David Lammy’s Transparency Paradox: The Crime Data His Ministry Does Not Want Britain to See

Sixteen late declarations, corporate access to the Foreign Office, judge-only courts, wrongful prisoner releases and a legal battle over convictions by nationality.

Jul 20, 2026

The most consequential political scandals are not always discovered in envelopes of cash or leaked messages. Some are conducted through official channels, defended by departmental lawyers and hidden beneath the antiseptic language of exemptions, administrative cost and procedure. They reveal something larger than private misconduct: who controls the evidence, who is expected to accept official conclusions on trust and what happens when the public asks to inspect the machinery for itself.

David Lammy’s Ministry of Justice is now at the centre of precisely that test. The disputed information is not a list of defendants, private case files or personally identifying records. It is aggregated criminal-justice data: the total number of convictions recorded against people of different nationalities, divided into broad offence groups. Such information is not sought merely to satisfy curiosity. Government itself uses statistical patterns to identify operational pressures, allocate resources and formulate policy. The political question is why the public should be denied a suitably anonymised version of data important enough for the state to collect.

On 22 June 2026, the Information Commissioner rejected the Ministry’s principal reasons for refusing the request. The Ministry had initially relied upon section 40(2) of the Freedom of Information Act, arguing that disclosure would involve personal data. During the Commissioner’s investigation it added section 12, claiming that locating and extracting the information would exceed the statutory cost limit. The Commissioner found that the Ministry had not demonstrated that section 12 was engaged and concluded that the personal-data exemption could protect cells containing fewer than five convictions, but not figures of five or more. The decision required a fresh response on that basis. During the investigation, the regulator had also issued the Ministry with a formal Information Notice on 5 May 2026.

The regulator therefore identified the conventional mechanism for reconciling transparency and privacy: suppress the smallest figures and disclose the larger aggregates. It did not order the publication of names or information capable of casually identifying individual offenders. Nor did it pretend that nationality was an uncomplicated measure. It ruled on whether the Ministry had justified withholding the requested statistical information under the exemptions it had chosen to invoke.

On 19 July, The Telegraph reported that the Ministry was appealing the Commissioner’s decision. At the time this article was finalised, no appeal filing or tribunal case page had been located in the public sources reviewed. The existence and reported purpose of the appeal are therefore attributed to that report rather than presented as independently proven from a published pleading. This distinction is not pedantry. It is the line separating documented investigation from political theatre.

What is established—and what is not

Confirmed - The Ministry refused aggregated convictions-by-nationality data using sections 40(2) and 12 of the Freedom of Information Act. The ICO rejected the cost argument and limited the personal-data exemption to figures below five.

Reported - The Telegraph says the Ministry is appealing the ICO decision. The article should be updated when the appeal notice, grounds or tribunal listing become public.

Not established - No publicly available document reviewed proves that Lammy personally ordered or authorised the appeal. What exists at present is ministerial and departmental responsibility, not evidence of a private instruction.

That limitation does not remove Lammy from the story. The government lists him as Lord Chancellor, Secretary of State for Justice and Deputy Prime Minister, with responsibility for the Ministry’s strategy, resources and the constitutional functions of the Lord Chancellor. If a politically explosive transparency case was escalated without his knowledge, the public is entitled to ask why. If he was consulted and approved the challenge, he should explain his reasoning. If he believes the figures should be released, he has both the authority and the platform to make that position known.

Ministerial responsibility cannot be switched off whenever a departmental decision becomes uncomfortable. Lammy need not have typed the legal submission himself to be accountable for the posture adopted by the institution he leads. His silence does not prove personal authorship of the appeal, but neither does it absolve him of the obligation to explain what his department is doing in his name.

The Information Vacuum Is a Political Choice

The government already accepts that foreign-national-offender data is a legitimate subject of official publication. On 31 March 2026, England and Wales held 10,487 foreign nationals in custody, approximately 12 per cent of the prison population. That total comprised 3,588 people on remand, 6,458 sentenced prisoners and 441 people held in a non-criminal category. Foreign nationals also accounted for 18 per cent of first receptions into prison during 2025.

