Demographics + Dating

A disproportionate discovery

I recently stumbled on an essay called “Math Explanation of the Shidduch Crisis.” The shidduch crisis is a phenomenon in Orthodox Jewish communities where there seem to be more marriage-age women than men - not enough guys to go around.

The essay laid out this elegant little model. Three variables: sex ratio at birth, population growth rate, and the average age gap between spouses. Plug in the numbers for Orthodox communities (high birth rate, men marrying women 3+ years younger) and out pops a female surplus. Simple demographics, no cultural hand-waving explanations required.

My mathy brain lit up. I love when social phenomena turn out to have tidy quantitative explanations. But then I thought: wait. Does this apply to secular America? What do the numbers actually say here?

The tension I’m trying to resolve

Here’s something that’s bugged me for a while. I genuinely believe men and women aren’t that different from each other. Similar distributions of intelligence, ambition, kindness, extraversion¹. I get annoyed when people attribute every behavioral difference to the Essential Gender Nature.

And yet. The heterosexual dating market feels asymmetric in ways that seem non-arbitrary.² My single female friends are picky because they can afford to be. My single male friends are grinding - optimizing profiles, reading dating advice, treating it like a skill to be developed. There’s an energy of scarcity on one side that I don’t see on the other.

I’ve always been a little uncomfortable with this observation. It pattern-matches to redpill nonsense about “sexual marketplace value” and I don’t want to lend credence to that framework.³

This came up recently when I wrote about why I stopped asking men out. I’d spent a while believing the feminist math - if women ask men out more, we get better outcomes! I tried it. It selected for passive men who let me do all the work⁴. So I shifted to a more traditional approach: signaling interest clearly but letting men take the visible step.

The response was... a lot. Comments and DMs telling me I was being a coward. That I needed to GROW UP and accept that getting what we want requires DOING UNCOMFORTABLE THINGS. That this was embarrassing. That I was making excuses.

Here’s the thing: I agreed with them for years. I *was* that person, telling other women to just ask men out already. The switch was illuminating, which is why I wrote about it.

But something about the pushback bothered me beyond the tone. It was treating a behavioral pattern as a character test. Women who don’t adopt male courtship strategies aren’t just making a different choice - they’re failing at virtue. They’re cowards. They lack maturity.

What if that framing is completely wrong? What if the asymmetry I observed isn’t about courage or effort or who’s Doing The Work? What if it’s not about gender differences at all? What if it’s just... math?

Three numbers

The dating pool imbalance in the U.S. might come down to three demographic facts.⁵

1. More boys are born than girls.

   For every 100 girls born in the U.S., about 105 boys are born. The CDC tracked this from 1940 to 2002 and found it consistently between 104.6 and 105.9. That’s roughly 80,000 extra male babies per year, every year. It adds up.

2. The birth rate has been declining.

   Births have dropped about 2% per year since 2015. Each year’s cohort of babies is smaller than the year before.⁶

3. Men typically partner with women a couple years younger.

   Pew Research found that husbands and wives are about 2.2 years apart on average (2022 data). This has shrunk over time - it was 4.9 years in 1880 - but it’s still there.

None of these facts are controversial. The question is what happens when you combine them.

Why the age gap matters

If everyone dated someone exactly their own age, the sex ratio at birth would be the only thing affecting the dating pool. 105 men per 100 women, a modest male surplus, end of story.

But the age gap means you’re comparing different birth cohorts. A 30-year-old man dating 28-year-old women is competing against men born in his year for women born two years later.

In a declining population, earlier cohorts are larger. More babies were born 30 years ago than 28 years ago. So there are more 30-year-old men than 28-year-old women.

In a growing population, it flips. Earlier cohorts are smaller. Fewer babies 30 years ago than 28 years ago. More 28-year-old women than 30-year-old men.

The shidduch crisis happens in growing populations where men marry younger women. The male surplus happens in declining populations where men date younger women. Same formula, opposite inputs, opposite outcomes.

(Yes, 2.2 years is just an average - plenty of couples are the same age or have the woman older. But the average captures the center of mass of who’s competing for whom. As long as men are, on average, partnering with younger women, the dynamic holds.)

