Methods and data sources
Birth Rate World runs a simplified cohort-component projection, the same family of method that statistics offices and the UN use, for 194 countries and territories at once. This page explains every moving part in plain language, and what the model leaves out.
The model in one paragraph
Every country is split into 21 age groups (0 to 4, 5 to 9, and so on up to 95 to 99, then 100 and over) and into community groups. Each simulated year, people in each group either survive or die according to a life table, a fifth of each five-year group moves up to the next one, babies are added at the bottom, and migrants move between countries. That is the cohort-component method: follow cohorts of people as they age, and add the components of change (births, deaths, migration). Because deaths depend on how old a population is, a country with very low fertility first ages and then shrinks quickly, without any extra assumption.
Births
Births each year are the number of women aged 15 to 49, times a standard age pattern of childbearing, times the country's fertility rate in children per woman (the total fertility rate, or TFR). Women are taken to be 49% of each age group. The age pattern spreads births as follows: 5% at ages 15 to 19, 17% at 20 to 24, 27% at 25 to 29, 27% at 30 to 34, 16% at 35 to 39, 6% at 40 to 44 and 2% at 45 to 49.
A TFR of about 2.1 is "replacement": each generation of parents is replaced by a generation of children of the same size, allowing for some deaths before adulthood. A TFR of 1.05 means each generation is about half the size of the one before it, and a TFR of 0.7 means about a third.
For most countries (about seven in ten) the starting TFR comes directly from UN estimates or official figures. Where no TFR figure is in the data, the model derives one from the country's crude birth rate (births per 1,000 people) and its age structure.
Deaths
Deaths come from a life table: the chance of surviving each year of age. The model uses a single family of mortality curves (high risk in infancy, low risk in childhood and early adulthood, then risk rising exponentially with age) and scales it until the curve gives exactly the country's life expectancy at birth. Life tables are pre-computed for life expectancies from 30 to 100 years in half-year steps. This is coarser than a real national life table, but it captures the main fact that matters: older populations have many more deaths per person.
The 1950 to 2025 warm-up
The size of each age group in 2025 matters a great deal, because a country with many people in their thirties will have more births than one with the same TFR but an older population. Detailed age structures for every country are not part of the current dataset, so the model builds one. It starts each country in 1950 with a stable age structure and runs it forward to 2025 with a rough fertility history for its region: high fertility in 1950 (for example about 3 children per woman in Europe, 5.6 in East Asia, 6.6 in Sub-Saharan Africa), a smooth decline centred on a region-specific decade, then a glide to today's value. Life expectancy rises in a straight line over the same period. Japan, Thailand, Russia, Central Asia and some European countries get their own timing because their fertility fell earlier or later than their region's average.
The result is an approximate 2025 age pyramid that has the right general shape: baby booms and busts in the right decades, and populations that are older where fertility fell long ago. The pyramid for the selected country is shown in the Country tab, with the 2025 outline for comparison.
Scenarios
Three scenario buttons sit in the Rates tab. Country pages report numbers for the first two, plus a third "closed borders" case.
- UN-style: fertility converges (the default). Each country's fertility drifts toward a long-run target of 1.8 children per woman. High fertility falls toward it with a half-life of 35 years (half the gap closes in 35 years). Very low fertility recovers toward it much more slowly, with a half-life of 150 years, because recoveries from very low fertility have so far been small. Life expectancy rises toward 86 years with a half-life of 100 years. Migration continues at today's rates. This mirrors the broad shape of official projections, which expect fertility to converge, but it is not the UN projection itself.
- Freeze today's rates. Fertility stays at its 2025 value forever in every country. Life expectancy still improves and migration continues as in the default. This scenario shows the pure arithmetic of today's fertility, and it gets extreme quickly. Read why freezing today's birth rates is not a forecast.
- No migration. Switches off the baseline migration flows while keeping the other settings.
- Today's rates, closed borders (country pages and video clips). Fertility and life expectancy are fixed at 2025 values and there is no migration of any kind. It isolates what a country's own birth and death rates imply.
All the underlying settings (target fertility, half-lives, life expectancy, migration scale) are adjustable with sliders, and you can edit any country's starting fertility, life expectancy and net migration in the Country tab.
Migration
Migration has three sources.
- Baseline flows. Each country's UN net migration rate (people per 1,000 per year, positive for net arrivals, negative for net departures) is applied to its population. Worldwide, departures and arrivals are balanced: if the sum of all departures is larger than the sum of all arrivals, departures are scaled down to match, and vice versa, so migration never creates or destroys people.