Those numbers are important, but they are not a simple foreign-versus-British offending rate. A prison snapshot is shaped by remand decisions, sentence length, offence severity, deportation delays, age, sex and the underlying population denominator. First receptions measure admissions, not unique offenders, and cannot be compared casually with a stock population measured on one date. The presence of non-criminal detainees further demonstrates why crude interpretations are unreliable. The correct conclusion is not that the prison count proves a sweeping claim about “migrants”. It is that nationality is already operationally significant within the justice system and that properly structured conviction data is needed to distinguish real patterns from propaganda.

The Home Office has reached much the same conclusion. Its foreign-national-offender statistics page was updated on 21 May 2026 after an upgrade to the relevant data collection and infrastructure. It began publishing additional breakdowns for offenders leaving immigration detention and those removed from the United Kingdom, including common nationalities, offence groups and sentence-length groups. The department also acknowledged that further work was required before it could release expanded data on offenders entering immigration detention and foreign offenders living in the community.

This is a crucial admission. The government does not maintain that nationality and offence data is inherently unlawful, irrelevant or impossible to publish. It is actively improving some datasets and placing them into the public domain. Yet the Ministry of Justice resisted a separate request for conviction totals and is now reported to be taking the matter further after the regulator rejected its principal exemptions.

That selective architecture produces the worst possible public debate. People determined to portray every migrant as a criminal fill the absence of robust evidence with anecdotes, fabricated graphics and decontextualised prison figures. People equally determined to insist that migration has no measurable relationship with crime dismiss almost any request for data as inherently prejudicial. Both positions thrive because the state controls the information required to test them.

The answer to inflammatory or easily abused data is not concealment. It is publication with definitions, denominators, limitations and enough methodological detail to prevent dishonesty from becoming the only available interpretation.

If the withheld figures show no meaningful overrepresentation, publication would puncture exaggerated claims. If they show serious concentrations among particular nationalities or offence categories, government has a duty to understand and address them. If the records are too incomplete, fragmented or unreliable to support analysis, that administrative failure is itself a matter of legitimate public concern. Every credible outcome favours disclosure. The outcome most threatened by disclosure is political convenience.

The Lammy Review Established the Standard He Is Now Expected to Meet

The contradiction becomes sharper when set against Lammy’s own work. His 2017 review into the treatment of, and outcomes for, Black, Asian and minority-ethnic people in the criminal-justice system did not treat sensitive demographic data as something best kept within Whitehall. It treated the absence of accessible information as one of the obstacles to trust.

The official Lammy Review argued for bringing decision-making into the open and allowing outsiders to scrutinise the system. Its third recommendation said the default should be for the Ministry of Justice and criminal-justice agencies to publish the ethnicity datasets they held while protecting personal information. A further recommendation called for sentencing information to be broken down by offence and court, with demographic characteristics including gender and ethnicity.

The methodological distinction between ethnicity and nationality is real. Neither can be used as a substitute for immigration status, and neither explains causation. But the democratic principle is identical. A government should not embrace granular demographic breakdowns when they may reveal discrimination by institutions, then retreat behind sensitivity and administrative burden when different breakdowns may reveal disparities in offending or conviction patterns.

Lammy’s review did not say transparency was valuable only when it produced politically agreeable findings. It presented scrutiny as a mechanism through which unfairness could be detected, challenged and corrected. The present dispute tests whether that commitment survives when the evidence may be interpreted in ways that are uncomfortable for the government responsible for immigration and criminal-justice policy.

This is the central paradox. Lammy built authority by arguing that trust cannot be commanded from communities whose experience of justice is obscured by incomplete information. His department is now asking the wider public to trust its handling of nationality and crime while resisting the release of the data capable of testing that trust.

Sixteen Late Registrations—and a Previous Warning

The nationality-data case would demand explanation under any Justice Secretary. Under Lammy it sits beside a documented personal history of delayed disclosure. In 2022, the Parliamentary Commissioner for Standards found that Lammy had registered sixteen registrable interests outside Parliament’s 28-day deadline. The original complaint identified eight entries. Further examination found seven more, and Lammy identified an additional late entry before the matter concluded.

