Modern hypergamy is probably not a real thing.

(content warning: kicking the Culture War hornet’s nest pretty darn hard)

People like to argue about Jordan Peterson on the internet, so I’ll start this post with a Jordan Peterson video.

“Hypergamy” in the sense that Peterson is talking about has become a common theme in culture war debates over why young people are having so little sex, among other things.

The claim is that women, but not men, will only date, marry, or have sex with partners whose social status or attractiveness is higher than theirs. Peterson fears that this can lead to a sort of sexually Hobbesian state of nature in which a small number of high-status men get all the women, low-status men get none, and women are unhappy with their partners’ divided attention.

The proposed solution is “enforced” monogamy — enforced not by coercion, but by social norms — which supports a harmonious society in which most men are married, most women get the monogamous attention of one man, and children are raised in stable homes.

This argument is novel in that it claims to support socially conservative sexual norms without appealing to God or traditional ideals of purity.  And it’s least superficially plausible: When I think of polygamy, I think mostly of societies that are more violent, unequal, and generally unpleasant than monogamous societies.

Ignoring the straw man — that “enforced monogamy” means legal enforcement — the main objection to this argument is libertarian: Respect for autonomy means women must have the right to run their sexual lives as they see fit, and shouldn’t be judged for it; in particular, they shouldn’t be accountable for what men “get”, sexually.

That’s basically correct, I think, but it might not be the whole story.  Society has non-coercive norms about all kinds of things to support good choices and discourage bad ones — from workaday decisions like whether to recycle to intimate ones like how to parent one’s children.  We recognize that there are sometimes difficult tensions between personal choice and social responsibility.

However, I don’t think there’s a real dilemma in the case of sexuality, because my read of the evidence is that hypergamy, in the harmful sense that Peterson is talking about, doesn’t actually exist.

What is “Modern Hypergamy”?

It’s clear that Peterson isn’t talking about “hypergamy” in the historical sense:

https://en.wikipedia.org/wiki/Hypergamy

The dictionary definition of hypergamy involves women marrying men from a higher social class. That’s a real historical thing, but it may not tell us anything about modern sexual behavior. In most historical societies, women’s best way of increasing their own status and resources was to marry men of higher social status, so it’s not surprising that many of them wanted to do so. And in many cases, what they wanted didn’t especially matter, because their families made the decisions for them. The most straightforward way to solve these issues is to give women more opportunities for advancement, not to restrict sexual behavior.

But people who believe in what I’ll call “modern hypergamy” generally admit that their definition doesn’t match the historical definition.  So what is it they mean?

Jordan Peterson describes hypergamy like so:

Women mate across and up dominance hierarchies, men mate across and down.

You could read this as a statement about balanced asymmetry between heterosexual men and women; for example, the commonplace that men are, compared to women, relatively more attracted to youth and beauty, and women more attracted to status and wealth.

But that in itself wouldn’t lead to the kind of social disharmony Peterson fears.  If that’s all there was to it, women would try to date across and up status hierarchies, and men would be willing to date across and down.  But there would be no reason to think those outcomes would actually happen on a large scale, because most people would settle for what they could get.

Women with high social status would try to date men with similar social status, and if they couldn’t, they would date men with lower social status.  We’d see some sorting whereby very attractive women ended up with very rich men, and less attractive women with less wealthy men, but everyone would land somewhere.  Metaphorically speaking, the “sexual market” clears, resulting in relative social harmony without any need for conservative sexual norms.

In order to be a serious social problem, “hypergamy” has to mean something more like the following:

Men may have a tendency to seek sexual variety, but women have simple tastes in the manner of Oscar Wilde: They are always satisfied with the best.

That is how Francis Roger Devlin describes female hypergamy in an essay called “Sexual Utopia in Power.”  I’m linking to it indirectly because it appears to be hosted on a site that has links to white nationalists, and the author is chauvinistic in the extreme.  Nevertheless, the idea seems to have persuaded some people whose beliefs are more mainstream, so I’m going to evaluate the argument itself rather than the character of its creator.

Analyzing a concept like “hypergamy”, with an ambiguous meaning that can refer to a number of closely related but subtly different things, brings dilemmas that are hard to get around.  No matter how I define it, there will be some people who can correctly say that my analysis has nothing to do with their definition of the word.

