Assortive mating and income inequality

More than in the past, we have doctors marrying other doctors,  rather than nurses, basically because of an increase in assortative mating for education. Ceteris paribus, this would tend to cause greater income equality among families. Is it the main driver of increasing income inequality?

Not at all.  Most of the increase over the last 30 years has been among business executives and people working in finance.  Since 1979, 58% of the expansion of income of the top 1% of households has this origin.  For the top 0.1% of households, it’s been 67%.

Business executives are making more relative to the average worker than they used to. The average compensation of a Fortune 500 CEO was 20 times that of his average employee in 1950, 42 times larger in 1980, 350 times larger today.

Now if hedge funds and Fortune 500 companies were being managed by married teams, dynamic duos like Justinian & Theodora or Burns & Allen, you could attribute this change to increased assortative mating.  In fact, I should probably write a best-selling book on this counter-intuitive idea, even though (especially though!) it never, ever happens.

Now I’m about to say something a little dangerous – so get your nitroglycerin pills ready.

Maybe those finance guys and CEOs are delivering enormously more value than they did in the 1950s!

For those remaining readers that haven’t died laughing, increased assortative mating probably has contributed to income inequality.  Just not very much.  Changes in the tax code, outsourcing, automation,  smothering the board of directors in cream, and inattentive stockholders all matter more.

This entry was posted in assortative mating and tagged . Bookmark the permalink.

53 Responses to Assortive mating and income inequality

  1. refuguist says:

    Mr. Cochran,

    I’m glad you’re addressing this issue as it’s been puzzling me lately. I’ve recently read that individual income inequality, measured by the Gini ratio, is virtually unchanged over the past sixty years, while household income inequality, as we all know, has risen dramatically since 1970. At first blush this appears to be evidence for the assortative mating hypothesis (not genetically, of course, but as far as the income accounting is concerned). On the other hand you have the rise in CEO compensation, and then automation, which, in theory at least, sounds as if it would reduce the value of unskilled labor.

    I can’t seem to come to a conclusion about what it all means. Maybe you could give us some more detail about how you put the data together to reach your conclusion. It appears to me that you consider the increase in finance income to be the smoking gun. But does this necessarily override the individual vs household income inequality differences? The financial sector may have expanded excessively, but I would tend to see that as a separate issue.

    I trust your views, but if you have a little time, perhaps you could make your argument more explicit to help me connect the dots.

    • gcochran9 says:

      What you read is incorrect, for three reasons. First, it’s almost certainly based on Census data, which does not include capital gains. Capital gains accounts for a much larger fraction of very high incomes. Second, many people have used public-use Census data. Although the Census asks people their income (other than capital gains), the data released (and to a lesser extent their internal data) suffer from topcoding. This means that they tell you the fraction of people who have an income higher than X, but no more. So if their top bracket is 1% and they tell you that the top 1% have an income higher than 350,000 a year, you’re missing all the income above 350K in your analysis. IRS data does a a better job, and with careful use of internal Census data, Census and IRS estimates agree pretty well.

      Over the last 20 years, almost all the increase in income equality has been caused by increased income in the top 1% of the distribution.

      That’s two reasons. The third reason is that the source you looked at was lying.

      • refuguist says:

        Thanks, Mr. Cochran, that’s enlightening.

        Would it then be correct to conclude that the increase in income inequality between unskilled and skilled labor, e.g. between the automated and the automators, is more or less insignificant compared to the increase in inequality between the owners of capital and the rest of the workforce? Put otherwise, insofar as automation is having an effect on income distribution, the effect is not to transfer income from the unskilled to the skilled, but from labor as a whole to the owners (and traders) of capital. Perhaps I’m jumping to conclusions.

        P.S. While I wouldn’t want to distract from your blog’s focus on genetics and history, it would be appreciated if you wrote or linked to a few reliable, fact-based articles on economics and politics. I’ve found it all but impossible to locate analysis that’s not ideological or incompetent, and like most people, I haven’t the time or resources to properly crunch the data myself. You seem to have a good grip on these matters.

