She had a saying. If she wished anyone harm, she would wish them a lot of money.

My mother repeated it often and always attributed it to Jane. I was very small when Jane died.

This week the country was told again that its problem is the cost of living. John Healey delivered his first Budget as Chancellor on 28 October. The Treasury’s own Budget document has a chapter titled “Breathing space”.

A Guardian conversation between the economists Karel Williams and Sukhdev Johal and Aditya Chakrabortty asks a different question. Its description says governments have tried spending more, spending less, cutting red tape, leaving the European Union and levelling up. None has raised growth per person.

I read “cost of living” as manufactured language. It names the symptom. It distracts from the real needs.

What the Budget bought

The Budget offers a £100 payment towards energy bills for up to 20 million households in England, in council tax bands A to D, from January 2027. It will abolish gas standing charges from 2027. The Treasury says 60 per cent of households will save an average of £45 a year. It temporarily cuts tariffs on more than 100 food and drink products, worth up to £21 a year on groceries. It freezes the prescription charge at £9.90. It raises the National Living Wage by 3.9 per cent from April 2027.

A household that gets the payment, the standing charge saving and the grocery saving gains about £166 a year on those figures. That is about £3.20 a week.

The Office for Budget Responsibility (OBR) is the official forecaster. It expects debt interest to be £14 billion a year higher, on average, than it forecast in March. That is money the government cannot spend on anything else. It expects real household disposable income per person, which is income after tax and adjusted for prices, to fall by an average of 0.4 per cent a year across 2026 and 2027.

The OBR forecasts growth of 1.2 per cent in 2026 and 1.0 per cent in 2027. It then forecasts 1.6 per cent in 2028, 1.8 per cent in 2029, 1.7 per cent in 2030 and 1.6 per cent in 2031. At 1 per cent a year an economy’s output takes about 70 years to double. At 1.6 per cent it takes about 44.

The longer record is in the official documents. The Treasury says real household disposable income per person has grown by around 0.8 per cent a year since 2010. Before the financial crisis the long-run average was around 2.5 per cent. The OBR says growth in that measure hit a historic low over the ten years to 2025-26, averaging a quarter of one per cent a year. It expects the level in 2029-30 to be 1 per cent above its pre-pandemic level.

Conservative, coalition and Labour governments all promised growth over those years.

When justice and fairness were the aim

When I was younger the relentless pursuit of growth was not a thing. The aim and focus appeared to be justice and fairness. That is my memory. It is not a statistic.

Then came the rise of the ultra rich. Anyone who asked how the rich got their money was accused of being a communist.

Many of the same people professed belief in God. They paid lip service to living a simple life. The teaching they professed says no one can serve both God and money (Matthew 6:24). It tells of a rich man who built bigger barns and was asked that night whose they would be (Luke 12).

The saint and the communist

Dom Hélder Câmara was archbishop of Olinda and Recife, in north-east Brazil, from 1964 to 1985. He is widely quoted as saying: “When I give food to the poor, they call me a saint. When I ask why the poor have no food, they call me a communist.” The wording varies between sources.

A £100 payment is a food parcel. So is a prescription freeze. So is a tariff cut on baked beans. Nobody who has been hungry would refuse one.

The second question is the one that draws the name. A standing charge is income for the supplier and a cost for the household. It is one payment seen from two ends. Why is it due before a single unit of gas is used? Why does the income left after essentials keep shrinking?

Williams, Johal and three colleagues set out their answer in When Nothing Works (2023). They argue the cost of living crisis is the visible part of a deeper failure of what they call foundational liveability. Incomes left after essentials are squeezed. Public services are failing. Social infrastructure is decaying. Higher wages and growth, they say, are not enough on their own.

Growth for whom

Growth per person is an average. Take a street of ten people. Nine earn £20,000. One earns £200,000. The average income is £38,000.

Let the economy grow by 3 per cent and let all of it go to the one at the top. The average rises to £39,140. The median, the person in the middle of the street, still earns £20,000.

The street is invented. The question it asks is real: growth for whom.

Bill Jordan, the social policy writer, published Automatic Poverty in 1981. Its first part is headed “The Problem of Economic Growth”. The publisher summarises the argument: Britain’s decline reflects an advanced stage of industrialisation, with production increasingly mechanised and output static.

