguably exploitative and business-destroying.

    As far as I know, most commenters here argue for equality of opportunity, not enforced equality of outcome - since there are numerous personal factors which affect outcome. Some of those factors can’t be changed and others can, by choice and effort.

    As far as I know, no commenter here who is critical of the Left (as most are) is supporting old-style “lord it over the plebs” conservatism, or the numerous exploitative excesses of capitalism. Not everyone supports all the views of Jordan Peterson, but I think quite a few would agree with his passionate statement of dismay at the reality of the lower performing, unhappy, unhealthy end of the spectrum of human potential and achievement. This was part of a podcast:

    Quote from Jordan Peterson:

    “You know, one literature I know extremely well is the literature on IQ. Because I am also a practicing clinician and have dealt with people across the entire cognitive spectrum, from barely cognitively able to function independently, to genius level cognitive ability, I have some sense of the actual differences between people – and they are absolutely massive.

    “Who wants that??!

    “I mean, its necessary. Everyone can’t have every gift. If we are going to have gifts at all, that means some people are going to have them and some people aren’t. A fair bit of that is very uncomfortably rooted in our biology. And no-one can be happy about that, because it contributes to that arbitrary unfairness we were talking about earlier."

  9. Extreme levels of wealth inequality leads to potentially destructive social tensions, but the article seems to have been written as if any inequality of outcome is a problem. The author seems to assume that the overriding goal is to “equalize the life chances of children born to rich and poor parents.”. I think this refers to the chances of achieving desired outcomes for all such children, not equality of opportunity, in which no-one faces barriers due to poor educational opportunities, or discrimination - a condition in which, naturally and desirably, the smartest, wisest, hardest working people do achieve higher than average outcomes.

    Part of the reason for the social tensions with high inequality is that parents and indeed entire social classes work to protect their own, by steering resources of their own and of the public to benefit themselves, their class-mates and their children. This is understandable behaviour, but I think the state has a responsibility to prevent such actions from leading to inequality of opportunity - especially for children and adolescents.

    Nations face a fundamental challenge when they achieve their desired high standard of living, with access to contraception and reduction of biases against women: births per woman drops below the 2.1 replacement rate - at least among the most productive/wealthy people creating and enjoying these successes. This is arguably less of a problem than a very high figure leading to unsustainable population growth, but it is a serious problem in that the workforce which built the civilisation does not sustain itself in later generations.

    If the productive momentum can’t be sustained by immigration - taking the more productive people from other nations - and if diminishing population and ability to support the elderly is not OK, then this is an existential problem for the nation. Nations compete with other nations as well as cooperate with them, and those which are less productive than other nations can’t contribute or compete, and therefore wind up in debt and ultimately dependent on or invaded by other nations.

    Assuming the high incomes of these educated inter-marrying people correlate positively with their actual contribution to society (though not not everyone who accrues a lot of money does contribute in a real sense) and that these marriages of educated, productive, people produce children at a higher rate than the societal average, then this “assortative mating” contributes decisively to the solution of one of the greatest challenges faced by developed nations.

    Inequality of outcome is ubiquitous in the lives of humans and other species. Policies based on the utopian insistence that this cannot be allowed invariably results in distortions and inequality of opportunity.

    Natural selection driven by suffering and death, or at least failure to reproduce, by a large proportion of children was one of the two mechanisms why humans evolved brains more capable than those of our chimp-like anscestors. The other mechanism was sexual selection - picking a mate who has a better than average chance of raising children both in quantity and with the qualities required to survive and repeat the process.

    I am no expert on evolution, but it seems to me that sexual selection as a means of improving the genetic fitness of a population, for whatever environment it finds itself in, involves less death and suffering than with natural selection. Of course both mechanisms occur together. 6 million years or natural selection alone evolving a spectactularly more capable brain could only occur with most children dying young, or at least failing to bear children if they survived as adults. I think most of this would be due to dying before even young adulthood.