        • gcochran9 says:

          “increase in inequality between the owners of capital and the rest of the workforce”

          Pretty much.

          As for talking sense about economics and politics, not on this blog, generally speaking. Maybe somewhere, sometime, but I would spend a lot of time saying “I don’t know”, and how interesting would that be? Sure, probably nobody else knows either, but they don’t let that stop them. Me it stops.

          General dishonesty. Yep, it’s a problem.

    • anon says:

      Proliferation of single, female-headed households… seems like it would have a significant statistical effect.

  2. georgesdelatour says:

    Assortative mating probably reduces social mobility more than it exacerbates income inequality.

  3. Patrick Boyle says:

    Yes, yes everyone knows that the big money is in finance. But finance is boring and dull. It can never fire your soul like the quest for valid inference or a beautiful algorithm.

    Maybe that’s why I’m so damn poor.

    Most comedian’s wives were not funny and most of the emperor’s wives were more of a danger than an asset. Livia, Messalina, and Agrippina set the deadly pattern. Theodora and Gracie were the exceptions that proved the rule.

  4. JayMan says:

    And some people are convinced assortative mating is a horrible thing…. :\

  5. marcel says:

    Maybe those finance guys and CEOs are delivering enormously more value than they did in the 1950s!

    Of course they, but as ever, the question is cui bono?
    Who is receiving the enormous increase in value?

    Res ipsa loquitur!

  6. Robert Ford says:

    A perfect time to ask Greg about this:
    Thoughts? My guess is that you’ll say “Stupid.”

    • Sandgroper says:

      I looked at that and figured that the classifications are too broad to be very informative (both high school or less, both some college, both college), so I went here:

      …and the numbers are stupid. Not the very broad trend, people in professions need higher IQ than taxi drivers, obviously, but when you look at the detail, the numbers are rubbish. There’s a sampling problem, and a classification problem, or else there’s some problem in the way that IQs have been estimated from wordsum scores – you need a higher IQ to be a stenographer than a civil engineer? Really?

      Part of the problem may be that engineers are notoriously bad on verbals, but that cannot be all of it. I have not known too many stenographers who are math geniuses.

      I’m tempted to say that rather than stupid, the numbers are just rubbish. Well, you’d need to be stupid to reproduce those numbers and think they mean anything, or just totally uncomprehending.

      The other thing is that now, you simply cannot enter a science or technology profession without post-graduate qualifications – just a bachelor’s degree will get you a technician’s job, or a teaching job in high school. So it will not even be enough to count the fields people are getting a first degree in, to make sense of the data, you need to classify by post-grad qualification.

      • Sandgroper says:

        The other thing is – I don’t know if this happens in America (I guess I could find out) but it is evident that it happens somewhat de facto anyway – not all college degrees are equal for the purpose of obtaining employment. In at least some other places, there are lower status universities that equip people with education in less intellectually demanding things. No one says that straight out, but everyone with a brain knows it – you can just look at the international university rankings, and you can look at the things they teach. Employers pay attention to those things. And among the top universities, some disciplines are much more intellectually demanding than others – that is stating the blindingly obvious. That is not always reflected in pay levels (simple examples – you need more brains to study physics than engineering, but engineers earn more, on average. There are lots of science PhDs who can’t get solid, well paid jobs. In Australia, tradesmen can earn more than engineers. During the recent sustained mining boom, unskilled building labourers were earning more than engineers.)

        So I don’t think that interrogating data at such a coarse level is going to tell you much. Yes, there is assortative mating, and there always has been, but you won’t see much of an effect in terms of change in household income inequality coming from that. The data are way too fuzzy.

      • RobertFord says:

        Ok, thanks a lot. I suspected as such. Appreciated.

  7. dave chamberlin says:

    It is my opinion that increased assorted mating of the very high IQ couples will keep us sufficiently supplied with geniuses until we fix stupid.

  8. Steve Sailer says:

    Herrnstein theorized that there had been an increase in assortative mating since Olden Days in a 1971 Atlantic article. That was toward the tail end of a four decade long period of low inequality.

