The claim: Visa grants equal arrivals.

Misuse

permission counted as arrival

Direction

restrictionist

What the data says

Two counts, two publishers, and they are not two versions of one number. In the year ending June 2026 the Home Office granted 727,487 entry clearance visas, excluding visitor and transit routes. Over the year ending December 2025 the ONS estimated that 813,000 people immigrated to the UK long-term. Both totals are on this page and nothing here subtracts one from the other: they cover different twelve months, they come from different publishers, and one counts permissions issued while the other estimates people who moved.

The biggest visa figure on this page is visits, and none of it is migration. The Home Office granted 2,185,158 visitor visas over the same year. A visit is not long-term international migration, which means a move intended to last twelve months or more, so those grants appear in no migration figure on this site at all. Anyone reading a visa total as an arrivals total has to decide what to do with that number first, and there is no answer that makes it an immigration figure.

The long-term routes are published separately, each on its own count. Sponsored study grants came to 383,455, work grants to 234,841 and family grants to 57,590. The study and work totals count main applicants and their dependants together, which is what the site's note on main applicants and people exists for. This page does not add these three up, and a reader should not either: a route total and a headline total are published on their own definitions, and the definitions do not line up merely because the labels do.

Why the claim is misleading

A grant is permission, and permission can go unused. The clearest case the data layer holds is the Ukraine schemes, where 290,044 out-of-country visas were granted between March 2022 and the end of June 2026. Out of country is in the name of the measure: it counts permission issued to people who were not in the UK, which is a different event from anyone landing. The grant total is not an arrivals total and nothing on this page turns it into one, and that is not a peculiarity of a humanitarian route. It is what every visa figure is: a count of permissions.

Some grants involve no arrival by design. An extension is a grant made to someone already living here, on a route they are already on. It adds to a grant count and moves nobody across a border. The Ukraine figure above is out-of-country grants only, and the extensions granted under those schemes are counted apart from it, which is the publisher being careful. The hazard is a headline "visas granted" written without that care, because the phrase itself does not tell a reader which of the two it is.

The unit is a grant, not a person. Someone granted two visas in the same year is counted twice, and a visa granted to a main applicant with dependants is counted for each of them. This site states the basis of every figure for that reason, and the caution is the one it applies to asylum figures as well: main applicants and people are different counts of the same reality.

And the two counts do not even share a clock. A visa granted in March may be used in September, next year, or never, and the ONS estimate counts a person as an immigrant on the basis of a move intended to last twelve months. So a grant and the arrival it permits can fall in different years, which is one reason the site's own glossary says these two figures will never match and are not supposed to.

What the claim gets right

Grants are the more recent number, and that is a real reason to reach for them. The Home Office publishes quarterly and the ONS twice a year, so the grant totals on this page cover a later twelve months than the immigration estimate beside them does. Anyone asking what has happened since is asking a question the immigration estimate cannot yet answer, and the grant figures can, read against their own earlier years rather than against the estimate. Setting one year-ending window beside another and calling the difference a change is a caution this site publishes in its own right.

And the two really are connected. A visa grant is what makes a long-term arrival lawful for somebody who needs one, so the claim is reaching for a real relationship rather than inventing one. It goes wrong at the word "equal": a connection between two counts is not an identity between them, and this page sets no figure against another to say how close it is.

The mirror claim fails too

"Visa grants tell you nothing about immigration" makes the opposite mistake. They are the permissions a long-term arrival is made under, they cover a more recent period, and they carry route-level detail. Both versions work by collapsing "these are not the same count" into "these are unrelated", which does not follow.

Better questions to ask

• Grants or people? The unit is a permission, and one person can hold more than one.
• Visits or long-term migration? Visit visas are granted in far larger numbers than the long-term routes, and none of them are in any migration figure.
• New permission or extension? An extension is a grant to somebody who is already here.
• Which publisher, and which twelve months? Home Office grant totals and ONS immigration estimates run on different definitions and different periods.
• Issued, or used? Nothing in a grant total records that the journey happened.

Source notes

Entry clearance, visitor, sponsored study, work and family visa grants, and the Ukraine schemes total: Home Office, immigration system statistics, year ending June 2026, published 27 August 2026. Long-term immigration: ONS, long-term international migration, provisional, year ending December 2025, a different publisher on a different period and a different definition.

This page does not subtract one of these figures from another, or form a share from them. The Home Office counts visas granted and the ONS estimates people who moved and stayed at least twelve months. The difference between the two totals is not a count of unused visas, and this site publishes no such figure, because these counts cannot produce one.

All common claims