Precision is necessary here. These were not sixteen proven bribes and should not be described as such. They were sixteen late registrable interests covering a mixture of paid speaking engagements, hospitality, overseas visits and unpaid positions. The rectified register records examples including paid appearances for commercial organisations, NFL and boxing hospitality, visits associated with Tufts University and the Bilderberg meeting, and unpaid roles. The Commissioner concluded that the failures resulted from “inattention to detail”, accepted that there was no deliberate intention to mislead and resolved the case through rectification after Lammy apologised and revised his office procedures.

Those mitigating findings belong in any honest report. They do not make the pattern irrelevant. Parliament’s registration rules are time-sensitive because disclosure delayed by months cannot perform the same function as disclosure made while an interest is current. Journalists, voters and other MPs are supposed to be able to assess a member’s outside earnings, hospitality and associations when relevant decisions and interventions are taking place—not after pressure has forced the record to catch up.

The Commissioner also considered an earlier 2019 inquiry into late registration. Lammy had apologised then and described procedures intended to prevent a recurrence. When similar failures emerged again, the Commissioner said the decision on whether to refer the matter for more serious consideration had been “finely balanced” and warned that a future allegation would be referred directly.

Nothing in that decision proves personal corruption. It proves something narrower and directly relevant to the present controversy: the minister now politically responsible for a department resisting disclosure has himself twice required parliamentary scrutiny over information that was not disclosed on time. He should understand better than most that reluctant, delayed transparency is not a substitute for openness.

Corporate Access to the Foreign Office

The same concern over who receives access—and under what safeguards—arose while Lammy was Foreign Secretary. In a March 2025 address to the British Chambers of Commerce, he announced that the Foreign Office would deepen its relationship with business. He said he had written to chief executives asking what diplomatic support they required, invited private-sector personnel to gain experience inside the department and announced that a British diplomat would be seconded to BAE Systems.

There is nothing inherently corrupt about secondments. Government can benefit from expertise in finance, technology, manufacturing, energy and trade. Civil servants may also gain a more realistic understanding of the industries affected by diplomatic and regulatory decisions. But the case for external expertise does not eliminate conflicts of interest; it makes transparent safeguards more important.

BAE Systems is not a neutral educational institution. It is a major defence contractor with direct interests in military procurement, strategic partnerships, export licensing and conflict. Shell is directly affected by sanctions, climate policy, energy diplomacy, taxation and relationships with resource-producing states. A placement scheme involving employees from such companies raises obvious questions about access to officials, policy formation, commercially valuable knowledge and the risk that public institutions become more permeable to organisations already capable of securing ministerial attention.

In August 2025, the legal firm Leigh Day confirmed that The Corner House had sent a pre-action letter challenging the transparency and governance of proposed exchanges involving Shell and BAE Systems. The concerns included potential improper corporate influence, unmanaged conflicts, insufficient public information and the need for clear whistleblowing arrangements. A pre-action letter is not a judicial finding. It does not prove that either company improperly influenced policy. It demonstrates that the proposed structure raised sufficiently serious public-law concerns for campaigners to threaten judicial review and demand the underlying terms.

The proper questions were never difficult. Which roles would secondees occupy? What documents and meetings could they access? Would they be excluded from policy affecting their employer? How would commercially sensitive information be protected? What declarations would be required, and who would enforce them? Would competing firms receive comparable access? What knowledge could participants carry back to their companies at the end of the placement?

A government confident in the arrangement should have published the framework, conflict rules, selection criteria and information barriers before the programme became controversial. Once again, the institutional instinct appeared to be the reverse: announce the benefits, retain control of the operational detail and provide clarity only when outside pressure makes opacity more costly than disclosure.

