What the words mean
Migration statistics use ordinary words in precise ways. Most public disagreements about the numbers are really disagreements about which number is being discussed.
A term earns a place here if it appears in the figures this site publishes and if misreading it produces a real error. That is the whole test. This is not a dictionary of immigration policy, and it deliberately omits terms that sound technical but change nothing about how a number should be read.
The entries are grouped rather than alphabetised, because these words are defined by their relationships to one another. Reading the first group first will make the rest easier.
How the counting works
Four ideas that cause more misreadings than any individual statistic. Almost every claim this site checks fails on one of them.
Flow and stock
A flow counts events over a period: people arriving during a year, claims made in a quarter, decisions issued in twelve months.
A stock counts a state at a single moment: people waiting for a decision today, people in supported accommodation on 31 March, people living in the UK who were born abroad.
Why it matters. The two answer different questions and cannot be compared or subtracted. 101,911 people claimed asylum during 2025, a flow. About 93,293 people were receiving asylum support on one day in June 2026, a stock. Those figures are of similar size and describe entirely different things. Someone who claimed in 2023 and is still waiting appears in the second but not the first.
The common mistake. Reading a stock as though it were an annual intake. See the claim check on who is in asylum accommodation.
Long-term international migration
A move to or from the UK intended to last 12 months or more. This is the basis for immigration, emigration and net migration figures.
Why it matters. It excludes visitors, short courses, seasonal work under twelve months, and business trips. Millions of people enter the UK each year who are not counted in migration statistics at all, because they are not migrating by this definition. In the year ending June 2026 the Home Office granted 2,185,158 visitor visas, none of which appear in migration figures.
The common mistake. Assuming migration statistics count everyone who crosses the border. They count a specific and much smaller group.
Non-EU+
A nationality grouping, not a route and not a visa type: people who are not British and not nationals of the EU, Norway, Iceland, Liechtenstein or Switzerland. ONS sorts long-term arrivals into three such groups, British, EU+ and non-EU+, and the three add up to the all-nationalities total.
Why it matters. ONS publishes its breakdown by reason for coming for non-EU+ arrivals only. Every reason share on this site, study, work, asylum and the rest, is a share of that group rather than of all immigration, and non-EU+ arrivals are the larger part of the total but not the whole of it. There is no published reason breakdown for everyone.
The common mistake. Reading a non-EU+ share as a share of all immigration, or hearing "non-EU+" as a statement about how someone arrived rather than about which passport they hold.
People, main applicants and cases
This site's single most important caveat, and the one most often responsible for two sources appearing to contradict each other.
- A main applicant is the person who makes an application.
- A dependant is a partner or child included in that application.
- People counts main applicants and dependants together.
- Applications or cases counts main applicants only.
Why it matters. The same underlying reality produces two very different numbers. In the year ending June 2026 there were 85,891 people claiming asylum, and 69,848 main-applicant applications. Both are correct. Neither is "the" figure.
The Home Office headlines people for asylum claims, but grant rates and case queues are usually published on main applicants. Mixing the two produces answers that are wrong by roughly a fifth, and the error is invisible unless you check the basis.
The common mistake. Dividing a people-based numerator by a case-based denominator. Every figure on this site states its basis for exactly this reason.
Provisional, revised and official statistics in development
Provisional means the publisher expects the figure to change as more data arrives. Revised means it already has. Official statistics in development is an ONS designation for statistics still being built and tested, which applies to current migration estimates.
Why it matters. These are not decorative labels. Net migration for the year ending December 2024 was first published as 431,000, revised to about 345,000, and now stands at 331,000, as ONS brought in new data and methods. Anyone who quoted the first figure was quoting the best available number, and was also going to be wrong within a year.
The common mistake. Treating the first published estimate as a fixed fact, and then treating its revision as evidence of manipulation rather than of how estimates work.
Migration measures
Immigration
People arriving in the UK for a long-term stay of 12 months or more. Around 813,000 in the year ending December 2025.
The common mistake. Treating immigration and net migration as the same thing. They differ by the entire size of the outflow.
Emigration
People leaving the UK for a long-term stay elsewhere. Around 642,000 in the year ending December 2025. It includes British citizens leaving, and this site publishes no figure for their share.
The common mistake. Forgetting it exists. Emigration is the half of net migration that rarely appears in coverage, and changes in it move the headline figure just as much as changes in arrivals.
Net migration
Immigration minus emigration. Around 171,000 in the year ending December 2025.
Why it matters. It is a balance between two much larger flows, not a count of anybody. Nobody is "a net migrant". It can fall because fewer people arrived, because more people left, or both.
The common mistake. Reading it as the number of people who entered the UK. See the full check on net migration read as a count of arrivals.