  9. misdreavus says:

    So this means that libertarians never have any clue what they’re talking about, huh.

    I’ll keep that in mind from this point forward.

  10. little spoon says:

    I guess elite schools are more selective these days as they have more applicants, so they can afford to raise their SAT standards. Thus, the average Yale grad couple may have a higher SAT percentiles now than the average Yale grad couple of 1982. However, I think the SAT of the 1980s correlated with IQ better than the modern SAT. Not sure.

    Even if assortative mating is the same as it was in the 1950s, couples now are more likely to be double earners. In the 1950s, an IQ 150 man might have chosen a lower paying career like violin playing and then the couple would be living on a violin player’s income. But if two high IQ people are earning, that increases the chances that one of them has a high paying job or that the combination of their salaries exceeds the 1% mark. So in that sense, perhaps the same level of assortative mating now contributes to more income inequality than before.

    I would think that assortative mating is increasing at the extreme of intelligence. In the early part of the 20th century, it is true that intelligent people probably found each other through socialization and got married, but they didn’t have MIT and Harvard admitting the top .1% of women and putting them in proximity with the top .1% of men in large numbers. Thus, I think there are probably more .1% IQ couples now than in the past. But I can’t claim that they are reproducing.

    • Steve Sailer says:

      I think Galton would have been pleased that institutions like Yale Law School and Harvard Business School are now coed and put together a lot of high IQ couples.

      • gcochran9 says:

        If he knew the facts, he would have been horrified. What do you think the TFR of women in those places is? 0.6?

        See next post.

      • FredR says:

        I never understood why he made such a big deal about how great increasing assortative mating would be, particularly in that unpublished novel “Kantsaywhere” he wrote. Shouldn’t the idea be to increase the frequency of desirable traits, rather than just spreading the variance?

  11. TC says:

    Gregory Clark pooh-poohs the threat of assortative mating in The Son Also Rises, and he explicitly criticizes Murray’s Coming Apart on several pages. He doesn’t think “social mobility,” properly understood as “status” mobility, has changed much in the past century.

    Few people noticed that it’s essentially The Bell Curve for social democrats. In fact, it’s the only substantial *hereditarian* counter-narrative to TBC that I’m aware of. He doesn’t think the SAT mattered all that much for social mobility. There is “always already” a hereditary elite … is essentially the argument. Ashkenazis, for example, were overrepresented among physicians as early as the 20s and 30s, even more so than the scions of Ivy League graduates.

    It’s bracing work, but Clark’s perspective is too macro, zoomed-out, basically Olympian, to satisfactorily explain what happened with the WASPs. They didn’t regress to the national mean, they’re still elite, but they don’t rule the country like they used to. Murray is useful in that regard.

    • georgesdelatour says:

      There can be very high status occupations that are relatively low paying (the Archbishop of Canterbury is paid £67,000 per year, for instance). And well-paying occupations that are considered somewhat disreputable (Larry Flynt and Richard Desmond are certainly wealthy, but is their chosen profession “high status”?).

      Originally no one got paid for being a Member of Parliament in Britain. “Noblesse oblige” and all that. Then in the 1906 General Election, 29 Labour MPs were elected, many from very poor backgrounds. The Labour leader, James Keir Hardie, was the son of a ship’s carpenter and a domestic servant, for instance. So salaries were introduced.

  12. panjoomby says:

    i like to call it “income diversity.”

  13. Orthodox says:

    Demographic change, family breakdown, severing the gold link to the dollar in 1973 (breaking the restrictions on the financial sectors ability to create money), globalization that benefits capital and punishes labor in wealthy countries. The last two being the most important if you want to finger the financial sector and financial assets as the culprit.

  14. dave chamberlin says:

    “Over the last twenty years almost all the increase in income has gone to the top 1%.”
    Once upon a time that factoid would have pissed me off. It doesn’t anymore primarily because I just see the world as it is rather than what I wish it would be. I am trying to figure out what changed me. It is more than just losing my youthful idealism, I think I’ve been influenced by reading the words of very bright people who honestly admit what they don’t know. After enough exposure to that then the contrast of ignorant bullshitters becomes much more recognizable. Hmmmm. If all you ever hear are babblers of simplistic right or left wing ideology that are forever proposing simple answers to complex problems how are you to know there is another way of looking at the world. Maybe blogs like this are having more effect, at least on the people bright enough to follow, than anyone imagines.

