Why are birth rates falling? What the data shows

In the late 1990s the middle country in the world had 2.8 children per woman. Around 2020 it had 2.0. The number of countries below 2.1, the level at which each generation replaces the last, went from 69 to 114. This article looks at what went along with that fall, where births rose instead, and whether paying families seems to make a difference. The data can show what changed together. It cannot prove what made births fall.

Children per woman, not births per person

The measure here is the fertility rate: the number of children a woman would have if she lived through today’s birth rates at each age. The more common birth rate, births per 1,000 people, also falls just because a population gets older, so it mixes two different things. The fertility rate does not.

Richer countries have fewer children, but that is only the start

The strongest single link is with income. Across 199 countries, the richer the country, the fewer children per woman: r = −0.83, on a scale from −1 (one goes up as the other goes down) to +1 (they rise together). Almost everything else about a country moves with income too, so the useful question is what still goes with fertility among countries of similar income.

Comparing countries with similar income (GDP per person), fewer children go with more internet use (−0.40) and more years of schooling (−0.20). More children go with higher infant mortality (+0.28) and with more religious attendance (+0.23). Women’s paid work and city living barely matter once income is allowed for.

Each measure against fertility today, comparing countries of similar income:

Internet use (197 countries)−0.40

Years of schooling (187 countries)−0.20

Women's political empowerment (168 countries)−0.12

People living in cities (199 countries)+0.04

Women in paid work (177 countries)+0.10

Religious attendance (105 countries)+0.23

Infant mortality (188 countries)+0.28

What went along with the fall

A snapshot shows how countries differ. To see what went with falling births, we compare change with change: the average for 1995–1999 against the average for 2018–2022, in 217 countries. Countries that started with more children had more room to fall, and poorer ones had more catching up to do, so each link allows for where a country started on both.

Fertility fell further where internet use spread more (−0.29), where more people moved to cities (−0.22), where more women took paid work (−0.22), where women gained more political power (−0.19) and where schooling grew more (−0.16). Falling infant deaths, often named as the first step, show almost no link once the rest is allowed for (+0.09).

Internet use (175 countries)−0.29

People living in cities (190 countries)−0.22

Women in paid work (174 countries)−0.22

Women's political empowerment (162 countries)−0.19

Years of schooling (153 countries)−0.16

Income (GDP per person) (190 countries)−0.03

Infant mortality (184 countries)+0.09

Religious attendance (46 countries)+0.32

No single change stands out. Each link is modest, and they all tend to happen together: as a country gets richer, its people move to cities, stay in school longer, go online and see women take on more paid and public roles. This data cannot pull those apart. Internet use, for one, is more likely a sign of when a country changed than a reason on its own. Religious attendance is the one measure that held up the other way: where attendance fell less, fertility fell less (+0.32), but that rests on only 46 countries with surveys in both periods.

Where births went up

Fertility was higher around 2020 than in the late 1990s in 20 countries. 14 of them are former communist countries. 10 of those had fallen below 1.5 children per woman in the late 1990s, after their economies collapsed in the decade before:

Kazakhstan: 2.03 to 3.05
Bulgaria: 1.18 to 1.77
Czechia: 1.18 to 1.73
Romania: 1.31 to 1.77
Mongolia: 2.45 to 2.86
Latvia: 1.15 to 1.56
Slovenia: 1.25 to 1.60
Kyrgyz Republic: 2.76 to 3.08
Russian Federation: 1.24 to 1.50
Estonia: 1.33 to 1.59
Georgia: 1.73 to 1.98
Hungary: 1.40 to 1.59

Three readings fit this, and the data here cannot choose between them. Some may be recovery: births put off during hard times, had later. Some may be later motherhood settling down, since the rate drops while women delay and recovers once they stop delaying. And several of these countries, Russia and Hungary among them, brought in large payments for families with children. And rising from very low levels may simply be easier than rising from average ones.

Rich countries: does paying families help?

Governments worried about low birth rates often pay families more. Across 38 rich countries, family benefits spending has almost no link with fertility (−0.05). Within each country over the years, as spending rose or fell, fertility did not move with it either (−0.07 across 42 countries, and +0.02 comparing each year with the same year elsewhere). That does not show the money does nothing: governments may spend more because births are falling, which would hide any effect. But there is no sign of a large one.

What does line up among rich countries is timing. Where mothers are older at birth, there are fewer children (−0.48), and the same goes for later marriage (−0.33). Women’s paid work goes slightly with fewer children (−0.22). And religion, which goes with more children worldwide, shows little link among rich countries (−0.14).

The usual explanations, against this data

More children survive, so families have fewer. True of the long history, but over the last 25 years falling infant deaths show little link with falling fertility once other changes are allowed for.
Women’s schooling and work. Consistent with the data: more schooling and more paid work went with bigger falls.
Cities and the cost of children. Consistent: fertility fell more where people moved to cities.
Religion and changing values. Partly consistent: more religious attendance goes with more children worldwide, though not among rich countries.
Economic shocks. Consistent with the former communist countries, which fell sharply in the 1990s and partly recovered.
Family payments. No sign of a large link in rich countries, at the levels governments spend.
Contraception. Almost certainly important, but not measured on this site.

What this cannot tell you

It describes countries, not families. It does not show why any one person has fewer children.
Everything here moves together over time, so no one change can be picked out as the reason.
A link is not proof that one thing changes the other.

These articles describe places, not people, and show links, not causes. Every article on PlaceStats.