Olivier De Schutter, the United Nations Special Rapporteur on extreme poverty and human rights, published The Poverty of Growth in 2024, with a foreword by Kate Raworth. He argues that growth often deepens exclusion and inequality rather than curing poverty. I had the phrase in my head before I found its author.

The Budget freezes income tax thresholds until 2031. The personal allowance stays at £12,570. A pay rise pulls more of a worker’s pay into tax. The new tax falls on income from savings and property. Who holds the savings and the property is the second question again.

What the saying says

The Book of Proverbs asks for neither poverty nor riches.

Jane McNiff knew the first. If she wished anyone harm, she would wish them the second.

The Budget offers a household up to about £3.20 a week.

Sixty years of being sorry

On 21 October 2026 the King, the Queen and the Prince of Wales stood in Aberfan and heard, in the King’s words, “stories of heartbreak, but also of heroism”. A few days earlier the Prime Minister had laid a wreath in the cemetery and said what the system has been slow to say: “I don’t think Aberfan was treated properly by the authorities.”

He is right, and the record is plain. A coal tip slid onto Pantglas Junior School in 1966 and killed 144 people, 116 of them children. The Tribunal of Inquiry found that the National Coal Board was to blame and that the disaster “could and should have been prevented”. Nine of its employees were censured. None was prosecuted. None, as far as the reports tell us, lost their job. Two years later the government made the Aberfan Disaster Fund hand over £150,000 to clear tips that belonged to the Coal Board. The money was repaid in 1997, with no allowance for the nearly thirty years in between.

Sixty years on, Wales still has some 2,600 disused coal tips, and the slides at Tylorstown in 2020 and Cwmtillery in 2024 have shown what that means for the people living beneath them. The UK Government is now providing around £140 million. That is welcome. It is also, to anyone who has watched the sequence before, the next step in a very familiar procedure.

That procedure is the subject of this piece.

What a seven-year-old kept

That is all I have of that day, and it is mostly my father’s face. The next thing I remember clearly is the ledge.

One of my last memories of my father, Edward McNiff, as a well man is of him on the television programme Nationwide, in February 1970. As I recall it, the filming fell on the day I went into hospital for surgery on my head, related to my ears, and he very nearly missed visiting time. We lived in Keresley, in Coventry, beside the Coal Board’s smokeless fuel plant, which we knew as the Homefire plant. Smokeless fuel was a Coal Board research programme in which Jacob Bronowski, the mathematician who later wrote The Ascent of Man, was a leading figure in the 1950s. My father’s complaint was that its output was filling the doctors’ surgeries with people with lung complaints. On camera he drew a finger along a window ledge and held up the dirt, reportedly soot fallout from the plant.

He was not the first to say it. On 20 June 1967, almost three years earlier, Mr Edelman had asked the Commons whether the minister was aware of “the hazards to health and the damage to amenities caused by the Homefire Plant at Keresley”, and whether a plant built to remove smoke should itself be fouling the atmosphere. He called it a plant that covered “the neighbourhood with a pall of black, soggy fall-out”. The minister, Robert Mellish, answered: “I much regret the conditions caused by this plant at present.” He added: “I can only apologise on behalf of those who operate the plant; I am not here to defend it.” The regret was recorded, the apology was offered, and the soot was still on the ledge in 1970. Fifty-one years after my father’s film, in 2021, the plant’s former owners and operators admitted in compensation claims that their own workers had been exposed to harmful dust, fumes and carcinogenic materials up to 1995, and that they had breached their duties to them. Over fifty former employees have been awarded compensation, according to ITV News. If the people inside the fence were being harmed, it is not a leap to wonder about the families on the other side of it.

He died of a lung-related illness in 1973. I will not claim a line between the ledge and the lungs. I am not able to prove one, and the point of this piece is that nobody was ever asked to try.

Shortly before he died came Lofthouse, near Wakefield, on 21 March 1973. Seven men were trapped 750 feet down when a new coalface broke into an old flooded shaft. My Uncle Tony, married to my father’s eldest sister Mary, worked at Lofthouse. I have never asked my cousins where he was that day, and I cannot tell you. The adults around me were angry. What I do not remember, because it did not happen where a child could see it, is anything being done for the families.