Visa grant
Permission to enter or remain in the UK under a particular route.
Why it matters. A grant is permission, not a journey. Some visas are never used. Some are extensions granted to people already here, who did not arrive that year at all. Visa figures come from the Home Office and count grants; migration estimates come from ONS and count people moving. The two will never match, and are not supposed to.
The common mistake. Treating visa grants as an arrivals count, or comparing a Home Office visa total directly with an ONS immigration estimate.
The asylum system
Asylum is one route within the wider migration system, and it has its own vocabulary. The sequence below roughly follows a claim through the system.
Asylum application
A request for protection from someone who says they cannot safely return to their country. Making a claim is not unlawful, and the route by which someone arrived does not determine whether their claim succeeds.
The common mistake. Treating applications, people and decisions as one unit. They are three different counts. See people, main applicants and cases.
Irregular arrival
Arrival without prior permission, or through an unauthorised route.
Why it matters. This describes how someone travelled, not what their legal status will be. A person may arrive irregularly and go on to be recognised as a refugee; another may arrive on a valid visa and later claim asylum.
A note on wording. This site uses "irregular arrival" as its default. Our own sources do not always: the Home Office publishes a dataset titled "Irregular migration to the UK detailed dataset (illegal entry routes)", renamed "Illegal entry routes to the UK detailed dataset" from the year ending December 2025 edition, and ONS reason splits include arrivals "via illegal entry". Where a source uses different words we quote its title as published rather than silently renaming it. The style guide sets out our reasoning, and distinguishes our precision rules from our value judgements.
Small boat arrivals
People detected arriving by small boat across the Channel: 33,374 in the year ending June 2026.
Why it matters. This is one route of arrival, not the asylum system. Most people who arrive this way do claim asylum, but they are a subset of all claims, and the figure is published on three bases at once: calendar year, year ending, and year to date. Those give three different numbers for what sounds like the same thing.
The common mistake. Using small boat arrivals as a proxy for asylum claims, or comparing figures on different bases without saying which is which.
Initial decision
The Home Office's first decision on a claim: a grant of protection or other leave, or a refusal.
The common mistake. Treating it as final. It can be appealed, and 39% of appeals decided in the quarter to June 2026 succeeded.
Grant rate
The share of initial decisions resulting in a grant of protection or other leave: 38% in the year ending June 2026.
Why it matters. The published rate is calculated on main applicants, and excludes withdrawals and administrative outcomes. It describes decisions made in a period, not the eventual outcome of the claims behind them.
A later rate exists, and it is a different measure. The Home Office also publishes, for each year of claim, the share granted at the latest outcome, following appeals and reconsiderations. It is materially higher. It counts a different group of people over a different period, so the two cannot be added together, and the refused-asylum check sets out what it does and does not show.
The common mistake. Recomputing the rate from people-based figures, which gives a different answer, or quoting the initial rate as though it were the later one.
Refugee
A person who has been granted protection. Not everyone who applies for asylum is a refugee, and not everyone granted leave to remain is granted refugee status specifically.
The common mistake. Using "refugee" for anyone who has claimed asylum, before any decision has been made. It prejudges the outcome in one direction, just as "failed asylum seeker" prejudges it in the other.
Backlog
There is no such thing as the backlog. There are several queues, at different stages, of very different sizes:
| Queue | Latest figure | What it is |
|---|---|---|
| Initial decision backlog | 40,168 people (30 June 2026) | Waiting for a first Home Office decision |
| Appeals backlog | 90,341 cases (30 June 2026) | Refused, appealing to the First-tier Tribunal |
| Total casework work in progress | 224,742 cases (June 2024) | Everything in the system. Not current: the breakdown was suspended |
Why it matters. These measure different things and move independently. The initial decision queue has fallen sharply while the appeals queue has risen, from 59,925 a year earlier to 90,341. Someone citing "the backlog" as one falling number, or one rising number, can be right about one queue and wrong about the system.
The common mistake. Collapsing them, or comparing a people-based figure for one queue with a case-based figure for another.
Asylum support and asylum accommodation
Accommodation, subsistence payments, or both, provided to people whose claims are being considered and who would otherwise be destitute: 93,293 people as at 30 June 2026.
Why it matters. It is a stock, counted on one day. It excludes unaccompanied asylum-seeking children, who are supported by local authorities and counted separately. Hotels are one form of accommodation among several, not the whole of it.
The common mistake. Reading the support figure as an arrivals figure, or treating "asylum accommodation" and "asylum hotels" as synonyms.
Return
A departure from the UK of someone with no right to remain. Enforced returns are carried out by the state; voluntary returns are not. Refused entry at port is counted separately again.