    As far as I know, sexual selection contributes to the genetic changes (progress) which lead to modern humans highly productive brains by way of causing births to be more likely to result in long, reproductively successful, lives. For a given amount of progress achieved in this way, this involves less death and suffering than if natural selection alone was at work.

    Of course, the unsolved barrier to civilised society maintaining a population with productive genetics is that civilisation aims to reduce death and suffering for all, thereby reducing the death rate before adulthood which was the mechanism of natural selection. This article might be interpreted as an argument against sexual selection - the less cruel form of developing and maintaining desirable population genetics.

    So to the extent that educated and economically successful marrying and child-bearing couples mentioned here are genuinely productive (not using their intelligence, charm, attractiveness and efforts to exploitatively gain the fruits of other people’s productive efforts), be happy. They have found a way of bringing more children into society with the benefits of whatever genes are partly responsible for their productivity, plus their personal handed-down version of culture of care which is good for children.

    If we relied entirely on natural selection for maintaining population genetics at about the level we got from hunter-gatherers and farmers, before civilisation and industrialisation, then coupling would be random, more children would need to be born and a greater proportion of them would need to die or at least not have children, than would be the case with this successful use of sexual selection.

    Human happiness and the physical, mental and emotional characteristics which are needed to build and defend societies which provide such happiness, are partly determined by passing on good culture, political systems, knowledge etc. Geography and access to natural resources are important too. The rest, it seems, is genetic.

    None of this is nice or an argument for institionalised inequality of opportunity. The state, for instance, should not be setting limits on parents’ efforts at home to care for and educate their children. Nor outside the home.

    Assuming education is publicly funded, in principle, ideally, every child would have equal opportunities to succeed. It is a natural priciple of civilisation to devote more public efforts to the weakest. However, society depends in part on its brightest and hardest working people leading really productive lives - and somehow raising children at the same time - so there are arguments for spending more resources than average on children who are likely to contribute most to national productivity.

    I can’t see a clear path through these conflicting moral and practical concerns. My point is that the author seems to be arguing against sexual selection which can maintain and improve the productive genetics of at least an important fraction of the population in a manner which involves less effort, suffering and death than relying on natural selection (largely banished).

    The other alternative is state control of births to maximise the likelihood of genetically desirable outcomes (eugenics). The theory is simple and promising, but trying to imagine the implementation, in terms of practicalities and moral conflicts with equality of opportunity and other civilised ideals leads quickly to a prohibitive number of headaches and contradictions.

  10. So side-by-side quotations of what @DSzczesniak said compared to what @jdfree49 said counts as “cherry-picked,” and @Robin-Whittle’s polite and considered response to @DSzczesniak’s defense counts as “piling-on.” You must think “left-wing economic positions” are particularly weak if they must be shielded from such gentle criticism.

    You can defend those positions all you like, but it is not appropriate to expect people not to respond, particularly when you are twisting someone’s words in order to portray them as a terrible person, as @DSzczesniak did in his first comment.

    But I suppose you’re not adverse to personal attacks based on this gem. It’s quite revealing that you use the term “exclusive,” as if there’s something excluding certain undescribed people from living here or in @PeterfromOZ’s suburb. Mine is a beautiful suburb, a safe suburb, but not an exclusive one. There are retirees, young professionals, unemployed people on disability, lawyers, musicians, students, academics, construction workers, small business owners. The only qualification it takes to live here is to be able to afford rent, which my husband and I do despite him being a labourer and me being a student.

    This picture you want to paint of people you clearly perceive as “privileged” in the popular lefty sense obscures the reality of people’s lives. My father was in jail for drug trafficking when my mother was pregnant with me, I grew up in my grandparents’ basement, and I’m three generations out from illiterate goat-herders from the Atlas Mountains. But I can see the Opera House from my shower, it must be the result of crony capitalism!