What the system is built to do

A system that meets a disaster does a small number of things, and it does them in the same order every time. It holds an inquiry. The inquiry finds failure and names institutions. The government accepts the findings and promises that lessons will be learned. A bill or a set of regulations follows, usually late. The families are offered a sum, and the sum is argued over. Then the story is closed, with a commemoration.

What the sequence does not contain is a step where a named person pays a price. Aberfan is the cleanest example in British history because the findings were so unambiguous. The Tribunal said blame rested on the National Coal Board. It said the disaster “could and should have been prevented”. Lord Robens, the Board’s chairman, offered to resign in 1967 and was told by the minister to stay. The Board’s first offer to each bereaved family was £50. It rose to £500. The Disaster Fund, built from the donations of strangers, was then used to pay for work the Board should have done.

Lofthouse seven years later runs the same script on a smaller stage. A new coalface was cut towards an old flooded shaft that the Board’s own surveyors knew about and did not believe reached the depth of the modern workings. British Geological Survey records showed that it did. They were not checked. Seven men died, and the first rule requiring colliery managers to make “diligent enquiry into other sources of information” before opening new workings arrived in the Mines (Precautions Against Inrushes) Regulations 1979, six years after the water came in. I remembered that nothing changed. The truer and worse reading is that something changed, and that it took six years.

The families’ story ends the way Aberfan’s money did. The public gave £750,000 to a fund for the Lofthouse families. By 2024 every beneficiary had died, and the trustees told the BBC they were struggling to give away the last £1 million because of restrictions on who could benefit. Their chairman feared it would “just end up going back to the government”. The charity of strangers outlived the people it was for, and the state was waiting at the end of it.

There is a number worth sitting with. The harm happens in a day. The learning happens over years. The people who do the learning are not the people who were hurt, and the people who were hurt are mostly gone by the time the learning arrives. A system that accepts that gap as normal is not failing to protect them. It is working as designed, and the design assumes that the cost of delay is carried by someone else.

Who carries the cost

Follow the sequence past the inquiry and the same shape appears in the lives of the people who stayed. Aberfan was the Coal Board’s disaster. Lofthouse was the Coal Board’s disaster. In 1984 and 1985 the state met the same communities in a different role, and the strike that ended in March 1985 closed pits and removed far more than wages. The Social Security Act 1980, and the regulations made under it, had already changed the rules for strikers’ families. The Department assumed that striking miners were receiving £15 a week from the union, a figure that rose to £16 by the winter of 1984, and deducted it from their families’ benefit. By then the union money had run out. In the places I know, it removed the working life around which the whole street was arranged. Many of those communities still carry the effects.

One case shows what that meant at the graveside. Dave Roper, a miner at Silverwood Colliery near Rotherham, lost his newborn son Adam in the summer of 1984. He applied for a funeral grant and was told that because he was on strike he was not entitled to it. He had no money. An undertaker offered to ask a bereaved family whether the baby could be placed in a dead relative’s coffin, and a free funeral was arranged. Dave and his wife did not go. “I’m going to somebody’s funeral and they’ll think ‘Why the bloody hell is he here?’” he told the BBC forty years later. The Department of Health and Social Security reversed the refusal in October 1984 after growing pressure. Adam had already been buried in a stranger’s coffin.

The pit that stood near us in Keresley, Coventry Colliery, closed in October 1991. The Homefire plant on its site, the one my father had held up a finger of soot against, did not close until 2000 and was demolished in September of that year. The plant that gave my father his complaint outlasted the pit it was built on by nine years. I note that because it is the whole argument in a single fact: the work left the community first, and the plant that had been the subject of complaints stayed on after it.

That is the pattern I see across these places. The communities supplied the coal, and the coal supplied the economy. When the economy no longer needed the coal, the communities were given what the system gives anyone it is finished with: a review, a regeneration plan, and a promise that this time will be different. I cannot give you a single figure for what was invested compared with what was extracted. I can say that none of the people I grew up among would recognise the investment as proportionate, and that the burden of proving it was proportionate has never fallen on the people who made the decision.

The same logic is now visible in the tip safety money. The £140 million is real. So is the number of tips, around 2,600. So are the landslips at Tylorstown in 2020 and Cwmtillery in 2024, which are the reason the money is being found at all.