Fair enough.  For purposes of this post, when I say “modern hypergamy” I’m talking about the concept as described by Devlin and popularized by manosphere writers like The Rational Male.  Jordan Peterson’s take on the topic is less detailed, and I can’t say for certain how closely his beliefs match Devlin’s — I will admit to lede-ing with Peterson as a recognizable foil rather than addressing his arguments specifically.

Devlin believes that, without conservative sexual norms, the sexual market does not clear.  Women are only attracted to men who are more attractive than they are, so only the most attractive men have much luck dating:

Once monogamy is abolished, no restriction is placed on a woman’s choices.
Hence, all women choose the same few men. If Casanova had 132 lovers it is because 132 different women chose him. Such men acquire harems, not because they are predators, but because they happen to be attractive. The problem is not so much male immorality as simple arithmetic; it is obviously impossible
for every woman to have exclusive possession of the most attractive man. If
women want to mate simply as their natural drives impel them, they must,
rationally speaking, be willing to share their mate with others.

Thus, as many on the internet put it, twenty percent of men have eighty percent of all sex, and no one is happy: Not the other eighty percent of men who aren’t getting any, not the women who can’t get alpha males to commit, and maybe not even the jaded Lotharios who realize, in their old age, that they never knew love.

(Incidentally, Devlin seems to think “alpha males” are defined mostly by looks; Peterson seems to think they are defined mostly by wealth and status; The Rational Male and other manosphere blogger seem to think it’s a mixture of the two.  This distinction won’t matter much for most of the evidence I’m going to look at.)

The problem for Devlin is that he provides almost no evidence that human mating actually works this way, other than some loose analogies to the evolved behavior of other animals.  Of course, different kinds of animals have different evolved behaviors, and humans are not sage-grouses, so the only way to settle the question is to look at empirical evidence.

What sort of evidence, you may ask?

Sadly, there’s no New England Journal of Things People Argue About on the Internet, so we’re going to have to piece together data from a variety of sources.  In most cases, I’m not the first person to analyze these data for evidence of hypergamy; this post is to some extent an aggregation of studies other people have found.  But in each case, I did my own analysis from scratch.  I’ll start where everyone else seems to start, which — because, internet — is an OkCupid blog post.

The data in that post show that women on OkCupid rate 80% of men as less attractive than “average” (2.5 out of 5), whereas men rate only about 50% of women that low.  This is evidence for some kind of asymmetry, but it’s not clear that whatever is going on resembles hypergamy — women send more than half their messages to men whose photos are rated below 2.5.  That doesn’t seem consistent with the claim that women are only satisfied with the best.  Reproduced without permission:

But maybe the distribution of messages sent by women doesn’t matter very much — should we assume the typical pattern for online dating is for men to send messages, and women to either respond or not?  In which case, we should be looking at response rates:

—

If women’s responses to men were hypergamous, then we should expect to see a steeper curve for men than for women, with very low response rates for the least attractive men, and high response rates for the most attractive men — perhaps an upward-bending parabola, so that even average men get very low response rates.  Instead, we see the most attractive men getting less than two and a half times the response rate of the least attractive men, and the ratio is similar for women.  Even if we re-normed the curve so the men’s average score was 2.5, the slope would curve upward only slightly.  Notably, women’s response rates are far more egalitarian than the rate at which men message less and more attractive women:

In summary: There’s something weird going on with how women rate men’s attractiveness, but the data from the OkCupid blog post does not support the hypergamy hypothesis.

I was able to find at least one scholarly paper that studied online dating behavior; interestingly, it uses the PageRank algorithm to assign objective-ish desirability scores to the subjects.  In the interest of keeping this post only-a-little-bit-ridiculously-long, I won’t discuss it in detail, other than to note that it also found roughly symmetrical behavior for men and for women:

What about people actually having sex in the real world?

It’s hard to say exactly how relevant data from online dating sites is to actual mating.  The more direct way to find out who’s having sex with whom is to ask them, and the General Social Survey is the gold standard in “asking Americans tons of questions.”  It’s been around a long time and has tons of statisticians and social scientists working to make sure the sampling is representative, the questions are asked in reliable ways, and so on.

