The government says Britain’s modern slavery system can be exploited to delay removals, and it is rewriting the rules to stop that. Its own latest statistics show a system at record volume, a backlog that has started growing again, and a gap where the evidence for “misuse” would be.
On September 24, 2026, the Home Office published the National Referral Mechanism (NRM) statistics for the year to June 2026. They record 25,373 potential victims of modern slavery referred in twelve months, up from 24,108 in the year to March and, in the department’s words, “the highest number of referrals received in a year since the NRM” began. The government’s Immigration and Asylum Bill, introduced on June 30, 2026 and still before Parliament, contains provisions that would tighten who can be treated as a trafficking victim, on the stated ground that the system is open to improper use.
Concern about abuse of the system is a legitimate policy position, and RSN does not treat it as pretext. But a government that proposes to restrict a protection on the grounds of misuse should be able to show the misuse. The department’s own latest release suggests it cannot yet do so, or at least has not published the data that would allow anyone to check.
What the Home Office statistics record
The year-to-June 2026 bulletin sets out the following. Of the 25,373 referrals, 77 percent (19,622) went to the Single Competent Authority (SCA) and 23 percent (5,751) to the Immigration Enforcement Competent Authority (IECA). By nationality, UK nationals were the largest group at 22 percent (5,548), followed by Eritreans at 13 percent (3,220) and Somalis at 11 percent (2,741). Fifty-four percent of referrals (13,591) concerned exploitation that took place exclusively overseas, up from 52 percent a quarter earlier and the highest share on record. Exclusively UK exploitation accounted for 38 percent (9,564).
On outcomes, 23,573 reasonable grounds decisions were issued, 64 percent of them positive. Of 21,564 conclusive grounds decisions, 71 percent were positive. These headline rates conceal a sharp split between the two decision-makers. The SCA found reasonable grounds in 73 percent of its cases; the IECA, which sits within the immigration enforcement function, did so in 34 percent. At the conclusive stage the figures were 72 percent for the SCA and 49 percent for the IECA.
The gap between the two authorities is not, on its own, evidence of anything. They handle different caseloads, and the IECA deals largely with people already in the enforcement process. It does, however, show that the outcome of a claim depends heavily on which part of government decides it, and that is a fact any reform debate should address directly.
Who is making the referrals
The “misuse” argument carries an implicit picture: claimants and their advocates using the system to frustrate removal. The referral data complicates that picture. Fifty-eight percent of referrals (14,764) came from government agencies. Within that group, UK Visas and Immigration made 5,767 referrals, Immigration Enforcement 3,493 and Border Force 2,832. Those three Home Office bodies alone account for 12,092 referrals, about 48 percent of the total. Non-governmental organisations and the third sector made 954, or 4 percent.
RSN’s reading, offered as analysis rather than as a finding of the bulletin, is this: nearly half of all referrals originate with the department that is now describing the system as vulnerable to abuse. If the volume is a problem, a substantial part of the remedy lies in the department’s own first-responder practice, not only in the legal test that follows. The statistics do not say whether those referrals were well founded, and they do not say how many were later rejected. That brings the argument to its weakest point.
The evidence gap
The bulletin lists its own limitations. Publication was delayed by problems extracting a robust dataset, linked to a planned IT change. The data was taken from a live system between August 8 and 20, 2026, so figures may change. Backlog figures are “not comparable with previous bulletins” because the method changed. And the reasons for negative decisions are omitted from this release, because “recording changed in April 2026”; the Home Office says they will return on a quarterly basis.
That last omission matters most. The central claim in support of the Bill is that some claims are made in bad faith or late, to delay removal. Whether a negative decision rested on a finding of bad faith, on lack of evidence or on something else is exactly the information that would test that claim. It is absent from the release.
The bulletin does contain data on disqualification, the existing power to deny protection. There were 144 disqualification requests in the year: 137 on public order grounds and 7 on bad faith. Of 119 confirmed disqualifications, 116 were on public order grounds and 3 on bad faith. Against 25,373 referrals, three confirmed bad-faith disqualifications is a small number. It does not prove that misuse is rare, since the existing test may be hard to meet, which is itself an argument the government can make. But it is the only published figure on the point, and it does not show a large problem. The bulletin also notes that public order disqualification decision-making has been paused since December 23, 2025. It does not give the reason in the passage RSN reviewed, and RSN does not speculate about it.
