Every number here comes from Eurostat and links back to the dataset it came from. Where a figure is uncertain, contested, or simply not measured, that is stated next to it rather than hidden.
Shade the map by legal residence, new permits, asylum, or enforcement activity. Each measure states what it counts and, just as importantly, what it does not.
Monthly, 2014-01 to 2026-04.
Applications peaked at 162,050 in 2015-10. The most recent month on record, 2026-04, is 43,410 — a fall of 73%.
The series stops at the last month every reporting state has filed, rather than running on to the present with partial returns. Months not yet reported are left out, never plotted as zero — Eurostat records them as absent, and drawing them as nil would invent a fall that has not happened.
Monthly, end-of-month stock.
Arrivals falling does not mean the system has cleared. 1,183,425 people were waiting on a first decision in 2026-04, against a peak of 1,263,380 in 2025-03.
A backlog reflects decision-making capacity as much as arrivals. It rises whenever applications outrun decisions, and it excludes anyone waiting on an appeal.
First-time applicants, 2025-05 to 2026-04.
The composition is not the one most people picture. Applications say nothing about how somebody entered, and several of the largest origin groups arrive by air on a valid visa rather than crossing a border irregularly.
Share of first-instance decisions granting protection, latest four quarters, for nationalities with at least 2,000 decisions. The EU-wide average across all nationalities was 41.2% in 2026-Q1, a number that describes almost nobody.
Protection is assessed against conditions in the country somebody fled, so the spread is what the system is designed to produce. It does mean any single headline rate conceals a difference of an order of magnitude, and that first-instance figures understate the eventual total, because a substantial share of refusals are overturned on appeal.
Annual, EU-27.
In 2025, 491,950 people were ordered to leave the EU and 155,110 departures were confirmed — 31.5%.
Treat this as an approximation. A departure recorded this year may follow an order issued several years earlier, so the numerator and denominator are not the same group of people, and states differ in how they confirm that someone has actually gone.
In March 2022 the EU activated the Temporary Protection Directive for people displaced from Ukraine. It grants residence, work and access to services without an asylum claim — and because no claim is lodged, none of these people appear in the asylum statistics.
That makes them invisible in almost every chart on this site, and in most reporting about migration to Europe. It is the single largest omission in the conventional picture, and it runs in the direction that flatters nobody’s argument: the asylum numbers are genuinely falling, and the number of people under protection is far larger than those numbers suggest.
There are currently 4,384,025 beneficiaries — about 7.0× the 629,360 first-time asylum applications lodged across the EU in the last twelve reported months.
EU-27, 2022-03 to 2026-05.
Concentrated in central and eastern Europe. Poland and Czechia host far more people under protection than their asylum figures would ever suggest.
Overwhelmingly Ukrainian, which is what the directive was activated for.
It is tempting to sum temporary protection and asylum into one headline number for “people under protection”. This site does not, because they are different legal statuses measured on different bases — one is a standing stock counted monthly, the other a flow of applications. Adding a stock to a flow produces a figure that means nothing.
The honest treatment is to show both, at the same prominence, and to say plainly that neither one alone describes migration to Europe.
Source: Eurostat migr_asytpsm
This is the figure every tracker claims to show, and it is the one figure no statistical agency produces. People without authorisation are, by definition, absent from the registers that would count them.
What exists instead is a record of contact with authorities: people found to be present without permission, refused at a border, ordered to leave. Those are counts of administrative events. One person can generate several. Somebody never encountered generates none.
Researchers do produce estimates using indirect methods, with error bars wide enough to drive an argument through in either direction. Those belong in a section that shows the range and the method — not on a map as though they had been counted.
Elsewhere this site declines to show a count of Europe’s irregular population, because none is collected. Researchers do nevertheless try to estimate it. Here is the most credible such attempt — shown the way its authors published it, as a range, with its method and its age attached.
3,900,000 to 4,800,000 people, across the 28 EU member states of the time plus Iceland, Liechtenstein, Norway and Switzerland.
The hatched portion is 900,000 people — roughly 23% of the lower figure. Quoting either end as “the” number silently discards it.
Chiefly the residual method: subtract the number of non-EU/EFTA citizens known to be lawfully resident from the total estimated number present. What is left over is taken to be the unauthorised population — along with everyone the underlying surveys missed, miscounted or misclassified.
The gap between the low and high figures is roughly 900,000 people — about a fifth of the estimate itself. That spread is not a failure of the work; it is the honest width of what the method can support. Any single number quoted from this range has had its uncertainty quietly deleted.
This remains the most recent comprehensive estimate for Europe, and it was itself the first in about a decade. It predates the 2022 displacement from Ukraine and every subsequent shift in arrivals and enforcement. Presenting it as a current figure would be wrong, and no newer equivalent has replaced it.