Running the numbers

I wrote a little Python script to play with this.⁷ Here’s what falls out:

\(\begin{array}{l|c|c} \textbf{Scenario} & \textbf{Men per 100 women} & \textbf{% of pool male} \\ & \textit{(in each other's dating pool)} &\\ \hline \text{Sex ratio only (no decline, no gap)} & 105 & 51.2\% \\ \text{Add 2.2 year gap, stable population} & 105 & 51.2\%\\ \textbf{U.S. conditions (declining + 2.2 year gap)} & \textbf{110} & \textbf{52.4%}\\ \text{Growing population (5%) + 3 year gap} & 91 & 47.6\% \\ \text{U.S. declining rate + woman 2 years older)} & \text{100} & \text{50%}\\ \end{array} \)

With U.S. demographics, there are about 110 men for every 100 women in the dating pool. That means roughly 9% of men mathematically cannot find a partner, even if everyone paired off perfectly.⁸

The growth rate is the biggest lever. It can swing the ratio from 94 to 117 depending on whether the population is growing or shrinking.

Does this match reality?

The model predicts a ~10% male surplus in dating pools. Is that what we actually see?

Pew Research data shows that among young adults (18-29), about 63% of men report being single compared to 34% of women. That’s a huge gap - but it’s partly because men and women in that age range aren’t dating each other. The 28-year-old women are dating 30-year-old men, who aren’t in the 18-29 bracket.

The Institute for Family Studies re-analyzed the data accounting for age gaps, and found the real gap is closer to 10 percentage points - which lines up extremely well with the model.⁹

Other signals point the same direction with various magnitudes:

• 61% of single men are actively looking for a relationship or dates, compared to 38% of single women

• Single women are more likely to say they’re single because they “have other priorities right now”

• The online ecosystem of male dating advice, self-improvement communities, and yes, incel forums is vastly larger than its female equivalents

None of this proves the demographic model is the whole story. But it’s consistent with a world where men face more competition and women have more options.

Incentives, not character

If different places have different demographic splits, you’d expect them to develop different cultural norms around dating.

Jacob Falkovich explores this in his essay The Skewed and the Screwed (a personal favorite of mine). When one gender is in oversupply, the other gender gains selectivity. This has a significant hand shaping courtship norms and political attitudes about relationships.

The Orthodox shidduch crisis and the secular male surplus are the same phenomenon pointing in different directions. In Orthodox communities with high birth rates and large age gaps, women face more competition and the culture develops norms around that (shidduch coaching, pressure on women to be less selective). In secular America with declining births and smaller age gaps, men face more competition and... well, look at the dating advice industrial complex. Look at which gender is told to “level up” and “increase their value.”

This is the part I find genuinely illuminating: we tend to moralize about patterns that are actually just incentive responses.

Think about how American attitudes toward Irish and Italian immigrants changed over the 20th century. In 1900, “Irish Need Not Apply” signs were common. By 1960, JFK was president. Did Americans become more virtuous? Did Irish-Americans become more likeable? No - demographic integration happened. Irish and Italian families moved into the same neighborhoods, worked in the same factories, sent kids to the same schools. Proximity changed incentives. Outgroup became ingroup. The behavioral shift wasn’t about character improvement. It was about conditions changing.

I think a lot of dating behavior works the same way. Women who take a more receptive approach to courtship aren’t being cowardly - they’re responding rationally to a market where they have options. Men who grind on self-improvement and approach anxiety aren’t being desperate or even agenticly virtuous - they’re responding rationally to a market where they face competition. The behaviors make sense given the conditions.

The people telling me I was a coward for not asking men out? They were treating a demographic phenomenon as a character test. They assumed women’s tendency toward yin was a moral failure rather than a rational adaptation. But you can’t shame people into behaving as if they face different conditions than they actually face.

What to do about it

This is a descriptive post, not a prescriptive one. But if you’re a man who feels like you’re on the losing end of this math, there’s an obvious lever: date closer to your own age, or older.

