Who arrives, who leaves

Everything else on this page is a balance. This is the one count: in 2024, 303,397,238 people were living in a country other than the one they were born in, 37 of every 1,000 people alive.

How to read this: the world in 2024 shrunk to 1,000 people, one square each. The 37 filled squares are the people living outside the country they were born in; everyone else lives in the country of their birth.

Almost twice as many as in 1990, but the world grew too

1. 1990153.65 million29 of 1,000
2. 1995162.85 million28 of 1,000
3. 2000174.17 million28 of 1,000
4. 2005192.34 million29 of 1,000
5. 2010220.53 million31 of 1,000
6. 2015249.52 million33 of 1,000
7. 2020274.71 million35 of 1,000
8. 2024303.40 million37 of 1,000

How to read this: one row for each year the UN made an estimate, one square per 20 million people. The UN counts every five years, and once more in 2024; nothing between those years is drawn.

The number of people born elsewhere has almost doubled since 1990, from 153.65 million to 303.40 million. Out of every 1,000 people it rose more slowly, from 29 to 37, because the number of people living where they were born grew too. Put the other way round: about 96 of every 100 people live in the country they were born in.

Where they live

The United States is home to more people born elsewhere than any other country: 52.38 million, about 17 of every 100 of them in the world. But they are 15 of every 100 people living there. Where people born elsewhere are the largest share of everyone, the list is the Gulf and small places: Qatar (77 of every 100) and the United Arab Emirates (74), where most of the workers came from somewhere else, the same shape of work as in the pictures below. In 17 countries and areas, at least half of the people were born in another one.

Many of them send money to the families they left: money sent home is on the economy page.

Migration happens at an age. In 2020–2025 more people arrive in the United Arab Emirates aged 25–29 than at any other age, 121,000 a year, while from 45 on the same country loses people every year. This is the part of population almost nobody draws.

more arrive than leave   more leave than arrive

How to read this: left to right is age, from babies to 100. Blue squares above the line are ages at which more people arrive than leave; orange squares below it are ages at which more leave than arrive. Each country has its own square size, printed beside its name, because the numbers differ enormously.

Two shapes. The United Arab Emirates shows the shape of work: people arrive in their twenties and thirties, almost no children arrive with them, and from the forties on more leave than come. The United States shows the shape of settling: people arrive at every age, children included, most of all at 20–24 (196,000 a year). The first country is borrowing working years; the second is gaining families. Both are called immigration, and they are not the same thing.

The ages also explain what migration does to the country people leave. The ages that move are the ages that have children, so a country losing people at 20–34 loses their children too: births that never happen there. It is one reason a generation on the wall and the births beneath it rarely agree.

The pictures above start in 2020, because that is where the UN's own table starts: it says plainly that "the package does not include historical migration by age". For the 73 years before it there is one number per country per year, and the largest of them all is Ukraine, which lost 5.70 million people in 2022.

more arrive than leave   more leave than arrive

The largest single years a country lost people

1. Ukraine 20225,699,445
2. Pakistan 20162,290,411
3. Syria 20132,081,303
4. Rwanda 19942,078,111
5. Afghanistan 19811,998,086
6. Iran 19931,696,091

And gained them

1. United States 20161,889,129
2. Iran 19811,452,169
3. Rwanda 19961,367,700
4. Russia 20021,358,944
5. China 19671,246,654
6. Democratic Republic of the Congo 19941,232,105

How to read this: one row per country and year, one square per 200,000 people. Orange means more left than arrived, blue means more arrived than left. The number on the right is the balance itself: everyone who arrived that year minus everyone who left.

The two lists are not alike. Losses come as single catastrophic years: a war, a partition, a famine. Gains come steadily. The United States gained between 1.59 million and 1.89 million people in every one of the ten years 2010–2019, which over a decade is far more people than any single year on the other list. Leaving is an event; arriving is a decade.

• No ages before 2020. The UN's table by age is a projection only. Nothing here can say how old the people leaving Ireland in 1955 or Viet Nam in 1979 were.
• These are balances, not people, all but one. Net migration cannot say how many arrived, only how many more arrived than left, and a place with large movement both ways looks empty here. The one count, people born somewhere else, says who lives where, not when they came or how often they moved.
• "Born somewhere else" moves when borders move. When a country splits (the Soviet Union, Yugoslavia, Czechoslovakia) people who had moved from one part of it to another become born in another country without moving at all. And some countries have no count of people born abroad, so the UN estimates from the number of foreigners instead. The World Bank, which publishes these figures, gives both warnings itself.
• No world figure, ever. The world's net migration is zero by construction. The UN's own world row is 0 in every year, and its table by age comes to about 3,931 people a year for the whole world, which is the model's rounding and not a finding about anybody.
• Men and women are never drawn apart. The table splits by sex, but its male and female world rows each miss zero by 165,706 over 2020–2025, in opposite directions. Drawing that split would be drawing the model's own error, so every figure here adds the two together.
• Refugees are not counted apart. Someone who flees and someone who takes a job abroad are the same line in this data. Who is a refugee, where they went and whether they went home is a different body of records: who is forced from home.
• After 2023 it is a projection, the UN's medium variant, and migration is its least predictable part: the model assumes recent patterns hold, so no war and no law after 2023 is in these numbers.

How to read this: people a year, arrivals minus departures, the biggest gainers first. A minus sign means more people leave than arrive. The last column is the single largest year in the record, either way. The last two columns are the one count: people living there in 2024 who were born in another country, and how many of every 100 people living there that is. A dash means the UN has no estimate.

All four tables are the United Nations' own. The one by age is published only inside a workbook: its periods are five years long and in whole people, so everything here is divided by five to give people a year, and its male and female rows are added together first.

Checked before it was drawn

Verified 2026-09-30 · 73 canaries The table by age is checked to hold exactly 9,504 rows — each of the UN's 297 locations, in 16 periods, for two sexes — and published values are reproduced from it before this page is built. The yearly figures are checked to cancel out across every country in every year: that is the claim the first paragraph makes, and the build fails if it stops being true.

Born somewhere else

The number of people living outside the country they were born in is the UN Population Division's International Migrant Stock, published by the World Bank with its licence record, which is checked before it is used. The shares are this site's: that number divided by the UN's own count of people in the same country and year, and checked against the Bank's own share (they agree to within 0.05 of a point). No estimate for: Taiwan, Mayotte, Cook Islands, Falkland Islands, French Guiana, Guadeloupe, Kosovo, Martinique, Montserrat, Bonaire, Sint Eustatius and Saba, Niue, Réunion, Saint Barthélemy, Saint Helena, Anguilla, Saint Martin, Saint Pierre and Miquelon, Western Sahara, Tokelau, Guernsey, Jersey, Wallis and Futuna.

Sources: UN WPP 2024 · World Bank. Every source, its licence and how to cite it →