// Summary · May 2026

The Global Fertility DeclineSummary

From above-replacement to sub-replacement, trends, drivers, and what the data does and does not show. Key findings from the full 10,500-word report.

Total fertility rate (TFR) is the average number of children a woman would have at current age-specific birth rates. Replacement is 2.1. The global rate has more than halved since the 1960s.

In 1970 the global TFR was 4.8. By 2024 it was 2.25. The decline unfolded in three waves: the Western inflection of the 1960s-70s (contraception, women entering the workforce), the East Asian transition of 1975-2000 (Japan, Korea, Taiwan, China industrialising rapidly), and the post-2015 ultra-low emergence, during which Korea fell from 1.24 to 0.72 and simultaneous record lows appeared across France, Australia, the United States, Japan, and China.

The world divides into three fertility bands. Sub-Saharan Africa sits above 4.0 (Niger: 6.6, Nigeria: 4.8), with the trajectory downward. The sub-replacement cluster (1.5-2.1) includes most of the Anglosphere: the United States (1.58), Australia (1.48), the United Kingdom (1.55), France (1.56). The ultra-low band (below 1.5) is dominated by East Asia and Southern Europe: Japan (1.15), Italy (1.20), Spain (1.16), China (~1.0), South Korea (0.80).

South Korea's 0.80 is the lowest recorded TFR of any country in history. Several countries including France, Australia, the United States, and Germany are below replacement but growing via net migration, a buffer that is finite and whose political sustainability is contested.

Two distinctions matter for reading fertility data correctly: tempo versus quantum, and period versus cohort.

Tempo is the timing of births within a woman's reproductive life. Quantum is how many children she ultimately has. Period TFR conflates the two: if women are merely delaying births, the headline rate falls even if lifetime fertility is unchanged. Cohort fertility (completed lifetime births) is the truer measure but only knowable once women reach 49.

In Australia and France, the gap is largely tempo. Australia's period TFR was 1.48 in 2024, but completed cohort fertility for women born in 1975 was 2.03, close to replacement. Women are having children later, not necessarily fewer. In Korea and Japan, the decline is quantum. Cohort fertility there is approaching or below 1.5 and trending lower, with no recovery in sight. Korea's collapse is driven heavily by rising childlessness, not by mothers having fewer children: more women are not becoming mothers at all.

No single driver explains the global trend. The evidence is strongest for economic and educational factors, real but harder to measure for cultural factors, and weakest for biological factors.

Female educational attainment is the most robust correlation in the cross-country fertility literature. More years of schooling, particularly tertiary education, reliably reduce fertility across regions and time periods. The mechanism is straightforward: education raises the opportunity cost of childbearing and raises the age at first birth.

Housing costs have emerged as the second strongest driver. Couillard (2025) attributes roughly 51% of the US fertility decline between the 2000s and 2010s to rising housing costs. A 10% rise in house prices is associated with a 1.3% drop in the birth rate in Australia and a 0.02 fall in TFR in Korea. The mechanism operates directly (children require space) and indirectly (high costs push back marriage and first birth).

Childcare costs, economic insecurity, and secularisation contribute further. Countries with publicly funded childcare (France, Nordic countries) have held fertility higher than peers. Religious adherence correlates strongly with higher fertility within every country it has been measured. Israel (TFR 3.0) is the most striking case. The smartphone hypothesis is the most amplified recent explanation; the underlying evidence suggests a real effect on teen fertility (largely a public health success) but no proven effect on adult fertility, which accounts for 80% of reproductive-age women.

Six contested questions, with the evidence weight for each.

Economic stagnation from workforce decline — The dependency ratio concern is real. Whether it constitutes catastrophe depends on productivity assumptions. Honest answer: serious adjustment challenge, magnitude uncertain.

Recovery — No advanced economy that has fallen below 1.5 has durably returned to sub-replacement and held it. Brief partial recoveries have occurred; sustained recoveries have not. Evidence leans toward concern.

Policy effectiveness — Evidence leans toward pessimism. Policy may shift timing; it has not demonstrably moved completed lifetime fertility. Korea's $270B failure is the canonical case.

Constraint versus preference — The desire-reality gap (French women want 2.3 children, have 1.56) supports the constraint reading. Expanded female opportunity supports the preference reading. Both are real and not mutually exclusive.

Replacement migration — Currently working in receiving countries. Source countries are entering their own demographic transitions. The pool is finite and shrinking.

Smartphones — Plausible, amplified, unproven for adult fertility. The careful read: worth taking seriously, not yet established.

The 1.5 threshold is the key analytical boundary. Above it, populations shrink modestly and can be stabilised through migration. Below it, shrinkage accelerates and recovery becomes empirically rare. Below 1.0, each generation is more than halved. No country that has entered the ultra-low band has durably recovered above it.

Israel (TFR 3.0) is the OECD outlier and the only developed-world counter-example. Ultra-Orthodox fertility is around 6.4; secular Israeli Jewish fertility is still above replacement at 2.7. The Israel case proves that high fertility in a developed economy is not impossible. The combination of cultural, religious, and demographic-political factors does not transplant easily to other contexts.

The honest reading: policy can affect the timing of births and may produce modest braking effects in some contexts. It has not been shown to durably reverse a structural fertility decline in any maximalist case.

Three major institutions disagree by amounts that translate into over a billion people by 2100. The UN WPP 2024 (medium variant) projects a peak of 10.3 billion in 2084. The IIASA/Wittgenstein Centre (SSP2) projects a peak of 9.7 billion in the 2070s. The IHME/Lancet projection is the most pessimistic: peak at 9.7 billion in 2064, declining to 8.8 billion by 2100, with 97% of countries below replacement.

The core disagreement is the assumed TFR floor. The UN model assumes countries partially recover once they reach very low TFRs. IHME assumes much less recovery. The UN's own projections have revised downward in each major release: the 2019 peak estimate was 10.9 billion; by 2024 it was 10.3 billion. The reasonable summary: 2100 global population is somewhere between 8.5 and 10.5 billion.

Three things the data shows with high confidence. Three things it does not settle.

What the data shows. Sub-replacement fertility is the global majority condition, not a Western or East Asian anomaly. No advanced economy that has fallen into ultra-low fertility has durably recovered. Maximalist pro-natalist policy has not reversed the trend in any of the countries that have attempted it.

What the data does not settle. Whether 8.8 or 10.3 billion is the better 2100 estimate. Whether the consequences of population decline are net negative or net positive, given that economic, environmental, and social effects pull in different directions. Whether the decline reflects worsening constraints or expanding choices, and the policy implications of that answer differ enormously.