A debt that is never called in

The oldest moral arithmetic we have says what a wrong costs. In the law of Moses, a person who took what was not theirs gave it back in full and added a fifth (Leviticus 6). When Zacchaeus, the tax collector, was confronted, he promised to give back fourfold (Luke 19). The principle is that restitution is not the return of the sum. It is the return of the sum plus an admission that time, trust and loss have a price.

Set that against £150,000 taken from the Aberfan fund in 1968 and returned in 1997 with no interest. The sum was handed back at the same number, nearly thirty years later, as though the years had been free. It took until 2007 for the Welsh Government to give £1.5 million to the memorial charity and £500,000 to the education charity, as recompense for the money wrongly taken. That is thirty-nine years to find a figure that arithmetic could have supplied in an afternoon.

There is a philosophical point inside the figure. A system with no memory of its debts will repay them only when someone embarrasses it. That is not because the people inside it are bad. It is because the thing that would call the debt in, a person with a name and the power to be held responsible, has been removed from the structure. Hannah Arendt called this rule by Nobody. The Board is censured, the chairman stays, and the families are offered £500.

The ledge

This month Mair Morgan, the last surviving teacher of Pantglas Junior School, a woman who saw it happen, met the Prime Minister, who said that Aberfan was not treated properly. She has waited sixty years for a sentence that could have been said in 1967.

My father, Edward McNiff, held up a finger of soot on a television programme a few years before he died. I do not know what that dirt cost him, and neither does anyone else. He died in 1973, and the plant that made the dirt outlived the pit it stood on by nine years.

That is what I mean when I say the system does not change. Not that nothing is ever done. A great deal is done: tribunals, reports, regulations, wreaths, millions. What does not change is who pays for the delay between the harm and the fix, and who is never asked to. It is the same people each time. They live downhill, they live downwind, and they are the last to be told.

So when the next inquiry reports, and one will, do not ask what it recommends. Ask three things. Who was named. Who paid. Who is still waiting. If you can answer the third question with a street, you already know how the story ends.

On 20 October 2026, six years and five months later, the Chair of the UK Covid-19 Inquiry, Baroness Hallett, found the claim unfounded. “Quite the opposite was true,” she said. Her words for the care sector in the early weeks were “something of an afterthought”.

That is the finding. What interests me is what happens to it next, because the sequel is already visible. There was no ring around the care homes. There is about to be one around the inquiry, made not of lies but of dates, procedures and a vocabulary, and it can be mapped in advance.

Jean Adamson

Start with one of the people the process will answer to. Jean Adamson speaks for Covid-19 Bereaved Families for Justice UK. Her father, Aldrick, a Windrush pioneer, was 98 when he died in April 2020 after catching Covid in his care home. She says that discharging hospital patients into care homes without testing was a death sentence for family members. She is one of the families who have waited six years and six months for a public finding.

His age matters to the argument. The report called unacceptable resuscitation decisions based on age or disability, or made without a personal assessment, and at 98 that assumption falls hardest.

The coverage does not say whether his home took in discharged patients, so nothing here claims that is how the virus reached him. The record shows the month. No policy required patients to be tested or isolated before discharge until mid-April 2020, and he died in April.

Other families have read the wording too. James Telfer lost his mother, Jacqui, in hospital in January 2021. He welcomed the report, then said it does not go as far as concrete, measurable recommendations, and he said so before any government had replied.

What the report found

Baroness Hallett’s report counts more than 43,000 care home deaths involving Covid between March 2020 and July 2022. Her explanation is plain. A 2016 pandemic exercise had named social care a particular concern, but its recommendations were unfinished when the virus arrived, and in March 2020 all four governments looked first to the hospitals. About 25,000 patients went from hospital to care homes, most of them untested. Hospitals were served first for protective equipment, and care workers improvised with festival ponchos, painters’ overalls and bin bags. Resuscitation decisions based on age or disability, or made without a personal assessment, were called unacceptable. Nobody kept a central record of how many care workers died. Visiting limits were necessary at first, but their harms came to outweigh their benefits, and some people were not permitted to be with a dying relative.

The report traces all of this to one root: there was no proper pandemic planning for adult social care. It makes eight recommendations. They include better pandemic planning, financial support so that care staff can afford to self-isolate, guidance on visiting that starts from residents’ right to in-person visitors, mandatory end-of-life training and a voluntary register of care workers.