Why it matters. Returns count events, not people. One person returned twice appears twice. And the three categories must not be added together casually: port refusals are published separately from the enforced-plus-voluntary total precisely because they are a different thing.
The common mistake. Reading a returns total as a headcount of individuals, or folding port refusals into it to produce a larger number. See returns for what each category counts and why they do not add up.
Living in the UK
Settlement, or indefinite leave to remain
Permission to live in the UK without a time limit: 199,628 grants in the year ending June 2026.
The common mistake. Confusing it with citizenship. Someone with settlement is not a British citizen, cannot hold a British passport on that basis, and does not have the vote in general elections.
British citizenship and naturalisation
Becoming a British citizen. Naturalisation is the main route for adults: 245,520 grants of citizenship in the year ending June 2026.
Why it matters. Citizenship usually follows settlement rather than replacing it, so a person may appear in both statistics in different years. Neither figure is a count of arrivals: people granted citizenship in 2026 typically arrived years earlier.
Foreign-born population
People living in the UK who were born outside it: 16% of the population at the 2021/22 Census.
Why it matters. This is a stock accumulated over decades. Someone who arrived in 1975 is counted. It tells you nothing about current immigration levels.
A gap now partly filled. ONS discontinued its annual survey-based country-of-birth series in October 2022. It has since published a provisional rolled-forward estimate for mid-2024, an official statistic in development, so a later figure now exists. Neither it nor the census-based figure above is accredited at UK level: ONS accredits the England and Wales and Northern Ireland statistics in its census-based UK release and states that the UK ones have "just Official Statistics status, without accreditation". See the claim check on the born-abroad share for the number and its caveats.
Foreign national, or non-British citizen
People living in the UK who do not hold British citizenship: 10% of the population at the 2021/22 Census. As with foreign-born, ONS discontinued the annual series in October 2022, so there is no official figure for the current period.
Why it matters. This is a different measure from foreign-born and a smaller number. Many people born abroad have become British citizens; some people born in the UK are not British citizens. The two figures are not interchangeable, and which one a source is using is often unstated.
The common mistake. Treating "born abroad" and "foreign" as the same category.
Money
Asylum system cost
What the government spends running the asylum system: £4.9 billion in financial year 2024-25, covering the Home Office and Ministry of Justice.
Why it matters. This is an NAO estimate based on actual costs, and it is reasonably firm. The National Audit Office notes it does not capture costs falling on local authorities, so it understates the full figure.
The common mistake. Treating it as the fiscal impact of immigration. It is the running cost of one part of one system, and tells you nothing about taxes paid by migrants generally.
Visa and immigration fees
What the Home Office charges for applications to enter or stay: visas, settlement, citizenship and the rest. It collected £3.6 billion in fees in the financial year 2025-26.
Why it matters. Fees are set above the cost of processing for most routes, so the system is not simply cost-recovering, and the income is real money paid before anyone pays any tax. It is charged per application rather than per person, and a refused application is not refunded.
The common mistake. Setting fee income against the cost of the asylum system to argue that migration pays for itself, or that it does not. The two figures cover different populations: almost nobody claiming asylum is paying visa fees.
Immigration Health Surcharge
A charge paid up front by most people applying for a visa of more than six months, covering NHS use for the length of the visa. It raised £2.3 billion in the financial year 2025-26.
Why it matters. It is paid in addition to any tax and National Insurance the person then pays, and in addition to the visa fee, so the same person is charged for the NHS twice on two different bases. Asylum seekers do not pay it.
The common mistake. Reading a published surcharge total as the whole of it. The department's accounts print the figure in two places, what it keeps and what it passes to the Consolidated Fund, and a total quoting only the first is lower than the amount collected.
Net fiscal impact
Whether migration adds to or subtracts from the public finances overall, once taxes paid and services used are both counted.
Why it matters. There is no agreed figure. Across the pre-Brexit static studies summarised by the Migration Observatory, the estimated net fiscal impact of the migrant population is small in magnitude, under 1% of GDP, and comes out positive in some studies and negative in others. Those are separate studies of different groups, periods and methods, not one estimate with a margin of error, and no equivalent analysis covers the whole post-Brexit immigration system. The Migration Advisory Committee published a static and dynamic estimate for one route in December 2025, the 2022/23 Skilled Worker cohort, putting its lifetime net contribution at about £47 billion, and says it intends to extend the model to other routes.
The common mistake. Quoting one study's result as the answer. Both "immigrants pay more in than they take out" and "immigrants are a drain" fail this test equally, and each is checked on its own page: whether immigrants pay more than they cost and whether immigrants are a drain on the public finances.
This page was last reviewed on . Its figures are not a live count; every one was checked against its source on or after .