ImmigrationDBVerified immigration data

The cities gaining migrants are not the cities that are growing

By · Published · Updated

Question: where are foreign populations actually growing, and does that growth follow the cities that are growing overall?

Answer: growth is widespread, and it is almost unrelated to whether a city is growing at all. Across 423 cities with at least a decade of official figures, 394 saw their foreign population rise and 29 saw it fall. But the correlation between foreign-population growth and total-population growth is only 0.11 — and inside every country large enough to measure separately, it is near zero or slightly negative: −0.01 in Germany, −0.05 in France, −0.08 in Italy, −0.16 in Spain.

That is the finding. Shrinking cities are not being skipped. In Spain the relationship leans the other way: cities losing residents overall have been gaining foreign residents at least as fast as the ones that are growing.

What we measured

One row per city, for every city that publishes a usable run of foreign-resident counts: at least four observations spanning at least three years, from one publisher, in one unit. The median span is 10 years. After the validation described in the method note below, 423 cities across 18 countries qualify.

Foreign population is a stock, not a flow. It counts people recorded as foreign residents on one date. It is not arrivals, not immigration, not net migration. Somebody who naturalises leaves the count without leaving the city, and somebody born locally to foreign parents may enter or never enter it depending on the country. A rising number is not "more arrivals" and a falling number is not "people leaving" — this article never uses either phrase for a change in the stock.

The definition also differs by country. Most of these series count residents without the country's citizenship; some national systems count the foreign-born, and others count a broader migration background. That makes the level incomparable between countries and the direction over time comparable only where one publisher used one definition throughout. So every comparison here is a city against its own past, or cities within one country against each other.

Growth is expressed two ways: total percentage change between the first and last observation, and the compound annual rate, which makes an eleven-year run comparable with a nine-year one. Total population change over the same window comes from the same city record.

What the data shows

Foreign populations rose almost everywhere: 394 of 423 cities. The median city added foreign residents at 2.8% a year in Spain, 1.5% in Italy, 4.0% in Germany and 4.3% in France. Every one of the 120 German cities and all 46 French cities in the sample grew.

Vilnius leads at 15.6% a year, from 7,311 foreign residents in 2011 to 48,333 in 2024 — a near-sevenfold rise in a city whose total population grew 11% over the same period. That gap between the two figures is the article in one line.

Behind it, the ranking is dominated by two countries. Portuguese cities — Braga at 13.1% a year, Aveiro at 11.5%, Viana do Castelo at 14.0% — and Bulgarian ones, Burgas at 13.9% and Varna at 12.3%. These are compound rates from small starting counts: Haskovo's 13.4% a year is 191 residents becoming 763.

Two German cities in the top twelve make the independence point sharply. Wilhelmshaven grew its foreign population 13.1% a year while its total population fell 1.5%. Gera grew 11.1% a year against a total population change of −0.1%. Neither city is growing. Both are destinations.

Where it breaks down

The pooled correlation of 0.11 is weak, and splitting it by country removes even that.

Germany is −0.01 across 120 cities, France −0.05 across 46, Italy −0.08 across 80, Spain −0.16 across 90. Only these four clear the article's floor of 20 cities. The pooled 0.11 is therefore not a European pattern that the countries share — it is what appears when four national systems with different growth levels are stacked in one chart.

The practical reading: the total-population line on a city page tells you almost nothing about whether that city is attracting foreign residents. They are separate questions, and this dataset says they should be asked separately.

The 29 declining cities are concentrated and they need care. The steepest are Baltic: Liepāja −68%, Daugavpils −62% and Rīga −61% between 2000 and 2025, and Narva −40% between 2011 and 2026. These figures include the large post-Soviet non-citizen populations of Latvia and Estonia, which shrink through naturalisation and ageing as well as departure. A resident who takes citizenship leaves this count while staying exactly where they are, so the decline cannot be read as people moving out. The Baltic series are also the longest in the sample — 25 years for the Latvian cities against a median of 10 — so they cover a different era from the rest.

Elsewhere the declines are milder and more ordinary: Norrköping −24%, Nitra −23%, La Louvière −19%, Košice −16%.

What this means if you are moving

Do not use a city's population trend as a proxy for whether it receives people like you. On this evidence the two move independently, and some of the most receptive cities in the sample are places with flat or falling total populations.

Read a fast growth rate together with its starting number. A city going from 191 to 763 foreign residents is growing at 13% a year and still has a very small foreign community; a city adding 40,000 at 15% a year is a different proposition. Both appear near the top of the same ranking.