The bereaved did not speak with one voice. Jean Adamson called the report “a devastating account” and Hancock’s claim “a cruel lie”. Nicola Brook, the solicitor for more than 7,000 of the group’s families, called it a “superficial damp squib”. Matt Hancock is not commenting on it.

Why the hospitals came first

The Chair found that the slogan of protecting the National Health Service sometimes came at the expense of the people who used and provided care. There are two ways to read that.

The mathematical reading is Goodhart’s law: when a measure becomes a target, it stops being a good measure. Hospital capacity became the target. Discharging patients to untested care homes served that number, and the number was met. The residents were the cost.

The theological reading comes from Mark 7:9 to 13. A religious formula, a gift declared devoted to God, released people from the duty to support their own father and mother, and the verdict is “making void the word of God by your tradition” (English Standard Version). Protecting the NHS was necessary, and it was true. It was also the sentence under which a duty dissolved. That is a reading of the pattern, not a claim about anyone’s motive.

A hospital can be protected and a resident abandoned in the same sentence.

The mathematics of a timetable

The pandemic had no timetable. The process that follows it has several, and they are worth setting side by side.

The claim at the podium stood for 77 months before it was found unfounded. The Inquiry asks each government to respond within six months, but the clock is slower than it sounds. The healthcare report (Module 3) was published on 19 March 2026, the UK government’s response is dated September 2026, and its first progress update is promised for May 2027, fourteen months after the report. The longer precedent is Grenfell. The fire was on 14 June 2017, and the government expects to implement all 61 recommendations from the Grenfell Inquiry’s two reports by the end of 2029, twelve and a half years on. By its September 2026 progress report, 30 were complete and 31 were in progress.

The care sector’s own reforms keep the same slow rhythm. The first Fair Pay Agreement for care workers, backed by £500 million, is expected in April 2028, eight years after the first wave, with negotiations starting in April 2027. The Casey Commission is still working on the shape of a National Care Service. Meanwhile the government has spent more than £122 million responding to the Covid Inquiry, on legal advice and staff preparing evidence, separate from the estimated £200 million cost of the Inquiry itself. That is the cost of answering an inquiry, which is not the same as the cost of acting on one.

What governments do with a report

Governments have a settled way of receiving a report like this, and the steps follow one another so naturally that nobody has to plan them. First comes reception. The Prime Minister said the government would consider the findings carefully and respond in due course, and the Health Secretary agreed that adult social care had been undervalued for too long. Then comes acceptance: the government says it is acting on all 61 Grenfell recommendations. Then comes description. In a letter on the government’s response to an earlier report, the Chair wrote that it used “opaque language that conceals a lack of concrete action”. The findings are then routed into programmes that already exist, each with its own dates, and progress is reported on a schedule the government sets, with the first Module 3 update promised for May 2027. Last comes the limit. An inquiry cannot determine anyone’s civil or criminal liability, so the finding that the claim was unfounded carries no consequence for the person who made it.

None of this needs bad faith. A public dashboard is scrutiny, and it is how Grenfell’s 30 completed recommendations can be counted at all. The UK government’s response to the healthcare report sets out actions on every recommendation and promises a progress update in May 2027. The criticism is narrower. Each step is reasonable alone, but together they form a ring, because a process that reports on its own schedule fixes its own dates, and the next virus does not choose them.

What comes next

What follows is prediction, built from those dates. Expect the acceptance language to continue until 19 November, when the Inquiry’s report on Test, Trace and Isolate (Module 7) takes the news cycle. The formal response to the eight recommendations should arrive around April 2027 on the six-month clock. Module 3 shows its likely shape. There the UK government broadly agreed with the Chair and set out actions on every recommendation, yet said it did not intend to introduce one standard process at this time, and that a plan on another recommendation would take longer than the six months requested. Expect agreement on all eight, with the hardest carved out or deferred.

The order of delivery will follow cost. Administrative items come first: the voluntary register of care workers and guidance on visiting. The costly ones come last or conditional: funded isolation pay, mandatory end-of-life training and pandemic planning for care. Visiting is the one to watch. The report wants guidance built on residents’ right to in-person visitors except in the most grave or exceptional circumstances. The question is whether government answers with guidance that ministers can rewrite, or with a statutory right.

