It’s been a tough few months for UK universities. Not only has the concept of a university education come under repeated attack from the mainstream media, but many are grappling with strict new compliance rules set out under last year’s immigration white paper.
A UK student sponsor must now keep its visa refusal rate under 5% to pass its annual Basic Compliance Assessment (BCA) – slashing the threshold by half even as key recruitment markets continue to see high refusal rates. However, withdrawn applications do not count.
In the year to June 2026, 11,500 student visa applications were withdrawn before a decision, double compared to the year before and the most in any year since 2010. The published refusal rate is 4.9%. But count the withdrawals and it rises to 7.5%.
How we got here
UK visa refusals are surging as the Home Office clamps down on immigration. Visa refusals for high-risk markets have surged over the past few months – with Pakistani applicants particularly badly affected.
According to Home Office data, between January and March 2026, more Pakistani applications were withdrawn than were granted. Some 2,362 Pakistani applicants were granted a UK student visa, 1,531 were refused, and 2,920 withdrew their applications before the Home Office had decided their fate.
Across all nationalities, 6,970 student visa applications were withdrawn in those three months against 5,499 refused: the first time outside the pandemic quarter of April to June 2020 that withdrawals have outnumbered refusals.
The rule that gives those withdrawals their meaning took effect on June 1, 2026. For compliance assessments applied for on or after that date, a student sponsor must have a visa refusal rate under 5% – half that of the previous 10% threshold.
The refusal rate is measured over the 12 months before the assessment, and the sector as whole has been moving towards the line
Exceeding it does not cost the licence outright: a first failure brings an action plan, a cut of at least 10% in the sponsor’s CAS allocation and a final warning.
The refusal rate is measured over the 12 months before the assessment, and the sector as whole has been moving towards the line: 3.1% in the year to September 2025, then 3.8%, 4.3% and 4.9% in the three quarters since. The 4.9% for the year to June 2026 (all student and Child Student visa applicants, dependants included) is the highest since 2016.
That rise has already been reported. What has had less attention is the second column in the same Home Office table: applications withdrawn before a decision.
From 2019 to the third quarter of 2023, that column never passed 850 in a quarter. In the fourth quarter of 2023, it jumped to 2,470. Over the year to June 2026, 11,500 applications were withdrawn: double compared to the year before and six times higher than the year to June 2023.
Why a withdrawal is worth more to a university than a refusal
The reason is written into the rules. A CAS can be used once. The UK’s immigration rules treat a CAS as used when the application it supported was granted or refused, but an application that was withdrawn does not use it up.
The sponsor guidance applies the same test to the refusal rate: it counts a CAS only when the application it supports “has been decided”, and the refusal rate is the share of those decided applications that were refused.
A student who withdraws has applied with their CAS, paid the fee and given biometrics. But no decision was made so for the compliance assessment, that CAS does not exist. The application is counted in the national statistics, in the withdrawn column. It is not counted anywhere in the university’s refusal rate.
Universities have a second route, and it was written into the guidance in June 2026 – in the same update that cut the visa refusal limit to 5%. Paragraph 2.19 says that a CAS the sponsor withdraws before a decision “will not be included for the purposes of the refusal rate calculation”.
The difference is what happens to the student. When a university withdraws a CAS while the application is pending, the application is normally refused because the rules require the sponsor not to withdraw its offer. The student then has a refusal on their immigration history.
Both routes protect the university – but only one protects the student. That is why, when a university wants an application gone, the instruction is usually for the student to withdraw it themselves.
Both routes protect the university – but only one protects the student. That is why, when a university wants an application gone, the instruction is usually for the student to withdraw it themselves
The PIE News described this practice during that 2024’s January intake delays: “Students are being told to withdraw their applications themselves as if universities do so, the automatic visa refusal that follows could mark the applicant’s record.”
In 2024, the trigger for withdrawals was that students had missed their start date due to visa processing delays. Since then, the point at which a student is asked to withdraw has moved earlier at some universities, to the moment UKVI invites the student to a credibility interview.
The Home Office caseworker guidance says an applicant must not be refused on genuineness grounds without an interview except in narrow cases, which makes an interview invitation the clearest signal a caseworker gives that an application is in doubt. Some universities have decided to act on that signal before the Home Office does.
The refusal rate alongside withdrawals
The published refusal rate is refusals divided by refusals plus grants, leaving withdrawals out by definition. A simple second measure puts them back: refused plus withdrawn, divided by everything that reached an outcome. It is not a measure the Home Office publishes or the licence uses. It is the share of applications that did not end in a visa.
On that measure, the sector’s year to June 2026 is not 4.9% but 7.5%: 19,608 refused and 11,500 withdrawn out of 414,563 applications that reached an outcome.
Quarter by quarter the gap is wider. In the first quarter of 2026 the published rate was 13.4%. With withdrawals counted, it was 25.9%. One outcome in four recorded in the January intake window was not a visa.
Not every withdrawal is a refusal avoided. Students withdraw because they chose another country, because a visa decision did not arrive before their start date, or because their plans changed. The question is how much of the 11,500 withdrawals that can explain.