The backlog is moving the wrong way
The department’s operational record is the second half of the case. Cases awaiting a conclusive grounds decision stood at 5,656 at the end of June 2026, up 23 percent from 4,604 at the end of March. The median wait rose from 50 days to 84. The bulletin attributes the fall in decisions in April to June to staff being moved to a reasonable grounds backlog. A separate August bulletin put the figure at 5,948 on August 31, up 13 percent from 5,258 at the end of July, with the median wait down to 69 days. The government has “committed to eradicate” the backlog by December 2026.
That is a long way down from the peak, and the bulletin notes the backlog is 82 percent below it. But the direction since March is up, the method of counting has changed, and the deadline is under three months away. Delay is the very outcome the Bill is designed to prevent. A system that takes eighty-odd days to reach a conclusive decision is one in which a late claim can plausibly delay a removal, and part of that delay is administrative.
What the Bill proposes
According to the Home Office’s factsheet on the Bill, the legislation clarifies “what factors are to be considered as damaging to a person’s credibility,” makes clear where people can be disqualified on bad faith and public order grounds, and “removes the requirement to grant Temporary Permission to Stay (TPS) to confirmed victims for recovery purposes.” The factsheet refers to “the potential for improper use” and says the reforms aim to “reduce opportunities for improper use.” It adds that courts “interpret the UK’s international obligations in ways that are not aligned” with the system’s intended operation.
The clause text is more specific. As reproduced by Free Movement’s bill tracker, clause 38 inserts a new section 63A into the Nationality and Borders Act 2022. It covers adults who have “claimed to be a victim of slavery or human trafficking in bad faith.” Its examples include a claim “made at a time when the person was due to be removed” and information given “knowing the information to be false or fabricated.” There are exceptions for late claims where a notice deadline was met or there are “compelling reasons” for the delay, and the Home Secretary could add further circumstances by regulations under the affirmative procedure. A Green Party amendment to remove clause 38 entirely has been tabled but not debated. Whether disqualification would become mandatory is stated differently in the bill tracker’s marked-up text and in the explanatory notes it cites, so readers should consult the bill itself before relying on either account.
The other side of the argument
Critics, including anti-slavery charities, argue that the changes will deter genuine victims and weaken protections. RSN does not adopt that position, and the evidence available does not settle it. A government has a legitimate interest in preventing any asylum or protection mechanism from being used to postpone lawful removal, and a late claim is not proof of falsity but is a recognised indicator of the risk. The Home Office is entitled to say that it, and not the courts alone, should define the limits of a statutory scheme.
The courts have also scrutinised how the department handles the other end of the process. RSN reported in July on a High Court ruling on the Home Office’s treatment of trafficking-victim protections, and last week on a High Court finding that the department’s defence in a detention case implied the existence of accommodation that had been closed since 2020. Earlier this month RSN examined how a falling asylum decision backlog measured only part of the system. A consistent theme runs through these items: the department’s published account of its own performance is more limited than the policy argument assumes.
Proven, reported, and not established
Established by the Home Office’s own statistics: referrals reached a record 25,373 in the year to June 2026; 54 percent concerned exploitation exclusively overseas; about 48 percent of referrals came from UKVI, Immigration Enforcement and Border Force; the SCA and IECA reach very different positive rates; the backlog rose between March and June; reasons for negative decisions are omitted from this release; and three bad-faith disqualifications were confirmed in the year. Stated by the government as its rationale: that the system is open to improper use and that courts have read international obligations in ways that undermine it. Not established by the published data: how widespread bad-faith or delay-motivated claims are, and how many negative decisions rested on such findings. RSN has not seen evidence on either point, and says so.
The accountability question is therefore a narrow one. If the Home Office believes the modern slavery system is being exploited, it should publish the data that shows it, restore the reasons for negative decisions to its quarterly releases, and explain how it will measure whether the Bill’s changes work. Legislating first and evidencing later reverses the order a serious department should follow.
Sources
- Home Office, Modern slavery: National Referral Mechanism and Duty to Notify statistics UK, year to June 2026, published September 24, 2026.
- Home Office, NRM cases awaiting a conclusive grounds decision, August 2026, published September 10, 2026.
- Home Office, Immigration and Asylum Bill 2026 factsheet: modern slavery.
- UK Parliament, Immigration and Asylum Bill (Bill 105, introduced June 30, 2026).
- Free Movement, Bill tracker, clause 38 (disqualification from protection). Figures and clause text above were read through a summarizing tool; verify against the primary documents.