The workforce findings will travel inside the Fair Pay Agreement and the Casey Commission, so a recommendation that must be in place before the next pandemic will move at the speed of programmes built for pay and recruitment. The first public progress update, on the Module 3 pattern, should follow the response by about eight months, which points to December 2027, fourteen months after the report. For the claim itself there will be no formal consequence. The only sanction is the finding.

Three things separate a response from a ring: a statutory duty, a funded date and a named accountable officer. The form exists. The question is whether each government will use it.

The reply

Aldrick Adamson, a Windrush pioneer, was 98 when he died in April 2020. Weeks later the country was told that a ring had been thrown around people like him. It has taken six years and five months to say that was untrue.

The reply is due by April 2027, seven years after his death, and the pattern says what it will look like: accepted in principle, routed into other programmes, reviewed on the government’s own schedule. Jean Adamson called the discharge policy a death sentence. The test is not whether the reply describes the failure. It is whether it names a duty, funds a date and puts one person’s name against it.

Until it does, there is still no ring around the care homes. There is one around the inquiry.

Systems on Trial examines the machinery beneath the narrative.
ianmcniff.substack.com

This piece argues that the missing person is the story. It uses the numbers, the law and the scriptures that every side cites, and asks what each looks like when a human being is put back inside it. A debate that strips the person out of the description does not become more rational. It becomes more extreme, because abstractions are all that remain to fear or to pity.

The word does the work

Military-aged. The phrase makes a man a combatant before anyone has asked him anything. On 25 August 2026, in a report of his remarks, Farage said “the only people who make it on the boats are young fit males” and spoke of “perhaps even an army within this country that is against us”. The Migration Observatory’s figures say otherwise: about 12 percent of small-boat arrivals in 2025 were children. The United Nations High Commissioner for Refugees (UNHCR) guidelines on military service add a second answer. Disproportionate or arbitrary punishment for refusing state military service can itself be persecution (paragraph 14). Forced recruitment of a child under 18 is persecution (paragraph 12). A man of military age may be exactly the man his state wants to conscript. Whether a particular man is such a case is a question for a decision-maker. The label answers it in advance.

Invasion. On 31 October 2022 Suella Braverman, then Home Secretary, described the arrivals on the south coast as an “invasion”. An invasion has an enemy. The five nationalities that made up two thirds of small-boat arrivals in the year to June 2026 were Eritrean, Sudanese, Afghan, Somali and Iranian. Under the word “invasion” they become an army.

Illegal. The Home Office counts small-boat arrivals within “detected illegal entry arrivals”: 88 percent of 38,000 in the year to June 2026. UNHCR’s position is that it is “not illegal to seek asylum”, and it prefers “irregular” to “illegal” for how people travel. Yet “illegal immigration” and “refugee” now run together in headlines as if they were one category. Whether that was designed or allowed, it works the same way: a lawful claim arrives pre-labelled as a crime.

Return. Return is a logistics word. It has no one in it. A person sent back is a person put in a place, and the word hides what the place is.

The mathematics of this is simple. A percentage keeps the numerator and forgets the denominator is made of people. “Seventy-six percent are men” is true. It is not an argument for anything until someone says who the men are.

What the law asks

The case against boat arrivals often rests on the rule of law. That deserves a straight answer from the text.

Article 31(1) of the 1951 Refugee Convention says states “shall not impose penalties, on account of their illegal entry or presence” on refugees who enter without authorisation, provided they present themselves without delay to the authorities and show good cause. The Convention anticipates unauthorised entry. It does not require a visa, a queue or a legal route as a condition of protection. Article 33(1) forbids the opposite act: no state shall “expel or return (‘refouler’)” a refugee “in any manner whatsoever” to a territory where life or freedom would be threatened.

In the Convention, then, return is the thing that is prohibited. Article 31 has conditions, and states argue about how far they reach, in particular the words “coming directly”. But nothing in the text makes a refugee a lawbreaker for the manner of arrival. A speaker who invokes the rule of law against the boat should say which rule.

Who is in the boat

Numbers first, because the debate prefers them. The Migration Observatory reports that about 76 percent of small-boat arrivals in 2025, where recorded, were men over 18 and about 12 percent were children. It gives one reason for the male majority: the danger of unauthorised travel. Women and children sometimes follow later through family reunion. The male majority is the shape of a dangerous journey. It is not a finding about intent.