This article finds no evidence for hypergamy in the GSS, but it’s hosted by a Christian think tank with an obvious “everything bad is caused by people not getting married” agenda, so it’s reasonable to worry that they might be cherry-picking figures.  I already have the GSS data on my hard drive, ‘cuz that’s how I roll, so I checked the numbers myself.  Here is the full distribution, by sex, of answers to the question “How many sex partners have you had in the past 5 years”, for respondents in the year 2016 who either had no partners or had partners exclusively of the opposite sex:

The distributions for men and women are almost exactly the same, with a very slight trend toward men reporting more partners overall.  The distributions for frequency of sex in the past year are also very similar, with a slightly larger fraction of women than men reporting no sex:

What’s interesting is how many opposite-sex partners people report since their 18th birthday:

This chart shows a number of interesting things:  First, I am apparently part of the top twenty percent of men making life miserable for everyone else, which is kind of amazing because I lost my virginity in my mid-twenties and I’m the kind of person who already has the General Social Survey data on their hard drive.

Second, the top twenty percent of men really do report having more partners than the top twenty percent of women — which could be evidence for female hypergamy, except men in every other quintile also report having more partners than the equivalent quintile of women.

It’s not quite mathematically impossible for the mean number of partners to be higher for men than for women — they could be having sex with women outside the United States, or who have since died and thus left the sample, and so on.  But these genuine differences should be much smaller than what we see in the chart; I think there is a simpler explanation.

 

People lie about sex.

Or at least, they mis-remember.  Brown and Sinclair find that men and women respond differently when asked how many sexual partners they have had in their lifetimes — men estimate, and women enumerate.  Estimation tends to overstate the true number of sexual partners, and enumeration tends to underestimate it.

(Incidentally, I just tried estimating and then enumerating, and indeed I overestimated slightly.)

When people are asked how many partners they have had in shorter time periods — within the past year, for example — the differences reported by men and women are much smaller.  That pattern holds true in the General Social Survey data above; the lifetime data shows men having significantly more partners than women all across the distribution, but the past-five-years data does not.  Likewise, the Center for Disease Control finds that that men report a few more sexual partners than women over the past year, but many more partners over the course of their lives.

The GSS and CDC data are unusually good because we can see the entire distribution.  Most studies report only summary statistics, and the whole “men report more partners than women” thing usually makes those impossible to interpret.

For example, this paper reviews the literature and concludes that “empirical studies consistently show the same pattern: women are found to have relatively narrow (low variance) distributions, meaning most of them have close to the average number of sexual partners, while men have much wider (high variance) distributions.”  If that were true, it would be strong evidence for hypergamy.  But let’s take a look at the appendix:

Seven of the nine studies reviewed report means for men and women.  In all seven cases, the sex differences in means are basically impossible.  Of course the variance in number of partners is higher for British men than for British women — the men report having had three times as many partners on average!

Likewise, the authors of this paper note that the different Gini coefficients for men and for women are entirely caused by men reporting more partners than women, but don’t seem to realize the Gini coefficient is an almost meaningless metric under those conditions.

Conversely, summary statistics are sometimes improperly used as evidence against hypergamy — people sometimes point out that if the 80/20 rule held for men but not for women, then women would have a higher median number of partners than men.  Doesn’t the fact that men report higher median numbers of partners than women disprove hypergamy, or even show that “reverse hypergamy” or “male hypergamy” exists?

No, because that reasoning assumes equal means.  If men report unrealistically higher means than women, we can’t draw conclusions by comparing medians.

The Brown and Sinclair study isn’t perfect.  The sample is drawn only from college students, it focuses on the most promiscuous ten percent of those students, and it fails to consider that more promiscuous people might use estimation because they have more partners, which makes it harder to enumerate them.

So we’re left with some uncertainty about the data:

• Neither the short-run data nor the long-run data show any straightforward evidence for hypergamy, but the long-run data in particular is known to be untrustworthy.
• “Men estimate, women enumerate” seems like the simplest explanation, but it’s hard to be sure about that.  If that’s the real explanation, it does not support the hypergamy hypothesis.
• Since we know people aren’t answering truthfully, it’s at least possible that female hypergamy exists, but for some reason people lie in precisely the pattern that would hide that fact from us.  This seems very unlikely to me, but I’m open to theories.

