How many babies are born in the world?
About 132.1 million babies were born worldwide in 2023: roughly 361,700 a day, and 4.2 every second. The map above shows where, for every year from 0 CE to 2100. Below it, the questions people ask about those numbers, answered from the same dataset the map draws.
How many babies are born every day?
About 361,700 babies are born somewhere in the world every day: roughly 15,070 an hour, 251 a minute, and 4.2 every second. That is one new person about every 0.2 seconds.
None of that is a count. No global feed of birth registrations exists: many countries report births once a year, and some do not register most of them at all. The daily figure is the UN's estimate of 132.1 million births in 2023 divided by 365.2425, and the per-second counter on the map is the same division with a clock attached. Real births are not spread evenly anyway: they are mildly seasonal, and where deliveries are scheduled, weekday mornings are busier than weekend nights. The average is also falling: the world's busiest year for births was 2012, at about 146 million.
Which country has the most babies?
India, by a wide margin: about 23.22 million births in 2023, or 17.6% of every baby born on Earth. China is second with 8.91 million and Nigeria third with 7.51 million. Together the ten largest account for 52% of the world's births.
Population size only half explains the ranking. India and China have populations within a few per cent of each other, yet India records 2.6 times as many births, because fertility in China is far lower. The order is not fixed either: on the UN's medium projection, India still leads in 2100, but DR Congo has risen to second, and China, second today, has fallen to 8th.
What is the fertility rate?
The total fertility rate is the number of children a woman would have if she lived through her whole childbearing life at one year's age-specific birth rates. It describes a year, not a person: it is assembled from the rates observed at each age during that year, so it can be compared between countries without their age structures getting in the way.
It is not the crude birth rate, which counts births against everybody in the population. Colombia and France show the difference. Both are below replacement, at 1.65 and 1.64 children per woman, yet Colombia records 13.5 births per 1,000 people against France's 9.6, because far more of its population is of childbearing age. Replacement level is about 2.1: the point at which a generation just replaces itself.
Why are fertility rates dropping?
The fall is the clearest pattern in the whole UN record. Taiwan went from 6.48 children per woman in 1950 to 0.87 today, and 129 of 236 countries and areas are now below the replacement level of about 2.1, holding 67% of the world's population between them.
There is no single cause, and the mixture is different in each country. Demographic research points to people having children later, to longer education and paid work, to reliable contraception making family size a decision rather than an outcome, to urbanisation (children cost more in a city and contribute less), and to the price of housing and childcare. Preferences have changed too: a small family is now the norm across most of the world. Those are explanations from the literature, not quantities this dataset measures. What the data here can show is the fall itself, country by country and year by year.
Which countries have the lowest and highest fertility rates?
The range is wider than at any earlier point in the record. At the bottom sits Macao at 0.66 children per woman, with Hong Kong and South Korea just above it. At the top, Somalia at 6.13, followed by Chad and Niger, more than 9.3 times as many children per woman as at the other end of the table.
The two lists barely overlap geographically. Almost every country at the bottom is in East Asia or Europe; almost every one at the top is in sub-Saharan Africa. The consequence shows up in where babies are born rather than where people live: the 129 countries below replacement hold 67% of the world's people but account for only 45% of its births.
Birth rates around the world
The crude birth rate is the number of live births in a year per 1,000 people. Worldwide it was 16.3 in 2023, down from 36.8 in 1950. Between countries it runs from 46.4 per 1,000 in Central African Republic to 4.2 in Holy See.
Birth rate and fertility rate are not interchangeable, and treating them as one number is the most common mistake made with these statistics. The birth rate divides by everyone, so it rises and falls with a country's age structure as well as with its childbearing. A country whose large generation of young adults is just reaching thirty will post a high birth rate on modest fertility, and the same country will post a low one thirty years later on exactly the same fertility. Use the birth rate to see how fast a population is being added to; use the fertility rate to see how many children people are having.
How global births are changing
For eighteen centuries the number barely moved. Around the year 0 roughly ~9.3 million babies were born annually; by 1800 it was 40 million, a fourfold rise across the whole period. Then it went almost vertical, reaching 91.79 million by 1950 and peaking at 146 million in 2012. In 2023 it was 132.1 million, already below that peak.
On the UN's medium scenario annual births stay near 132.5 million in 2050 and fall to about 110 million by 2100. Those are projections built on assumptions about fertility, mortality and migration, not forecasts with a guarantee, and their uncertainty grows the further out they run. Looking the other way, the numbers get softer too: figures before 1950 are historical reconstructions, and those before 1800 are modelled rather than recorded, carrying an uncertainty of roughly a third.
Will the world population start shrinking?
Falling births do not shrink a population on their own. Four things decide the direction: births, deaths, migration and the age structure that is already in place. A country whose fertility fell twenty years ago still has large numbers of people entering their thirties, so it keeps producing births for decades afterwards. Demographers call that population momentum, and it is why the turning points arrive in sequence rather than together.
On the UN's medium scenario, world population peaks at about 10.29 billion in 2084 and is slightly lower again by 2100, while annual births have been falling since 2012. The gap between those two dates is the momentum, measured. This dataset holds population and births, not deaths or migration, so it can show the first of the four forces and the net result of all of them, but not each one separately. It would be wrong to read a birth series as if it were a population forecast.
Explore births by country
All 236 countries and areas have their own page: annual births from 0 CE to 2100, the crude birth rate, children per woman, population, and a label on every value saying whether it is a UN estimate, a UN projection, a historical reconstruction or a modelled estimate.
- India
- China
- Nigeria
- Pakistan
- Indonesia
- DR Congo
- Ethiopia
- United States
- Bangladesh
- Brazil
- Germany
- Japan
- France
- United Kingdom
- Spain
- Italy
- South Korea
- Mexico
Figures are for the land inside present-day borders in every year, including years in which those borders did not exist. "France in the year 800" means the area that is France today.
Where does the data come from?
Births are estimated as population × crude birth rate ÷ 1,000. From 1950 both inputs are the United Nations World Population Prospects 2024, whose estimates end in 2023; 2024–2100 is its medium scenario. For 1800–1949 population comes from Gapminder and the birth rate from Gapminder's own reconstruction. Before 1800 no source records births anywhere in the world, so population comes from HYDE's gridded reconstruction and the birth rate is modelled from mortality and the measured growth of the population series itself.
Every value on this site is tagged with which of those four it is, and printed only to the precision its uncertainty supports: a modelled estimate reads "~9.3 million", never 9,304,187. The pre-1800 model is tested where it can be, against Gapminder's independent reconstruction at 1800, and the size of its error is published rather than corrected away.