The F-35 Carve-Out: A Risk Acknowledged, a Supply Route Preserved

Lammy’s September 2024 arms-export decision provides the clearest example of government acknowledging a grave risk while preserving the mechanism considered strategically indispensable. Addressing Parliament, he said the assessment had left him unable to reach any conclusion other than that, for certain exports to Israel, there was a clear risk that the items might be used to commit or facilitate serious violations of international humanitarian law. The government suspended 29 licences at the time—often rounded in political discussion to approximately 30—from a total of around 350.

The suspension was not an embargo. Most licences remained in force, and exports into the multinational F-35 programme received a specific exception, apart from components going directly to Israel for use there. The formal notice to exporters recorded that carve-out. Ministers argued that the United Kingdom could not halt its contribution without disrupting the global programme and producing extraordinarily serious consequences for British and allied security.

The uncomfortable fact is that the exception was preserved despite the risk assessment, not because the risk had disappeared. The government’s position was that it could not identify and segregate every British component eventually destined for an Israeli aircraft once parts entered the international spares pool. In a parliamentary answer, ministers acknowledged that the United Kingdom could not track or control individual components after their entry into the global programme; preventing any British-made part from reaching Israel would require suspending exports to the entire programme.

That is not a minor administrative footnote. The United Kingdom contributes approximately 15 per cent of the value of each F-35 aircraft. The policy therefore accepted that British components could reach Israel through an international pool while maintaining that direct exports for Israeli F-35 use were suspended. The route remained open because closing it was judged to impose an unacceptable strategic cost.

By 28 February 2026, the government reported 395 non-suspended extant licences involving Israel as a destination or potential destination: 203 categorised as military and 192 as non-military. The official release cautions against treating every licence as equipment for Israeli combat operations. Eighty-nine of the military licences supported re-export to third countries, 37 concerned testing, research, demonstration or manufacturing, nine involved academic or civilian purposes, and 68 had the Israeli government as an end-user or potential end-user, including missile-defence, submarine, trainer-aircraft and global F-35 programme items. The document also records that six further licences were suspended or amended after the initial 29.

The distinction matters because a dramatic total stripped of categories would mislead. Yet the official breakdown does not erase the central policy fact. The government identified a clear risk, suspended a defined group of licences and maintained the F-35 pathway because the supply chain could not be selectively controlled without broader consequences.

The High Court dismissed Al-Haq’s legal challenge in June 2025, and in November the Court of Appeal refused permission to appeal. The litigation therefore did not establish that Lammy’s decision was unlawful. Politically, however, the judgments leave responsibility where ministers said it belonged: with the executive, accountable to Parliament and ultimately the electorate. That makes transparency about licences, routes, end users and the limits of government control more—not less—important.

Judge-Only Courts Are Proposed, Not Yet Law

The original version of this investigation risked presenting Lammy’s criminal-court reforms as an accomplished transfer from juries to judges. The current parliamentary record requires more careful language. The Courts and Tribunals Bill is active, but as of 20 July 2026 it had been reintroduced at report stage in the House of Commons and had not received Royal Assent.

The proposal remains constitutionally significant. Under the government’s Courts and Tribunals Bill factsheet, defendants would lose the existing right to elect Crown Court trial for either-way offences. Cases assessed as likely to attract a sentence of three years or less could be allocated to a new Bench Division and tried by a judge alone. Lengthy or technically complex fraud cases could also be heard without a jury. The government says the reforms would reserve juries for the most serious offences and cut delay in a system burdened by a historic backlog.

The backlog is real. Victims wait years, witnesses disengage, evidence weakens and defendants remain under unresolved accusation. Refusing all reform would not preserve justice; in many cases it would preserve paralysis. Lammy is entitled to argue that a fair trial does not invariably require twelve jurors and that professional judges can decide cases lawfully and efficiently.

But the constitutional trade-off must not be buried beneath operational language. A likely sentence of up to three years is not trivial. It represents the loss of liberty, employment, family life and reputation. The state allowed court capacity to deteriorate over many years, then proposed reducing direct citizen participation as part of the cure. Whether that trade is justified should be debated as a transfer of adjudicative power, not marketed as a scheduling adjustment.

