Why are fertility rates dropping?
Fertility is falling in almost every country in the UN record, and in some of them faster than any demographer predicted. The causes are reasonably well understood, they are not the same everywhere, and they are not the kind of thing a birth dataset can measure. This page keeps those three facts apart.
- 129of 236 countries below replacement
- 67%of world population lives there
- −5.61children per woman since 1950: Taiwan
- 16.3world births per 1,000 people, 2023
First, the size of the fall
In 1950 only a handful of countries were below the replacement level of about 2.1 children per woman. In 2023 there are 129 of 236, and they hold 67% of the world's population. The fall has been steepest where it started latest: several countries compressed a transition that took Europe a century into a single generation.
The table below is the twelve largest falls since 1950 among countries of more than 20 million people, read straight from the UN's own figures for the two years. Small territories are left out because a population of thirty thousand can move several children per woman on very few births.
| # | Country | 1950 | 2023 | Fall | Births |
|---|---|---|---|---|---|
| 1 | Taiwan | 6.48 | 0.87 | −5.61 | 132,000 |
| 2 | South Korea | 6.06 | 0.72 | −5.34 | 236,000 |
| 3 | Saudi Arabia | 7.58 | 2.28 | −5.30 | 546,000 |
| 4 | Iran | 6.97 | 1.70 | −5.27 | 1,170,000 |
| 5 | Philippines | 7.13 | 1.92 | −5.21 | 1,840,000 |
| 6 | Thailand | 6.36 | 1.21 | −5.15 | 591,000 |
| 7 | Peru | 6.94 | 1.98 | −4.96 | 539,000 |
| 8 | Syria | 7.62 | 2.71 | −4.91 | 522,000 |
| 9 | Malaysia | 6.43 | 1.55 | −4.88 | 436,000 |
| 10 | Türkiye | 6.47 | 1.63 | −4.84 | 1,070,000 |
| 11 | China | 5.81 | 1.00 | −4.81 | 8,910,000 |
| 12 | Mexico | 6.71 | 1.91 | −4.80 | 2,040,000 |
What the fall is attributed to
These are the explanations demographic research converges on. They are attributions from the literature, not quantities measured in this dataset, which holds births, birth rates, fertility rates and population, and nothing about why anyone had a child. Their relative weight differs by country, and in most places several act at once.
Childbearing starts later
The age at a first birth has risen almost everywhere. Later starts compress the years available for further children, and they depress the fertility rate of any single year even when the number of children women eventually have falls by less. It is a timing effect that makes current rates read lower than completed family size will turn out to be.
Education and paid work
Longer schooling and higher labour-force participation among women are among the most consistent correlates of lower fertility across countries and over time. Both delay childbearing and raise what economists call the opportunity cost of a birth: the earnings and career progress a pause costs.
Contraception and control over timing
Reliable contraception turns family size from something that happens into something decided. It does not by itself set the number people choose, but it closes the gap between the number they want and the number they have, and historically that gap ran in one direction.
Urbanisation
On a farm a child is labour; in a city a child is an expense, in space as much as in money. The share of humanity living in cities has more than doubled since 1950, and urban fertility is lower than rural fertility in nearly every country where both are measured.
Housing and childcare costs
Where housing is expensive relative to earnings and childcare is scarce or costly, surveys repeatedly find people having fewer children than they say they want. This is the cause most often argued about, because it is the one that policy can plausibly move, and the evidence that policy does move it is mixed.
What families are expected to look like
The number of children considered normal has fallen, and it has fallen in countries with very different economies, welfare systems and religions. Norms also travel: fertility often falls in neighbouring countries and language groups at similar times, which is hard to explain through local economics alone.
The causes are not the same in every country
Reading one explanation across the whole world gets most countries wrong. Taiwan's fall from 6.48 to 0.87 happened alongside rapid urbanisation and a large rise in women's education, over a few decades. Much of Europe fell below replacement earlier and more slowly, and has stayed in a narrow band since. Fertility is still above 4 across much of sub-Saharan Africa, where the transition began later and is now clearly under way.
So a country's position in the table is not evidence for any particular cause acting in it. What the data supports is the shape and the timing of each country's fall; the reasons have to come from research into that country, and the honest form of the claim keeps them separate.
What the projections assume
The UN's medium scenario assumes fertility continues converging towards a low level in most countries, and it produces world births of about 110 million a year in 2100 against 132.1 million in 2023. That is an assumption carried forward, not an observation, and recent history is a warning about how uncertain it is: several East Asian countries fell well below what earlier projections had them reaching.
This site shows the medium scenario only, marks every year from 2024 onward as a projection, and does not draw a prediction interval around it. The medium variant does not come with one, and inventing a band would give the projection a confidence it has not got.