Global births per woman have fallen from about 4.9 in the 1950s to 2.25 today, and 55% of countries are already below the replacement level. The workforce-to-retiree shift this produces is not a future scenario in the OECD area — it is already measured in the data.
Fertility has fallen in all ten of the world's most populous countries since 1990, though from very different starting points.4 China's rate has fallen furthest in relative terms, down 60% to 1.0 births per woman in 2023 — among the lowest recorded anywhere.4 Nigeria and Ethiopia remain well above replacement, but even there the rate has fallen — by close to a third in Nigeria and by nearly 45% in Ethiopia since 1990.4 India, once a byword for rapid population growth, has fallen even further in relative terms — down 50% to 2.0 in 2023, just below the replacement line.4
The global fertility rate has fallen by more than half since the 1950s, when it stood at around 4.9 children per woman.5 World Bank figures put it at 4.7 in 1960 and 2.2 by 2023;6 the UN's own 2024 estimate for the following year is 2.25.1 The two agencies use slightly different methods and vintages, which is why the figures don't match to the decimal — both agree on the direction and the broad scale of the fall. The UN's medium-scenario projection, built from historical trends rather than new measurement, has the world reaching the 2.1 replacement line by the late 2040s.1
In the OECD area, the number of people aged 65 or older per 100 people of working age (20–64) rose from 19 in 1980 to 31 in 20233 — a measured increase, not a forecast. Between 1960 and 2022, the 65-plus population grew at an annualised 2.2% while the working-age population grew at just 0.9%, and the OECD's over-65 share roughly doubled to around 18%.7 On the OECD's own projection, the old-age dependency ratio keeps climbing to 52 by 2060 if nothing else changes3 — a scenario, shown dashed below, not a new measurement.
Everything past 2024 in this piece is a scenario, not a measurement. The UN's own medium, high and low variants diverge by roughly half a child per woman globally by the mid-2030s, and its "peaked population" countries are modelled to see fertility recover to around 1.4 by 2100 — an assumption based on historical rebounds in 39 countries, not a certainty.1 How much of the OECD's projected dependency-ratio rise is offset by longer working lives, higher immigration, or automation is a policy question this data cannot answer on its own. And South Korea's fertility rate rose in 2024 for the first time in nine years9 — too recent and too small a move to say whether it marks a genuine turn or a one-year blip tied to a post-pandemic marriage rebound.
Dozens of governments now run cash, leave and childcare programmes aimed at fertility.10 Researchers disagree on how much these move the underlying rate, as opposed to just shifting when people who were going to have children anyway have them.
A 2025 systematic review covering high-income countries found cash benefits, childcare and tax refunds were the policies most consistently associated with higher birth rates, and that combining several policies at once had the largest effect.11 Economist Lyman Stone argues incentives demonstrably work, but require budgets far larger than most governments have committed — his estimate for the U.S. runs to roughly $5,300 per child per year to meaningfully move the national rate.12
Economists Melissa Kearney and Phillip Levine conclude that research on pro-natalist policy is mixed and generally does not find sizable fertility effects: Norway's expanded maternity leave had no discernible effect on birth totals, and cash schemes in Sweden, Spain and Alaska shifted the timing of births rather than raising the total a woman ultimately has.13