Compare within one country. Because "foreign" is defined nationally, the honest comparison is Braga against Aveiro, or Wilhelmshaven against other German cities — not Braga against Wilhelmshaven.

And treat a falling count carefully before reading it as decline. In the Baltic cities it substantially reflects people becoming citizens of the country they already live in.

Limits

This measures a recorded stock of foreign residents, not migration. No arrival, departure or net-migration figure appears anywhere in it, and no statement here should be read as one.

Definitions differ between countries and the levels are not comparable. Even within a country a series can change basis; the ones that visibly did were removed, but a subtler revision could survive undetected.

Growth rates from small bases are volatile. Several of the fastest-growing cities started under 2,000 foreign residents, where a few hundred people move the percentage a long way.

The windows differ. The median span is 10 years, but the Latvian and Estonian series run 15 to 25 years and cover a different period from the rest, so their rates are not measuring the same decade.

Nothing here identifies a cause. The dataset contains no policy, employer, university, housing or transport variable, so this article does not explain why any city grew — only that it did, and that its total population moved separately.

Method

The sample begins with 517 cities holding a usable foreign-population trend and ends with 423, and the 94 exclusions are themselves a result worth stating.

The submitted ranking was led by five Dutch cities at 24–26% a year, which over eleven years implies roughly twelvefold growth. Checking the underlying series showed why. Every Dutch city's run ends on a value from a national statistical office file after a decade of Eurostat citizenship counts, and the jump at that seam is roughly fourfold — Assen goes from 3,166 in 2024 to 14,266 in 2025, Almere from 54,096 to 114,076. That is a change of definition, from residents without citizenship to a broader migration-background count, not a change in who lives there. Separately, 39 of the 42 Dutch cities carry a 2023 observation three to five times both its neighbours — Amsterdam records 201,158 in 2022, 611,531 in 2023 and 236,146 in 2024 — which is a broken extraction for that year, not a population event.

Every series was therefore tested for two faults: a change of publisher at the final observation accompanied by a jump of more than 50% up or down, and any single year more than twice both of its neighbours. 94 cities failed and were removed — all 42 Dutch cities, 21 French, and small numbers in Spain, Italy, Portugal, Switzerland, Germany, Finland, Norway, Belgium and Sweden. Reporting their growth would have published a definitional artefact as a migration finding.

On the 423 that survive, Spearman's rank correlation is the primary coefficient, computed between total foreign-population change and total population change over each city's own window. Every city is weighted equally; population is not used as a weight. A country receives its own coefficient only with at least 20 qualifying cities, which leaves Germany (120), Spain (90), Italy (80) and France (46).

Sources

Frequently asked questions

Where is the foreign population growing fastest?

Among 423 cities with at least a decade of official figures, Vilnius leads at 15.6% a year, from 7,311 foreign residents in 2011 to 48,333 in 2024. Portuguese and Bulgarian cities fill most of the rest of the top ten. These are compound annual rates for a resident stock, not counts of arrivals.

Does a growing foreign population mean the city is growing?

Almost not at all. Pooled across 423 cities the Spearman correlation between foreign-population growth and total-population growth is 0.11, and inside the four countries large enough to report it is near zero or slightly negative: −0.01 in Germany, −0.05 in France, −0.08 in Italy and −0.16 in Spain. A city can be shrinking overall and still be a destination.

Why did foreign population fall in Rīga and Narva?

Foreign population is a stock, not a flow. In Latvia and Estonia it includes a large post-Soviet non-citizen population that shrinks through naturalisation, ageing and emigration. Rīga falls 61% between 2000 and 2025 and Narva 40% between 2011 and 2026 without anyone necessarily leaving on the naturalisation part.

Can these numbers be compared between countries?

No. Each statistical office defines "foreign" its own way — non-citizen in one country, foreign-born or migration background in another. Compare a city with its own past, and cities within one country with each other. This article never ranks countries against each other by level.

Why were 94 cities excluded?

Their published series break mid-way. All 42 Dutch cities switch from a Eurostat citizenship count to a national migration-background count, which multiplies the last value roughly fourfold; 39 of them also carry a corrupt 2023 point three to five times its neighbours. Measuring growth across that break would have reported a definition change as migration.

More research

More foreign residents, more crime? The answer changes at every border

Does a larger foreign population mean more recorded crime? It depends on the country: 0.06 in Germany and the US, 0.50 in Spain and Belgium, 0.59 in the Netherlands.

Clean air barely shows up in the rent

Do you pay more rent for cleaner air? Across 392 German housing markets the link is weak — Spearman −0.12 — and 31 districts sit below the median on both.