The world is running out of children

The world is running out of children.

Not in the way doomsayers said it would in the 1960s, when Paul Ehrlich was warning about a population bomb. The opposite. The global total fertility rate is now around 2.24 and falling fast. The US hit a record-low 1.599 in 2024. South Korea is at 0.72, the lowest of any country ever recorded in peacetime. Hungary is at 1.38 after spending roughly 6% of GDP trying to fix it. About 71% of humanity now lives in countries below replacement. The Lancet projects that by 2100, 97% of countries will be there too.

This is the most important demographic shift of our lifetime, and the conversation around it is a mess. The doom industry has it. The culture warriors have it. The contrarian Twitter accounts have it. What it does not have, mostly, is the calm, sourced, careful explanation that a story this big deserves.

That’s what this project is for.

The number nobody is amplifying

Here is the single most important data point in the field, and almost nobody quotes it: the UN’s 2025 survey across 14 countries found that people, on average, say they want about two children, and end up having about 1.4.

That’s a 0.6-child gap between desire and reality. Reframe everything around that number and the question changes completely. The question is not “should people have more kids,” which is a culture-war fight nobody wins. The question is why aren’t people having the kids they say they want. That is a problem-solving frame. It’s the frame the OECD and the IEA’s Mind the Fertility Gap paper both center, and the one this project will live in.

The second number worth sitting with: no developed country has reversed fertility decline back to replacement using policy. None. Not Hungary, which spent 6% of GDP. Not South Korea, which has spent over $280 billion since 2006 and watched its TFR fall from 1.13 to 0.72 every single year since the programs began. Not France, whose 4%-of-GDP family-policy regime sustained a TFR near 1.9 for two decades and is now down to 1.6. Not the Nordic gold standards. Norway, Sweden, Finland, and Iceland are all in collapse together.

This is genuinely new in human history. Every previous fertility decline was either temporary or reversed. This one shows no signs of reversing in any country that has lived with it for twenty years. That is the story. And the honest acknowledgment that we don’t yet know how to fix it is, I’d argue, the foundation any serious project on this topic has to start from.

The three lazy narratives, and why each of them breaks

Most of what’s loud on this topic falls into one of three boxes. All three are wrong in different ways.

The first is the right’s easy answer: ban abortion, pay birth bonuses, exhort women to do their patriotic duty. Romania ran the cleanest version of this experiment in 1966. Ceaușescu’s Decree 770 outlawed abortion and contraception, and the period TFR doubled almost overnight, from 1.9 to 3.7. Within a decade it had fallen back. The “decree babies” generation famously filled Romania’s orphanages. Cohort fertility ended up modestly elevated and the human cost was catastrophic. The modern, gentler version of the same answer is South Korea’s $280 billion. It produced the most expensive demographic failure on record.

The second is the left’s easy answer: immigration will fix it. It will, mathematically, for about thirty years. Then it stops. The world fertility rate is now 2.24 and falling toward replacement. By 2050 the global pool of young migrants will be small, and the receiving countries that depend on it will be competing for a shrinking flow. Alice Evans, who is as far from a nativist as anyone writing on this, has been clear about this for years. Immigration buys time. It does not fix the underlying equation.

The third is the tech-bro answer: AI and automation will solve it. Productivity gains will indeed soften some of the fiscal arithmetic of an aging population. Fewer workers per retiree matters less if each worker is dramatically more productive. But productivity does not solve who carries the culture forward. It does not solve who maintains civic institutions, who staffs the human networks that hold a society together, who teaches the next generation what their grandparents knew. A society with brilliant machines and no children is not a problem AI can fix. It is the problem itself.

What we actually know might help

Here is what the evidence-based literature, read honestly, suggests can chip at the gap. Not solve it. Chip at it.

Earnings-related parental leave at high replacement rates appears to produce real quantum effects, not just timing shifts. Estonia’s 2004 reform, which gave leave at 100% of pre-birth income capped at three times average salary, moved second and third births meaningfully, especially for educated women, per Klesment and colleagues in the European Journal of Population. It is the cleanest natural experiment in the literature.

Affordable, high-quality childcare on the French and Quebec model has a sustained positive effect of roughly 0.1 to 0.2 children, per the IFS analysis of the French regime and the 2024 cohort analysis of the Quebec Parental Insurance Plan.

Housing matters more than almost anyone realizes. A University of Toronto paper estimates roughly half of the recent US fertility decline is attributable to rising housing costs. American apartments have shrunk from 1,300 to 1,000 square feet since 2010, and family-sized units are vanishing from the stock. Singapore’s “no flat, no child” effect is empirically real.

Remote work is the strongest new signal in the field. A 2026 paper by Aksoy, Barrero, Bloom and coauthors, across 38 countries, finds an estimated +0.32 children when both partners work from home at least one day a week. In the US, +0.45. The authors estimate WFH accounts for about 8% of current US births. That is the largest accidental pro-fertility policy in the developed world.

Then there are the cultural outliers. Israel sustains a TFR near 3.0 across both religious and secular populations. In Georgia, the Patriarch’s 2007 baptism promise produced a +0.3 children-per-woman effect that a 2025 paper in the Journal of Population Economics finds real and persistent. And Mongolia moved from below 2.0 to roughly 3.0 in a decade with no major cash policy at all, by conferring high public status on motherhood. These are the cases policy alone cannot replicate, and that anyone serious about the topic has to grapple with honestly.

Add it all up and the highest-EV package (leave, childcare, housing, workplace flexibility, plus whatever cultural support a society can muster) might plausibly move us from 1.4 to 1.7 or 1.9. That is meaningful. That is not 2.1.

What this project is

This project is an attempt to cover the most important demographic story of our lifetime with the calm, the rigor, and the intellectual honesty it deserves. Hannah Ritchie’s posture toward climate data, applied here. Tim Urban’s patience with hard ideas, applied here. Kurzgesagt’s commitment to getting the numbers right, applied here.

What we are not doing: doom. Tribal pronatalism. Manosphere bait. Lectures about civilizational decline. Easy ideological answers from any direction. If you’ve come for a side, this isn’t going to be your favorite newsletter. If you’ve come because you actually want to understand what is happening and what, if anything, helps, you’re in the right place.

You’ll get one careful essay a week. Numbers always in context, sources always cited, no hedge-words, no melodrama. When the evidence is ambiguous, I’ll say so. When a study is fragile, I’ll flag it. When a country looks like a success story and isn’t, I’ll show the receipts.

Subscribe below. Read the next piece. Send it to one person who’s been thinking about this and can’t find anyone serious to read. That’s how this gets built.