How have births changed through history?

flat, then vertical, then down Annual births went from about 9.29M in 1 CE to 39.7M by 1800 — a fourfold rise across eighteen centuries. They then rose 3-fold in the two centuries after, peaked in 2012, and have been falling since.
10M100M 1 CE1000150018002026
Estimated births worldwide each year, 0 CE to 2026, on a logarithmic scale — a linear one would show the first eighteen centuries as a flat line on the floor. The shaded band marks modelled years. The marker sits on 2024.

Four phases

World births and the crude birth rate per 1,000 people, by phase.
YearsWhat happenedAnnual birthsPer 1,000
0 – 1800 Eighteen centuries of almost nothing 9.29M → 39.7M 40 → 41
1800 – 1950 The rise 39.7M → 91.8M 41 → 37
1950 – 2012 The peak 91.8M → 146M 37 → 20
2012 – 2026 The turn 146M → 132M 20 → 16

0 – 1800 — Eighteen centuries of almost nothing

Births roughly quadrupled in 1,800 years — slower than the world population now grows in a decade. The birth rate barely moved, because it could not: it was pinned near the death rate.

1800 – 1950 — The rise

Deaths fell first, births stayed high, and the gap between them became the population explosion. More babies were born because more of the babies already born survived to have their own.

1950 – 2012 — The peak

The absolute number kept climbing even as the rate collapsed, because the number of people of childbearing age was growing faster than fertility was falling.

2012 – 2026 — The turn

Fertility fell far enough, in enough of the world, that the absolute number of births began to fall too. World population is still rising; the number of babies is not.

The counter-intuitive part

The birth rate and the number of births move together for most of history and then come apart. Between 1950 and 2012 the rate fell from 37 per 1,000 to about 20, while the number of babies rose by roughly 59%. A shrinking share of a much larger population is still more people.

That is also why the current decline matters more than it looks. Births are now falling in absolute terms, from a population that is still growing — which means the rate is falling fast enough to overcome the momentum that carried the numbers up for sixty years.

Everything before 1800 on this chart is modelled and everything before 1950 is reconstructed. The shape is well established; the exact heights are not. An uncertainty of about a third applies to the ancient end — see the methodology, which also reports how far the pre-1800 model lands from an independent reconstruction where the two can be compared.

Play the whole timeline

Related questions

Figures on this page are generated from the same dataset the map draws, and update whenever it does. See the methodology for how each era is estimated and how much it should be trusted.