- Overcrowding spill. When a country exceeds its capacity (see below), 3% of the excess leaves each year for countries with room, weighted toward countries with more room and higher existing immigration.
- Your own filters. In the Migration tab you can add flows of a chosen size, from one place to another, for a chosen community and period.
Migrants are mostly young adults, following a fixed age profile that peaks at ages 20 to 29. They keep their community group and carry their home country's fertility with them, which then fades toward the host country's level with a half-life of 40 years by default. This reflects a common pattern in research on immigrant fertility, which tends to move toward that of the host society over time and across generations.
The capacity rule
To keep long runs from producing physically absurd densities, each country has a capacity: a maximum density (600 people per square kilometre by default) times its land area, or 1.3 times its 2025 population if that is larger. As a country approaches its capacity, any fertility above replacement is squeezed toward 2.1. Past capacity, fertility drops below replacement in proportion to the overshoot, and the overcrowding spill moves people elsewhere. This rule is a modelling device inspired by the Game of Life idea of overcrowding, not an empirical law. It only matters in long scenarios where some populations grow many times over.
Community groups
Within each country the population is divided into nine broad groups by region of origin: European and European descent, Sub-Saharan African and diaspora, Middle East, North Africa and Central Asia, South Asian, East Asian, Southeast Asian and Pacific, Latin American and Indigenous, and two Jewish groups (non-Haredi and Haredi). Most countries have a single home group; multi-ethnic countries are seeded with approximate shares from census-based estimates.
The groups exist so that migration and fertility differences can be seen on the map. Research consistently finds that fertility is driven far more by country, income, education and religiosity than by origin. For that reason a group's fertility normally stays close to the country it lives in: it starts at most a modest step away from the country average and fades toward it over time, and the country's total births always match its overall rate in 2025.
The model can also treat a group as "insular", meaning it keeps its own fertility wherever it lives. By default only the Haredi (ultra-Orthodox) Jewish group is insular, starting at 6.5 children per woman, which is close to the figure reported by the Israel Democracy Institute; in the UN-style scenario that rate drifts toward the target like every other. You can switch this on or off for any group in the Communities tab. These are descriptive settings for exploring demography, not statements about any community.
How the map is drawn
The world is a grid of 800 by 348 cells drawn from Natural Earth country shapes. Each cell belongs to one country. As a country shrinks, cells go dark in proportion to its loss versus 2025, starting from the edges of a noisy pattern so the shrinkage looks organic. Growing countries brighten. In the Communities layer, cells are coloured by community share, so blobs of colour grow, shrink and move as shares change. Other layers show change versus 2025, current natural change (births minus deaths), median age and density.
Data sources
- Population, birth rates, death rates, net migration, life expectancy and fertility: UN World Population Prospects 2024 estimates for 2024 and 2025, rounded (population.un.org/wpp). For China, Japan, South Korea, Taiwan and Hong Kong the starting values use recent official national figures instead, because these change quickly and are closely watched.
- Haredi fertility (6.5) and share of Israel's population (about 13.6%): Israel Democracy Institute, Statistical Report on Ultra-Orthodox Society in Israel (en.idi.org.il).
- Community shares in multi-ethnic countries: rough estimates based on national censuses.
- Map: Natural Earth country boundaries (naturalearthdata.com), via the world-atlas package, rasterised to the cell grid. Land areas come from the same shapes.
Small territories are merged into the country they belong to for the map. The full data and the build scripts are open source.
Limitations
- Scenarios, not forecasts. Rates in the real world change. Fertility in South Korea went from about six children per woman in 1960 to below one; no model can know what the next fifty years bring. The longer the horizon, the more the results reflect assumptions rather than evidence.
- Approximate age structures. The 2025 age pyramids come from the warm-up, not from census counts, so the timing of births and deaths in the first decades can differ from official projections.
- Simplified mortality. One curve shape for every country, scaled to life expectancy. No pandemics, wars or famines.
- One sex ratio, one fertility pattern. Women are 49% of every age group everywhere and the age pattern of childbearing is the same in every country.
- Migration is mechanical. Flows follow today's net rates, the capacity rule and your filters. Real migration responds to economies, policy and conflict.
- Rounded inputs. Rates are rounded, and some small countries use estimates.
- Five-year age groups stepped yearly. Moving a fifth of each group up every year is a standard approximation, but it smooths out sharp single-year booms.
For comparison, in the UN-style scenario the world population goes from about 8.1 billion in 2025 to a peak near 9.49 billion around 2098, somewhat below the UN's own 2024 medium projection, which peaks at about 10.3 billion in the mid-2080s. Treat every number on this site as an illustration of what given rates imply.
Found an error or have better data? Get in touch.