  15. Peter Connor says:

    Most of the change has been a product of crony capitalism. Before Congress made it impossible to take over companies without innumerable hurdles, it put limits on how bad corporate governance could be. Then banking, normally not a very safe or profitable endeavor, got a blank check from Congress and the Fed, so huge salaries for losing money. Finally, we have all the directly subsidized entities, green and otherwise.Kinda like ancient Rome.

    • gcochran9 says:

      Your picture is largely incorrect.

    • Greying Wanderer says:

      “Before Congress made it impossible to take over companies without innumerable hurdles, it put limits on how bad corporate governance could be.”

      Other way round. In the 80s hostile takeovers were made easier and combined with the rise of “shareholder value” as the sole metric they pretty much forced mass off-shoring.

      (nb libertarian economics is mostly rich people lying)

      • georgesdelatour says:

        I’d agree with that. Michael Milken’s Junk Bonds were used to fund hostile takeovers. Once the takeover was successful, the new owner would sell off most of the company; (1) in order to pay back the bank loan funding the takeover, and (2) to make a quick profit on the deal. James Goldsmith’s takeover of Crown Zellerbach in 1985 followed exactly this asset-stripping script.

      • Peter Connor says:

        The folks who lobbied successfully in Congress against all these mergers and acquisitions were the same incompetent executives who had entrenched themselves in underperforming corporations with hand picked Boards.

        • gcochran9 says:

          Relative CEO salaries were much smaller back in the days when hostile takeovers were extremely rare, just not done. Economic growth back then was both A. much higher and B. more evenly distributed.

          So you are, on balance, wrong. Will you now fall down in a swoon?

    • anon says:

      Be sure not to conflate community banking, a traditionally staid and straightforward undertaking, with investment banking (i.e. Wall Street).

  16. thinkingabout it says:

    Assortative mating of highly sciencey men and women produces autistic children.

  17. Toad says:

    If we make a hypothetical that IQ can be inherited, then in that case wouldn’t consanguineous marriages be the most assortive mating of all?

  18. peppermint says:

    Look, I love red pills as much as any sperglord with no real connection to any kind of community. So I will take seriously the notion that CEOs deliver more value than they did in the ’50s.

    How, man? How?

    How do you explain the expensive mistakes like Carly Fiorina and Tim Cook?

  19. Luke Lea says:

    Not sure if you have seen these data but he top incomes belong to guys running hedge funds and financial institutions according to these two guys at U. of Chicago:

    It would seem (if I remember correctly) that the vast majority of the people among the top 12,000 income tax payers were either entrepreneurs or their heirs:

    12,000 top income in America

  20. Steve Sailer says:

    Dear Luke:

    Thanks. So it looks like in 2007 there were 18,000 households making over $10 million that year. They accounted for 8.3% of taxable income and paid 9.9% of income taxes. They paid a lower rate on average than people making 200k to 500k, which seems nuts. I don’t understand any principled reason why income tax marginal tax rates are progressive in the middle and upper middle ranges and then stop going up above a few hundred thousand in income. It would seem sensible to keep adding a few percent to the tax rate with every order of magnitude increase in income.

  21. Peter Connor says:

    Capital gains, municipal bonds, qualified dividends, and various methods of shifting income forward in time.

    • gcochran9 says:

      Mostly cap gains. Stock options. Cap gains are taxed a far lower rate than regular income, which is why we see fraudulent methods of turning salary into cap gains, such as back-dating options. Which, as it turned out, virtually every corporation did, which is why nobody was every thrown in the slammer for it.

  22. Sandgroper,

    Read the actual commentary accompanying that post. That is by far the most viewed post I’ve ever put together, yet 90% of commenters don’t get the obvious joke (obvious because I explicitly describe it as such) because they just look at the table without looking at anything else.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s