Withdrawals now follow refusal risk. In 2023 they did not.
If withdrawals were mostly students changing their minds, they would be spread across markets in rough proportion to application volumes. If they were mostly applications pulled to avoid a refusal, they would cluster where refusals cluster. The Home Office data can tell the two apart because it reports withdrawals by nationality.
In the year to June 2023, withdrawals looked like students changing their minds. The largest single source of withdrawals was China, with 288, and China’s refusal rate that year was 0.2%. France, Germany, the US and Spain, all with refusal rates under 1%, were in the top 10.
Across the 66 source markets with 500 or more applications, the share withdrawn was between 0.3% and 0.7% in every band of refusal risk. There was no relationship between how likely a market was to be refused and how likely it was to withdraw.
By the year to June 2026, that had changed. Markets with a refusal rate under 2% withdrew 0.4% of their applications, whereas markets between 2% and 5% withdrew 1.8%.
This trend becomes more start the more high-risk the market. Markets with refusal rates between 5% and 10% withdrew 3.4%, whereas markets at 10% and above withdrew 8.4%. Across the 57 markets with 500 or more applications, the correlation between a market’s refusal rate and its withdrawal share is 0.77, on a scale where 1 means the two rise and fall together and 0 means no relationship. In the year to June 2023, there was no relationship at all (the correlation was actually slightly negative).
If every market had withdrawn at the rate the low-risk markets did, the year to June 2026 would have seen about 1,700 withdrawals. Instead, it saw 11,500. The other 9,800 or so are concentrated in just four countries.
YEAR TO JUNE 2026, MAIN MARKETS
| Market | Issued | Refused | Withdrawn | Published rate | With withdrawals | Share of all withdrawals |
| Pakistan | 21,327 | 4,778 | 3,594 | 18.3% | 28.2% | 31.3% |
| Nigeria | 33,080 | 4,344 | 2,964 | 11.6% | 18.1% | 25.8% |
| India | 83,958 | 3,846 | 1,731 | 4.4% | 6.2% | 15.1% |
| Bangladesh | 10,021 | 1,635 | 808 | 14.0% | 19.6% | 7.0% |
| Ghana | 4,155 | 885 | 297 | 17.6% | 22.1% | 2.6% |
| Nepal | 17,150 | 827 | 231 | 4.6% | 5.8% | 2.0% |
| China | 86,614 | 102 | 198 | 0.1% | 0.3% | 1.7% |
| United States | 16,850 | 70 | 88 | 0.4% | 0.9% | 0.8% |
| All markets | 383,455 | 19,608 | 11,500 | 4.9% | 7.5% | 100% |
Published rate = refused / (issued + refused). With withdrawals = (refused + withdrawn) / (issued + refused + withdrawn). Source: Home Office, Immigration system statistics, year ending June 2026. Basis: Student and Child Student routes, main applicants and dependants.
Pakistan and Nigeria between them account for 57% of the year’s withdrawals and 17% of its applications.
For Pakistan, the first quarter of 2026 reads 39.3% refused on the published basis and 65.3% with withdrawals counted: two in three Pakistani applications that reached an outcome did not become a visa. For Nigeria, the same quarter reads 21.4% and 38.9%. For Bangladesh, 27.2% and 45.0%.
This is the pattern you would expect if applications were being pulled to avoid refusals – not if students were changing their minds – and it is what people in the sector mean when they talk about self-regulation. It is consistent with universities and agents managing the number the licence is judged on, one application at a time.
Processing delay is the other explanation. The same markets that get refused get interviewed, and the same markets that get interviewed wait longest for a decision. A student who withdraws because the January start date has passed looks identical in this data to one who withdraws because the university asked.
The same markets that get refused get interviewed, and the same markets that get interviewed wait longest for a decision
Previous reporting from The PIE described exactly that: Pakistani students withdrawing not because they wanted to but because a decision never came.
The Pakistani figures show both forces at once. In the fourth quarter of 2025 the refusal rate was 39.4% and 374 applications were withdrawn. In the first quarter of 2026 the refusal rate was the same, 39.3%, and 2,920 were withdrawn.
Refusal risk did not change between those two quarters. What changed was that the January intake arrived, with its interviews, its delays and its deadlines. The data cannot say how many of the 2,920 Pakistani students who withdrew their visa application were asked to do so by their university and how many ran out of time before their start date. Only the Home Office can let us know for certain.
Nothing in writing
For this piece, I looked for the written version of the “tell us if you are called for interview and we will withdraw you” policy on the public pages of about 115 UK universities, pathway providers and private colleges. I could not find one.
What universities do say in public about the UKVI interview is the opposite: attend it. The University of Lincoln’s page says: “If you are asked to attend an interview, you must do so.”
Queen Margaret University tells applicants: “Failure to pass the interview will result in a visa refusal, and the university is unlikely to be able to issue a new CAS for a further application.”
Similarly, the University of Westminster asks students refused after a credibility interview to inform its visa compliance team, and says it “is unlikely to issue a second CAS for a new visa application unless your administrative review is successful”.
Everywhere I looked, the published position of every university is that the interview is the student’s to pass and the refusal is the student’s to carry.