One more data set, because this graphic is really cool.

I mean, just look at it:

That’s a map of sexual partnerships at a pseudonymous high school from a study creatively called “Chains of Affection: The Structure of Adolescent Romantic and Sexual Networks.”  This is about as reliable as data on sexual partnerships gets; every sexual partnership within the sample was verified through direct interview.

The authors weren’t interested in sex differences, so they provide no summary statistics at that level, but fortunately someone counted them all:

Here again we see a familiar pattern: “sexual inequality” is similar for men and for women.

The network diagram is notable for something it doesn’t show us — “harems” in which one male “monopolizes” multiple females.  There’s that one guy who slept with nine women, but seven of those women had other partners.  “Inverse harems” seem to be slightly more common; there are several females in the data set who had multiple male partners that did not, in turn, have other female partners.

Contrary evidence and asymmetrical balance.

Throughout this post, I have carefully limited the scope of my argument to a very specific definition of “hypergamy” based on the idea that women aren’t willing to “date down” and men are.  I’m concerned exclusively with that definition of hypergamy because that’s the kind that could, theoretically, justify monogamous sexual norms as a way to prevent sexual dystopia.

As I noted in the second section, there are other kinds “hypergamy” that could exist without causing major social problems.  For example, if women tend to be more attracted to wealth and status than men are, but they’re nevertheless reasonable about whom they’re willing to date when they find out Idris Elba is already taken, we might call that “hypergamy” of a kind.

But this type of hypergamy does not cause any obvious social problems.  Men and women would sort themselves roughly by how attractive the other sex finds them; men with wealth and status would date physically attractive women, men with average wealth and status would date average women, and so on down the line.  The world I’m describing is not egalitarian — realistically speaking, any world where not all people are equally attractive is going to be that way — but it’s balanced, in that most everyone has a realistic chance to find a partner, even without monogamous sexual norms.

When I search the internet for evidence-based arguments for female hypergamy, essentially all of them are for hypergamy in this narrower sense.  I’m not accusing anyone of committing the equivocation fallacy, because I haven’t seen any specific person explicitly defend the “women won’t settle” version of hypergamy by providing evidence for the “women like status” version of hypergamy.  But any argument for female hypergamy that hopes to be convincing should be explicit about exactly what definition it’s defending.  I have yet to see an argument that provides strong evidence for the “women won’t settle” version of the idea.

I am aware that some feminists object to even the weaker “women like status” definition of female hypergamy, but I don’t think it should be seen as threatening to feminism.  What this definition says is merely that women are neither more or less shallow than men when it comes to attraction, albeit maybe shallow in slightly different ways.

This is not to say I’m fully convinced that women are attracted to wealth and status and men are attracted to youth and beauty — or at least, I’m not convinced that it’s as strong a tendency as many seem to think.  I’m saying simply that I’m not terribly interested in exploring that argument right now, because it doesn’t seem relevant to what kind of sexual norms we should have.

 

The 80/20 rule revisited.

For what it’s worth, I am also not arguing that sexuality is egalitarian or fair.  Of course, sexual “inequality” is more ambiguous than economic inequality: Most people want to be richer, but not everyone wants more partners.  Nevertheless, there are clearly people who are disadvantaged when it comes to finding partners and who suffer because of it.

To put numbers on it: Depending on exactly how I slice and dice the data in the General Social Survey — which time period I use, whether I exclude anyone who has had sex with a prostitute, and so on — I estimate that the most promiscuous 20 percent of men and women each have somewhere between 60 and 80 percent of all sexual partners, not too far off from the Pareto principle.  If Sexghanistan were a country, its Gini coefficient would be around 0.60, similar to highly unequal places like Honduras and South Africa.

But this form of inequality exists primarily within each sex, not between the sexes.  “Chad” and “Stacy” exist, in a sense, but Chad is not surrounded by a harem of desperate Stacies; rather, Chad and Stacy are mirror images of each other’s promiscuity.

I will say that, having recently discovered that I am Chad (or at least, in the top twenty percent in terms of lifetime partner numbers), I don’t feel as though I am a member of a sexual ruling caste.  The question of objective symmetry between men and women is separate from the question of “what is all this like, subjectively, for the people involved?”, which is a topic I may address in a future post.