Do Japanese People Not Want Children? | An Ideal of 2.25 Against a Reality of 1.14

Hello, this is Ryuta Hamamoto from TIMEWELL. A map colouring countries by their fertility gap went widely viral recently, with Japan in the darkest red. Elon Musk amplified it, noting he had been warning about this for years.

My first reading was that this showed how severe Japan's fertility decline has become. Going back to the underlying data, I had it wrong. This map does not measure low fertility.

Read correctly, it says something harder about Japan. Let me work through it.

What the map actually measures

The source is "The Global Fertility Gap," published by the Institute for Family Studies in February 2019 and written by Lyman Stone¹. The subtitle on the map reads:

Estimated Total Fertility Rate minus Average Ideal Fertility of Reproductive Age Women

Actual fertility minus the average ideal. It measures the distance between what women say they want and what they get, not the fertility level.

The proof sits inside the map. Niger is shaded almost the same red as Japan. Niger's total fertility rate is above six and its population is growing fast. If this were a map of demographic viability, Niger could not share a colour with Japan. Niger is red because ideals there are higher still, and actual births fall short of them.

The dataset behind it gathers 727 data points across 141 countries from 86 sources spanning 1936 to 2018, with the map showing the 2000 to 2018 average¹. The preferred survey question is of the form: if you had enough resources for it, how many children would be ideal for you personally to have. Stone concludes that in the vast majority of countries women end up with fewer children than they call ideal, and estimates 2.7 million "missing births" worldwide each year¹.

This distinction is not good news for Japan. "Low viability" would put Japan alongside other countries sinking together. "Cannot have the number they want" points at a country's own design.

Japan's numbers: an ideal of 2.25, a reality of 1.14

How large is Japan's gap? Here we do not need a foreign dataset. Japan runs a survey measuring precisely this.

The National Institute of Population and Social Security Research conducts the National Fertility Survey every five years. From the 16th round, fielded in June 2021²:

Measure	16th survey (2021)	Change
Average ideal children per married couple	2.25	Continuing gradual decline
Average intended children per married couple	2.01	Flat
Completed fertility, wives aged 45–49	1.81	Down from 1.86
Average desired children, unmarried men	1.82	Down from 1.91
Average desired children, unmarried women	1.79	Down from 2.02

And the reality. Japan's Ministry of Health, Labour and Welfare reported on 3 June 2026 that births in 2025 numbered 671,236, a tenth consecutive record low, with a total fertility rate of 1.14, also a record low and a tenth consecutive year of decline³. Deaths were 1,589,489, leaving a natural decrease of 918,253, the nineteenth straight year of decline³.

An ideal of 2.25 against a realised rate of 1.14. The distance is more than a whole child, and that distance is exactly why Japan is dark red on the map.

What strikes me is the size of it. Couples say two or more is ideal, and say they intend to have two, yet society delivers 1.14. Folded into that difference are both the people who never marry and the couples whose plans do not survive contact with reality.

One faint positive signal is worth noting. Marriages in 2025 came to 489,119, a second consecutive slight increase³. In a country where births remain closely tied to marriage, that is worth watching.

Why they cannot: the reason 52.6% gave

The strength of the Japanese survey is that it asks couples directly why they will not have the number they call ideal. Among couples whose intended number falls below their ideal, the most selected answer was²:

Raising and educating children costs too much: 52.6%

More than half. The survey also shows the reasons shifting with the wife's age. Where the wife is under 35, economic reasons dominate: childrearing costs, education costs, housing, work. Where the wife is 35 or older, physical reasons take over, notably not wanting to give birth at an older age, and wanting a child but being unable to conceive².

Those two sets are not separate. Postpone for economic reasons and you arrive at a biological constraint. When you are young there is not enough money; by the time there is, the age barrier has arrived. Consistent with that, couples who have undergone fertility testing or treatment rose from 18.2% to 22.7%, roughly one couple in 4.4².

So the substance of Japan's fertility gap is largely money and time, and the age constraint they produce between them. It is not a preference against children.

I am not dismissing value change. The same survey shows unmarried women's desired number falling below two for the first time, and sharp declines in support for norms like "married couples should have children"². But with married couples still naming 2.25 as ideal, preference alone cannot explain 1.14.

This is not changing values. It is a design failure

What follows is my own view.

I have heard "young people's values changed, so nothing can be done" many times. That explanation has a property its audience finds very convenient: it removes the possibility of action. If preferences changed, policy is powerless, and nobody has to answer for anything.

But the data does not describe a preference problem. Ideal 2.25, intended 2.01, actual 1.14. If people want and cannot, what stands in the way is not preference but constraint. And constraints, unlike preferences, respond to design.

There is a second reason I call it a failure. The same survey clearly shows where things did improve. The share of wives who stayed employed through the birth of their first child rose from the low fifties to 69.5% within five years, and 79.2% of those who stayed used parental leave². Build the system, change the operations, and behaviour actually moves. There is a domain where that has been demonstrated, and yet 52.6% still name cost. Knowing a lever works and not pulling it is a question of priorities, not of inevitability.