Because the estimate is what remains after subtraction, every inaccuracy in the inputs lands in the result. Undercount the lawfully resident population and the irregular figure rises; undercount the total population and it falls. The residual cannot distinguish an unauthorised resident from a survey error.
Frontex publishes detections of irregular border-crossings, and states that these count crossings rather than individuals, because the same person may cross more than once. A person recorded three times is three detections. This matters most along the Western Balkans route, where the same border may be attempted repeatedly. No EU system tracks individuals across crossings, so detections can never be summed into a population.
The honest summary: the best available estimate of Europe’s unauthorised population is a nine-year-old range spanning nearly a million people, produced by subtraction rather than counting. That is genuinely useful for understanding scale, and genuinely useless for the precise per-country figures that trackers like to display. Both of those things are worth knowing.
This is the question people actually arrive with, so it is answered here rather than avoided. It is also the question most often answered with invented numbers, in both directions.
Two things are true at once, and most accounts drop one of them. Across the research literature, immigration is not associated with rising crime. And in several European countries there is a real, measured gap in recorded offending between people of foreign and native background. Understanding why both can hold requires knowing what the crime statistics actually count.
The most cited synthesis pools 51 studies published between 1994 and 2014 examining immigration and crime across geographic areas. The overall association is negative — more immigration, slightly less crime — but the authors are explicit that it is very weak, and that results vary considerably depending on how crime is measured, the size of the areas studied, and the period covered. The honest summary is that immigration does not drive crime up, not that it drives crime down.
Swedish studies have repeatedly found the risk of conviction among people of foreign background to be roughly twice that of people born in Sweden to two Swedish-born parents. This is a real finding in the data and stating otherwise would be false.
Where studies capture differences in resources and social position between groups rather than looking at origin alone, the measured overrepresentation is lower. How much of the gap this accounts for differs between studies, and the research does not support a claim that adjustment removes it entirely. This is the point where the evidence is genuinely unsettled, and where confident assertions in either direction outrun what has been shown.
Police statistics record nationality, and sometimes country of birth. They do not record immigration status. There is no European dataset in which offences are broken down by whether the person was lawfully resident. Figures circulating online that purport to count crimes by "illegal immigrants" are not drawn from any such source, because none exists.
Bundeskriminalamt (BKA) (2024); Office for National Statistics (2025)
These are not debating points. Each is acknowledged by the agencies that publish the underlying statistics.
Illegal entry and residence are themselves criminal offences, and they sit inside the same police statistics as theft and assault. In Germany these are recorded as ausländerrechtliche Verstöße, offences for which there is almost never a German suspect, because a German cannot commit them. Counting them makes foreign nationals appear more criminal purely as an artefact of immigration law. The BKA publishes its figures both with and without this category for exactly this reason; excluding it, the non-German share of suspects is 30.5%. Any comparison that does not exclude these offences is measuring immigration status, not crime.
Bundeskriminalamt (BKA) (2024); Mediendienst Integration (2025)
Police statistics record people suspected of an offence, at the point police record it. Investigations may be dropped, charges never brought, defendants acquitted. German police statistics are, in the words of police representatives themselves, a record of police work rather than a picture of crime. If any group is disproportionately suspected — through profiling, through reporting patterns, or through living in more heavily policed areas — that appears in the numbers as offending.
Everywhere in the world, in every population ever measured, young men commit most recorded crime. Migrant populations skew young and male far more than settled populations do. A comparison that does not adjust for age and sex is substantially measuring that difference in composition rather than any difference in behaviour.
A rate needs a population to divide by. Offender counts by nationality are often available while the resident population by nationality is not. In the United Kingdom the census has no nationality field at all — the closest proxy is which passport a person holds, recorded for a limited set of countries. Offender counts by nationality can therefore be published without it being possible to turn them into rates. Dividing by a guessed denominator produces whatever answer the guess implies.
Crime enters the statistics when someone reports it. Willingness to go to the police differs between native and migrant populations, in both directions and for different offence types. That difference alone moves recorded crime rates without any underlying difference in what happened.
Migrants are concentrated in cities and in poorer districts, which have higher recorded crime for reasons that long predate any recent arrival. Comparing a heavily urban population with a national average compares urban crime with rural crime as much as anything else.
Brottsförebyggande rådet (Brå) (2021); Bundeskriminalamt (BKA) (2024)
Police records hold nationality, and sometimes country of birth. They do not record whether a person was lawfully resident. That means no honest chart can show crime committed by irregular migrants in Europe, because the data to build it has never been collected. When a site shows you that number anyway, the number was produced by the site.