The male surplus is created by the combination of declining population and men dating younger. If you personally reduce your age gap, you’re competing in a less crowded pool. A 30-year-old man interested in women 30-35 faces dramatically less competition than one focused on women 25-28.

This isn’t a character prescription. I’m not saying you should want to date older women, or that there’s something wrong with age-gap relationships. I’m saying: if the math is working against you and you want to change your odds, this is the variable you actually control.

(Interestingly, if we had the same magnitude age-gap on average, but in the opposite direction, with older women, the cohort size would be almost exactly equal, with less than 1% of the population unmatched. Take from that what you will)

What this explains (and what it doesn’t)

The dating market asymmetry I’ve observed isn’t evidence that men and women are fundamentally different creatures with different levels of worth or desirability. It’s not about who’s “trying harder” or who “has it worse” It’s not about who’s brave and who’s cowardly. It might just be a supply and demand artifact created by three boring demographic facts.

Small demographic differences can cause big cultural effects. A 10% oversupply of men doesn’t mean every woman has ten suitors and every man is desperate. But at the margins, it changes the vibe. It explains why my female friends can be choosy while my male friends feel competitive pressure. It explains different dynamics on dating apps. It explains why the dating advice aimed at men emphasizes self-improvement and competition while advice aimed at women emphasizes selectivity and standards.

This doesn’t tell any individual person why they’re single. Macro patterns don’t predict individual outcomes. And it certainly doesn’t justify treating anyone badly.

But it does make me feel less crazy for noticing the asymmetry. And it makes me more sympathetic to both sides - women aren’t delusional for having options, and men aren’t failing for feeling like they have to compete harder. Everyone’s just responding rationally to the actual conditions they face.

The next time someone wants to explain dating dynamics through gender essentialism or moral character, I’m going to point them at three numbers: 105, -2%, and 2.2 years.

No evo-psych required.

1. I lean towards the ‘same mean, different variance' hypothesis - men and women have similar averages on most traits, but men might have wider distributions (more at both extremes) causing over-representation in fields that select for very high values of a trait. But that’s another post.

2. This post describes population-level patterns in heterosexual dating with age gaps. None of it is prescriptive - if you’re queer, or date older partners, or aren’t looking for a relationship, the math doesn’t say you’re doing anything wrong. I’m focusing on the het/age-gap case because that’s where these demographic dynamics show up.

3. This sentence has caused quite a bit of controversy among early readers with less negative opinions of SMV than me.

4. I think, basically, this can be a good strategy for people of all genders WHEN PAIRED WITH the assumption that you’re willing and able to stop dating people who are even kinda poor fits for you, quickly and without emotional hardship. This is not me.

5. I say “might” because this is a simplified model. Three variables obviously can’t capture everything about dating markets. But sometimes simple models are surprisingly powerful, and I think this one is worth taking seriously.Immigration complicates this picture - immigrants add to the population in ways that don’t follow the birth cohort logic. The sex ratio of immigrants also matters. I’m bracketing this for simplicity, but it’s a real limitation of the model.

6. Immigration complicates this picture - immigrants add to the population in ways that don’t follow the birth cohort logic. The sex ratio of immigrants also matters. I’m bracketing this for simplicity, but it’s a real limitation of the model. Typically, immigration to the U.S. skews male, so is unlikely to resolve a male-surplus effect.

7. https://github.com/cynnamon-gh/Shidduch-extension

8. This assumes a simplified world where everyone’s trying to pair off 1:1 with someone in a typical age range. In reality, lots of people are happily single, not looking, queer, in non-traditional arrangements, etc. The model doesn’t predict how many people will be single overall - just the gap between male and female singleness rates among those who are looking.

9. A note on percentages vs. percentage points, since this confused me at first: The IFS data shows roughly 50% of women unpartnered and 60% of men unpartnered. That’s a gap of 10 percentage points (60 minus 50), not 10 percent. The model predicts this gap - it says that whatever the baseline rate of singleness is (due to all the other factors that affect whether people partner up, assuming they affect men and women equally), there should be about 9-10 percentage points MORE men single than women, purely from supply constraints. So the model and the data are measuring the same thing: the marginal difference, not the absolute levels.