The nationality-data dispute intensifies the concern. The same department seeking greater institutional discretion over how cases are allocated is reported to be appealing a regulator’s decision concerning the data by which the public might scrutinise the outcomes. The problem is not that judges are inherently untrustworthy. It is that systems concentrating decision-making require stronger external visibility, not weaker access to evidence.

AI Is Already Scaling Across Justice—But It Is Not Yet an Algorithmic Judge

The language surrounding artificial intelligence is another area where exaggeration would weaken the investigation. The Ministry is not currently claiming that an algorithm will determine guilt or sentence offenders. Its published projects centre principally on transcription, summarisation, search, legal assistance and court administration. Those distinctions must be retained.

The scale is nevertheless substantial. In June 2026, the Ministry announced that every probation officer in England and Wales had been equipped with Justice Transcribe, a system that records, transcribes and summarises conversations. The government estimated that the tool could free the equivalent of 18,750 calendar days of staff time each year. It also announced AI legal assistants, tools intended to identify trial-ready cases and assist listings, and a similar transcription system being trialled within immigration and asylum tribunals.

Justice Transcribe is not, on the evidence published, an automated sentencing engine. Its purpose is to create records that human professionals review and use. The Ministry has published a transparency record for the service, and its Justice AI Unit presents the system as a means of reducing note-taking and improving consistency. That is a more defensible description than claiming that artificial intelligence is already deciding who goes to prison.

The accountability issue lies in how apparently administrative tools become embedded within consequential workflows. A transcript becomes part of a case record. A summary can determine which facts a later officer sees first. A listing tool can influence how long somebody waits for trial or remains on remand. Semantic search can shape which guidance, precedents or risk information is surfaced. Errors need not be final decisions to alter the path by which final decisions are reached.

Government policy already recognises this danger. The Algorithmic Transparency Recording Standard is mandatory for central-government tools that significantly influence decisions with public effect or interact directly with the public. The standard requires information about purpose, data, integration, human review, risks and accountability. The central question is whether those records will remain complete, current and technically meaningful as justice tools move from pilot to scale.

This returns the investigation to the original paradox. A Ministry preparing to govern ever more of its work through data-rich systems is reported to be resisting a decision requiring a comparatively simple statistical disclosure. The state wants increasingly sophisticated tools to process the citizen. The citizen must retain an equally strong ability to inspect the state.

The Chagos Agreement Had Not Transferred Sovereignty

Material correction: Earlier wording stated that the Chagos agreement had transferred sovereignty to Mauritius. That was inaccurate as of 20 July 2026. The treaty was signed, but its own implementation depended upon entry into force and changes to UK law. The government bill intended to make those changes lapsed when the 2024–26 parliamentary session ended.

Lammy was Foreign Secretary when the United Kingdom and Mauritius signed the Chagos Archipelago agreement on 22 May 2025. Under its terms, Mauritius would become sovereign over the archipelago after the treaty entered into force, while the United Kingdom would retain long-term authority to operate the strategically critical Diego Garcia base through a 99-year arrangement.

The government presented the agreement as the only viable means of protecting the base from mounting legal uncertainty. It valued the financial commitment at an average of £101 million a year, with a £3.4 billion net present value, and argued that inaction could eventually make the facility inoperable. Those are the government’s stated strategic and financial assessments, not independently proven inevitabilities.

Crucially, Parliament’s own update states that the Diego Garcia Military Base and British Indian Ocean Territory Bill would make no further progress after the session ended on 29 April 2026. Government guidance likewise says the treaty’s sovereignty changes would occur only after entry into force and the necessary domestic legislation. The legal and political position as of publication was therefore not that Britain had already completed the transfer, but that the government had signed a treaty whose implementation had not been completed through the lapsed bill.

The correction does not remove the controversy. It changes its nature. The public is entitled to know whether and how the government intends to revive implementation, what happens to the treaty while the necessary legislation is absent, how costs are calculated, what termination provisions mean in practice and whether future parliamentary approval will reopen issues treated as settled in ministerial announcements.