The projections are worth holding in view. Japan's 2023 official projection has the population falling from 126.15 million in 2020 to 87 million in 2070, with the share aged 65 and over rising from 28.6% to 38.7%, and the total dropping below 100 million in 2056⁴. A society shrinking by more than 30% in fifty years will not run on today's design.

Musk's warning, and the criticism of it, together

On the post that started this. Musk has raised the subject before.

On 7 May 2022, responding to reports of Japan's largest recorded population drop, he wrote that unless something changes to cause the birth rate to exceed the death rate, Japan will eventually cease to exist, and that this would be a great loss for the world⁵. That August he added that population collapse from low birth rates is a far bigger civilisational risk than global warming⁶. In 2021 he gave $10 million to the University of Texas at Austin to support demographic research⁶.

The claim also drew direct criticism. Tobias Harris, then at a U.S. think tank, argued in 2022 that the real concern about Japan's demographic future is not that Japan will cease to exist but the profound social dislocations occurring as the population moves to a lower level⁵. United Nations projections still have world population rising to 9.7 billion by 2050 and about 10.4 billion in the mid-2080s⁷. "Global population collapse" and "one country contracting fast" are different claims and should be kept apart.

I find the criticism the more useful frame. Whether Japan physically disappears is not a question for any of our lifetimes. The question is whether healthcare, logistics and infrastructure hold as the working-age population keeps shrinking. Not extinction. Degradation. And that one is much closer.

Musk's warning did useful work in drawing attention. What deserves the attention, though, is not whether a nation disappears but how to unlock a population that wants more children than it can have.

What to do

Three things, in the order I would prioritise them.

First, reduce the uncertainty of the cost. The economic reason chosen by 52.6% is not only about money today. It is about not knowing what the next eighteen to twenty-two years will cost. People behave conservatively toward uncertain expenditure, and predictability can matter more than the amount. Cash transfers are legible, but one-off payments do not reduce uncertainty. For long-horizon costs like education, a design where the total is knowable in advance should move behaviour harder.

Second, make it possible earlier. Given a documented path where economic constraints cause postponement and postponement runs into biology, support delivered later works less well. Measures aimed at disposable income and housing costs for people in their twenties should outrank measures aimed at people in their forties. That is my reading of the age-stratified results, and I hold it as opinion.

Third, redesign work on the assumption of a smaller workforce. This is the part that belongs to companies. Even if fertility rose tomorrow, those children reach the labour market in twenty years. The workforce decline of the next two decades is already fixed. In a society heading for 38.7% aged 65 and over, processes designed around today's headcount will not hold.

That third one is our daily work. Taking processes that were scaled by adding people and rebuilding them for a world where you cannot. AI and automation are tools for that, not fashion. If you want to see how far your own operations could bend, our AI literacy assessment is a place to start the inventory, and organisational design is what WARP is for.

Closing: start from "they want to, and cannot"

The record. The viral map measures the distance between ideal and actual, not the fertility level¹. Japanese couples name 2.25 as ideal and 2.01 as intended², while the 2025 fertility rate was 1.14 on 671,236 births, a tenth consecutive record low³. The leading reason for falling short of the ideal was cost, at 52.6%². Population is projected at 87 million by 2070⁴.

The single thing I wanted to land is that Japanese people have not stopped wanting children. They want them. They cannot have them. Any policy that does not start from both halves of that sentence will aim at the wrong target. Reduce it to a values problem and the debate ends, but it has not been resolved. It has been abandoned.

Companies can only do so much; moving national policy is not our job. But rebuilding work for a smaller workforce is available to any company today. And it connects directly to whether working parents can have the number of children they want. Leaving processes built around long hours untouched while blaming demographics on policy does not hold together.

Start by modelling whether your core operations still run with 20% fewer people ten years from now. If they do not, that is where to begin. To talk through the design, get in touch and we can work through it together.

References

All figures come from primary statistics and their publishers. Musk's statements and the criticism of them are confirmed through press coverage.

Footnotes

1. The Global Fertility Gap — Lyman Stone, Institute for Family Studies — 25 February 2019 ↩ ↩² ↩³ ↩⁴

2. The 16th Japanese National Fertility Survey, Summary of Results — National Institute of Population and Social Security Research — fielded June 2021 (Japanese) ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸

3. Vital Statistics 2025, Monthly Report Annual Total (Approximate Figures) — Ministry of Health, Labour and Welfare — published 3 June 2026 (Japanese) ↩ ↩² ↩³ ↩⁴

4. Population Projections for Japan (2023 Revision), Summary of Results — National Institute of Population and Social Security Research (Japanese) ↩ ↩²

5. Elon Musk tweet on Japan doomed by low birthrate provokes anger — The Japan Times — 9 May 2022 ↩ ↩²

6. Elon Musk tweets "population collapse" — Newsweek ↩ ↩²

7. World Population Prospects — United Nations Department of Economic and Social Affairs ↩