Many universities also publish carefully worded positions around withdrawals. The University of Brighton’s deposit refund table has a row for “visa application withdrawn before UKVI decision”: refund yes, less a £200 fee.
The University of Staffordshire refunds a deposit in full “where a CAS is withdrawn by the university prior to a decision from the UKVI”. Both universities have a process for applications that end before the Home Office decides them. Neither says when it asks a student to start one.
The instruction that follows a UKVI interview invitation reaches the student or the agent directly, and no university has put its name to it in public.
Earlier this year, one PIE commentator described Pakistani students who “are being called for a UKVI visa interview, already a stressful and time-consuming step, only to find that their university has quietly cancelled their CAS”.
If interview invitations are being used this way, then a policy that decides the outcome for a large number of applicants a year is being run through private correspondence, with nothing a student could read before paying a deposit and nothing an agent could show a family before they applied.
The student’s choice
Consider the student who receives the interview invitation and has been told by their university to report it. They have three options and none of them are good.
If they tell the university and withdraw, they lose the place for that intake. The visa application fee, £558 from outside the UK, is not usually refunded once biometrics have been given, which by the interview stage they have. On the plus side, the student’s immigration health surcharge is refunded.
Whether the deposit comes back depends on the university: Brighton refunds less £200, Staffordshire in full if the university withdrew the CAS, and many policies are silent on a withdrawal the university itself asked for. What the student does keep is an immigration history with no refusal on it. A withdrawn application does not count as a refusal, nor does it have to be declared as one on a student visa application.
If they stay quiet, attend the interview and pass, students may be able to keep their place, and the university may never know. The Home Office writes to the applicant, and nothing in the guidance requires the sponsor to be told. The policy depends on students volunteering the information that will be used against them, and it rewards the ones who do not.
If they stay quiet, attend and fail, they carry a refusal that must be declared on every future UK visa application, they lose the fee, and at some universities they lose the deposit too: the University of Hull’s deposit policy of May 5 2026 lists “Credibility (eg following a UKVI interview)” among the refusal grounds that carry no refund.
The university’s refusal rate takes the hit it was trying to avoid. The student who was honest and withdrew and the student who gambled and lost both lose the same intake. But only the student who attended the interview has a refusal on their record.
This scenario is hypothetical – but the past few months will have seen many such students trying to make their way to a UK university
This scenario is hypothetical – but the past few months will have seen many such students trying to make their way to a UK university. After all, refusal rates of 32.1% for Pakistan and 24.4% for Bangladesh in the second quarter of 2026 are not the product of a few bad decisions.
But universities should keep in mind that an interview invitation does not mean an automatic refusal. The Home Office requires an interview so that a doubtful case can be tested rather than assumed.
Universities that withdraw on invitation are assuming the outcome, and they are assuming it about the students least able to absorb the cost of being wrong: the deposit, the fee, the flights, the accommodation contract, the year.
The agent in the middle
A university’s withdrawal policy cannot work without agents, and in most cases agents were not asked for their opinion.
An agent is usually paid on enrolment, so their short-term interest is for the student to attend the interview and take the chance. Their long-term interest is their own refusal rate with each partner university, because an institution running near 5% will look hard at agents whose students are refused.
Every student an agent persuades to withdraw costs them a commission, or at best defers it to the next intake. The reward for doing it is staying on the university’s partner list.
PIE commentator Muhammad Rizwan Riaz described the experience from Pakistan back in May: “No email. No phone call. Just a changed status on a screen.”
The complaint is not that universities are cautious. It is that the caution arrives without warning, without a written policy the agent can put in front of a family, and after the deposit has already been paid.
A published policy in the offer terms, a deposit refund when the university asks the student to withdraw, and written notice before any CAS is withdrawn would not stop a university protecting its licence. They would simply stop the cost of that protection falling on the people with the least information.
What it all means for the 5% threshold
The Home Office set the 5% visa refusal limit to make sponsors more selective, and sponsors have become selective in two ways.
One is the outcome the policy presumably intended: fewer doubtful applications are made at all.
The other is the subject of this piece: applications that are made, and look doubtful, are taken out before the Home Office decides their fate. The withdrawal column is where the second kind ends up, but the assessment does not count it.
The compliance test is applied sponsor by sponsor and counts main applicants only, so the fair sector-wide comparison is on main applicants: a published rate of 4.5% for the year to June 2026, with 10,875 applications withdrawn.
About 1,900 more refusals would have taken that rate to 5%. That is one withdrawal in six. On the wider basis used elsewhere in this piece, dependants included, the sector is at a 4.9% refusal rate and the gap is about 570 refusals, one withdrawal in 20. Either way, the number of withdrawals is far greater than the sector’s distance to the threshold.
Two things would settle how much of this is deliberate. The first is Home Office data on credibility interviews: how many invitations were issued, how many applications were withdrawn after an invitation, and how the interviewed applications were decided. The Home Office is likely to hold that data and does not publish it.
A Freedom of Information request should produce it, and I intend to make one. The second is for universities to put the policy in writing where students can read it. If the practice can be defended, and it can, it can also be published.