The process also attracted sustained criticism from Chagossian representatives. A House of Lords committee concluded that the community had a moral right to be heard and that the government should have engaged more meaningfully, while acknowledging the practical difficulty of identifying one unified Chagossian position. The legitimate security needs of Diego Garcia do not cancel the democratic obligation to distinguish confidential military detail from information about cost, sovereignty, consultation and legal structure.

The Pattern Is Selective Transparency, Not a Proven Criminal Conspiracy

No responsible investigation should manufacture a criminal conspiracy from a sequence of political controversies. There is no evidence presented here that David Lammy accepted a bribe in exchange for a ministerial act. The Standards Commissioner found no deliberate intention to conceal the late registrations. A threatened judicial review did not prove that Shell or BAE Systems captured Foreign Office policy. The courts did not find the F-35 decision unlawful. The judge-only court proposal remains a bill rather than law, and the Ministry’s published AI tools are not algorithmic judges.

Those limitations do not exonerate the wider record. They define the accusation accurately.

Lammy’s own review argued that the justice system should publish demographic data and expose decision-making to outside scrutiny. His Ministry then resisted a request for aggregated nationality conviction figures, lost on its principal exemptions before the Information Commissioner and is reported to be appealing. Lammy has an official record of sixteen late registrable interests after a previous late-registration inquiry. As Foreign Secretary, he promoted a deeper exchange of personnel between government and powerful corporations while campaigners had to demand the governing safeguards. He acknowledged a clear humanitarian-law risk in certain exports to Israel while preserving the multinational F-35 route because the government said it could not selectively control the destination of British components without imperilling the wider programme. He is sponsoring a bill that would move a category of cases from juries to judges, while his department scales AI systems across probation, courts and tribunals.

These events do not demonstrate one secret scheme. They reveal one persistent distribution of power. Government collects the information. Officials decide what the public can inspect. Ministers receive the confidential assessments. Corporations may be brought closer to institutional decision-makers. Judges are offered greater control over verdicts. Artificial intelligence becomes part of the record-making and case-management infrastructure. Long-term strategic commitments are announced through conclusions the public cannot fully reconstruct from the underlying advice.

The public is permitted to hear the verdict of government.

It is increasingly forced to fight for the evidence.

Publish the Evidence

The immediate remedy is neither radical nor complicated. The Ministry should publish the requested conviction tables in the form contemplated by the Information Commissioner: suppress figures below five, disclose the larger aggregates and accompany them with a methodological document substantial enough to prevent misuse.

That document should define nationality, explain how missing and dual nationality are handled, distinguish conviction counts from unique individuals, identify the period covered and explain the offence groupings. It should state clearly that nationality is not ethnicity, immigration status, asylum status, country of birth or evidence of causation. Where robust population denominators do not exist, it should say so rather than offering false precision.

The Ministry should also publish the original request, its full refusal reasoning, the appeal grounds if an appeal has been lodged, the authorisation route and the estimated cost. If another lawful exemption is now relied upon, it should be identified and defended in public rather than introduced through an opaque procedural escalation.

Publication must not become permission for collective blame. A nationality is not a criminal category. A prison population is not an offending rate. A conviction statistic cannot explain the social, demographic and legal factors beneath it. But those truths do not justify withholding evidence. They justify publishing it honestly.

Lammy’s own review established the principle: institutions earn trust by making their decisions visible enough to be challenged. That standard cannot apply only when transparency reveals disparities the minister already expects to find. It must apply when the evidence may disrupt the government’s preferred account of crime, migration and institutional competence.

End the secrecy. Publish the figures. Release the methodology. Let independent researchers test the claims advanced by every political faction.

Britain does not need another minister instructing the public what the evidence supposedly means while the state controls access to the evidence itself.

It needs the record in full.

Thanks for reading! This post is public so feel free to share it.

Sources

Primary and official sources were prioritised. The reported Ministry of Justice appeal is sourced to The Telegraph because no public appeal filing was located at the time of verification.

Thom Aster

Discussion about this post

Ready for more?