  11. I had quite the unfortunate upbringing, actually. I’ve seen many people who had a much better start do much worse than I have (so far). As many people have been trying to explain to you, individual factors make a huge difference on outcomes.

    The system works for everybody to a degree, making what we call “poor” straight-up decadent by the standards of a couple hundred years ago, but it works better for some than for others. It favours those who make the right choices: finish high school, get a job, don’t have kids until you’re married. Americans who do those 3 easy things have less than 5% chance of staying in poverty, and a 75% chance of joining the middle class. Additional factors such as intelligence, choice of university major, drive, industriousness, conscientiousness, ect play a significant role in an individual’s outcome. No system can equalize these factors without crushing the human spirit.

    Shaming the rich for marrying who they want to marry is insane. As is looking around at the most prosperous society in human history and claiming you need to be a billionaire to enjoy it. Your resentment is blinding you. It looks like it’s affecting your mental health, and I imagine that is bleeding out into other aspects of your life as well, including your employment prospects. You have value as a person and you have what it takes to succeed in this world. There is a place for you where you will be happy and successful, if you adopt the right values and put in the hard work. We all want to see you get there.

  12. How do you explain that California hasn’t had anything close to right-wing politics for decades, but is incredibly unequal and the poor live in squalor? Real life doesn’t match your narrative.

  13. So I thought this was a months-old article I had read already, brought into Quillette Circle by someone reading old articles, so I didn’t look at it before commenting. Turns out it is new. Damn shrinking brain cells!! Here’s my fuzzy thoughts:

    I find the premise kind of odd: sure, it’s interesting to look at marriage trends, but the tone seems to imply that people’s choice in spouses is a problem, which just seems really personal and inappropriate.

    Wouldn’t this just reflect that it is disproportionately the wealthy who are getting married these days? It’s not like women’s desire to marry rich men have increased, and there’s a limited supply of those anyway. Women are still shacking up with poor guys, they just aren’t marrying them. Weddings are expensive and even though you can cheap out on them (I spent $500), it’s just not a priority for most couples who can’t afford the dream wedding.

    Isn’t it a good thing that people are marrying people they have more in common with? Isn’t this the equality in marriage that the left has been demanding? Or would they prefer we go back to men picking their housewives out of the girls they meet at church?

    So money that was already subject to income tax and that will be taxed again at the point of purchase should have a third (or more?) layer of taxation put upon it? Even though the author claims that they recognise that the advantages conferred by affluence are less about money than about upbringing? This seems misguided and punative.

    Overall, there were some good pieces here but I would have put them together differently: it seems to me there’s a case to be made here that the drop in marriage rates among the poor exaggerate wealth inequality. I think it’s more useful to council poor couples on how to increase their children’s chances of success than to try to tell rich people to stop marrying each other.

  14. i was raised by a single mother. i would have rather had both parents, but modern families dont allow for it. this is a definite part of women entering the workforce - they think they are unstoppable and can have the entire world because they are single mothers. no. the only thing that they can provide better than a male parental figure is empathy. the economy barely values their work the same, because like it or not, women cant do the sorts of jobs that men have historically been doing, because the jobs are insanely hard work - logging, mining, and construction all require a great deal of strength. sure, technology is making things easier for both genders, but like it or not, this is the result.

    sure, maybe some extra income is nice. but i would rather have more time to spend with my family while i was still growing up. maybe these mental illnesses would be less severe.

  15. Once again, the dirty secret of American meritocracy remains unmentioned. As education rates increased to effectively 100%, and transportation and media technology were popularized, we pulled the higher intelligence young people out of the farms and factories, and allowed them to rise to their proper income level. And once that is done, there will be very few high intelligence young people left to draw up out of poverty - intelligence is limited by genetics, and more intelligent people have more intelligent children. As low intelligence people tend to have low intelligence children. The low hanging fruit has effectively been harvested. The End.

Continue the discussion in Quillette Circle

43 more replies


Comments have moved to our forum