These are news reports, not data. They are kept apart from the statistics above on purpose.
A single incident, however vivid, tells you nothing about a rate. Coverage also clusters around dramatic events, so a run of headlines reflects what is newsworthy rather than what is frequent. Read this section for what happened; read the charts for how much.
Ceuta triggers two days of EU emergency talks
Andy Burnham promises to be relentless in tackling small boat crossings – video
Ceuta: Most who breached border have left, Spain says
Chaos of border crossing gives way to limbo for Moroccans on Ceuta’s streets
My highly educated Uber driver’s heartache underscores how transactional Australia’s visa regime has become | Julianne Schultz
EU to hold urgent meeting of interior ministers on Tuesday over Ceuta crisis
Germany: New 'compass classes' aim to integrate migrant children in education system
At least 57 people die as tens of thousands cross to Spanish enclave of Ceuta from Morocco
Why is Ceuta a migration flashpoint and how were so many people able to cross from Morocco?
Understanding Ceuta: A contested borderland in Europe’s African enclave
Spain's PM visits Ceuta after thousands breached border
Film: ‘The Lion at My Back’ shows the resilience of migrant women
Germany widens deportations to Afghanistan
Germany: Net migration falling due to changing policies and demographic patterns
French riot police used teargas on people trying to board small boat, charity claims
Spain sends extra troops and police to Ceuta after thousands of migrants arrive
Spain will send troops to Ceuta to help with border security
UK and German governments deploy AI in asylum process, as activists call out potential dangers
Channel crossing: French coast departures claim more lives as migrants attempt to reach UK
Germany deports 31 Afghan men to Kabul, including one with no criminal record
Ceuta: More bodies recovered as arrivals and deaths continue to rise
Over 200 migrants picked up by Greek authorities south of Crete
France: Homeless migrant families occupy plaza in front of emergency shelter offices
Refugees and migrants should not be exploited as political tools in campaigns, warns UN
Lithuania investigates potential 'migrant smuggling' tunnels from Belarus
AI tool will lead to more child refugees being treated as adults, charity warns
From the archive: ‘A dizzying maze’: how the UK immigration system is geared to reject – podcast
Morocco: Families search for those missing after attempting to swim to Ceuta
Greece: Series of arrests for smuggling migrants in vehicles
Eastern Libyan authorities claim they have deported more than 100,000 irregular migrants
Beaten, robbed and driven out: violent demonstrations target immigrants in South Africa
As Sudan’s war crosses borders and millions are made refugees, could Ethiopia’s plan solve ‘humanity’s migration issue’?
Out of Sight review – mountain idyll in the Austrian uplands is a gilded cage for distressed asylum seekers
Drawn from a fixed list of outlets — The Guardian, InfoMigrants, Deutsche Welle, Politico Europe — rather than an open search, and filtered to migration coverage. Headlines link to the publisher; nothing is rehosted here. Retrieved 2026-08-03.
If a number is not traceable to something in this list, it is not published on this site.
Eurostat · migr_asyappctzm · updated monthly
The number of asylum applications lodged in a reporting country in a given month, broken down by the applicant’s citizenship, age and sex.
Eurostat · migr_asypenctzm · updated monthly
The backlog: applications awaiting a first-instance decision at month end.
Eurostat · migr_asydcfstq · updated quarterly
Decisions issued at first instance, split into refugee status, subsidiary protection, humanitarian status and rejection.
Eurostat · migr_eiord · updated annual
People issued an administrative or judicial order to leave a Member State.
Eurostat · migr_eirtn · updated annual
People who actually left following an order, including voluntary and enforced returns.
Eurostat · migr_resfirst · updated annual
First residence permits issued to non-EU citizens, by reason: work, family, education and other.
Eurostat · migr_pop1ctz · updated annual
Resident population by citizenship, used here as the denominator for per-capita rates.
Eurostat · migr_asytpsm · updated monthly
People holding temporary protection at month end — a separate legal status from asylum, activated in March 2022 for those displaced from Ukraine.
Eurostat · migr_asyunaa · updated annual
Applicants recorded by the receiving state as unaccompanied minors.
Every figure on this site, as CSV. Regenerated on each refresh, so these files always match what is displayed. Empty cells mean not reported — they are never written as zero.
If a number here is wrong, it should be fixed rather than defended. The most useful report identifies the figure, the dataset it came from, and what the source actually says.
Two categories of error are worth distinguishing. A pipeline error is ours — a wrong filter, a bad join, a misread dimension. A source revision is Eurostat restating a figure, which happens routinely; the retrieval date on each section shows which vintage is displayed.
Because the data files are committed on every refresh, the change history is itself a record of what moved and when.