The same briefing reports that about 62 percent of initial decisions on small-boat arrivals from 2018 to 2025 were grants of protection, against 52 percent for all claims over the same period. Most of the people decided on were found to be refugees. The overall grant rate has since fallen. Home Office figures give 38 percent for the year to June 2026, down from 48 percent.

Then the journey itself. In September 2017 the International Organization for Migration (IOM) and UNICEF published Harrowing Journeys, drawing on about 22,000 interviews. It found that 77 percent of children and youth on the Central Mediterranean route reported direct experiences of abuse, exploitation or practices that may amount to human trafficking. In March 2023 the UN fact-finding mission on Libya reported “reasonable grounds to believe that migrants across Libya are victims of crimes against humanity”. It linked murder, enforced disappearance, torture, enslavement and sexual violence to arbitrary detention, and documented “enslavement, sexual slavery”. One investigator, Chaloka Beyani, said the European support given “has aided and abetted the commission of the crimes”. The mission did not find the EU or its member states responsible for war crimes.

After the crossing, asylum seekers in the UK are generally barred from working while their claims are decided. In November 2025 Lucila Granada, chief executive of the labour-exploitation charity FLEX, said people seeking safety are left with “the choice of exploitation or destitution”.

Sexual violence, the abuse of children and enslavement are documented on these routes. The findings come from United Nations bodies. They are almost absent from a debate that speaks mostly of totals and costs. A debate that held them in view could not describe the people in the boats only as a threat.

Thirty months of leave, and what the clock counts

From 2 March 2026 the United Kingdom (UK) began granting adults and accompanied children who claim asylum 30 months of leave instead of five years. On 25 September 2026 the United States (US) President set the refugee ceiling for fiscal year 2027 at 17,500, with places allocated primarily to Afrikaners from South Africa. On 1 October 2026 federal Medicaid eligibility ended for most refugees and asylees in the US who do not yet hold a green card. In North Carolina alone about 14,000 refugees were due to lose cover.

The case for a review cycle rests on one fact. Refugees do go home. UNHCR reports that nearly 4.4 million refugees returned to their countries of origin in 2025. Syria, Afghanistan and Sudan received 90 percent of them. Forced displacement fell in 2025 for the first time in a decade, to 117.8 million people. One person in 70 is forcibly displaced. Thirty-nine percent are children.

Exile is the other fact. On 4 June 2026 UNHCR reported that refugees in Eastern and Southern Africa remain in exile for a median of almost 16 years. Families of five or more remain nearly 19. Refugees registered before the age of five remain more than 18. The figure is regional and is not a UK figure. It sets the scale.

The arithmetic is short. Sixteen years is 192 months: 6.4 review cycles of 30 months. Nineteen years is 228 months: 7.6 cycles. The Home Office has also announced a 20-year wait for settlement. That is 240 months: eight renewals of 30 months. The Immigration Law Practitioners’ Association describes the settlement part of the model as not yet implemented, so the 20-year figure is a stated plan, not current law.

The Refugee Council estimates the cycle will add up to 1.1 million repeat reviews in the first ten years, at a cost of up to £725 million. At those upper estimates each review costs about £660. The Refugee Council also warns that the Home Office is already struggling with backlogs.

The US ceiling follows similar arithmetic. A ceiling of 17,500 against 117.8 million displaced people is one place for every 6,700 people. Against the 2.5 million refugees UNHCR estimated would need resettlement in 2026, it is 0.7 percent. In 1980 the US admitted 207,116 refugees, the highest annual figure in the Congressional Research Service series from 1980 to 2025. That is nearly twelve times the fiscal year 2027 ceiling.

The determination also allocates places by group. A group allocation and an individual test for a well-founded fear of persecution answer different questions. This piece takes no position on any group’s claim. The point is that the rule has changed, not only the number.

The host is the party on trial

The religious texts reviewed here do not treat the refugee as an exception to be managed. They treat the refugee as a founding figure. They put the host under judgement.

Judaism. Leviticus 19:34 tells Israel to treat the resident stranger “as one born among you”. The reason given is memory: Israel was a stranger in Egypt. Numbers 35 sets aside six cities of refuge for a person who kills without intent. Verse 15 extends them to the stranger and the sojourner. The refugee stays until the death of the high priest, then may return home (35:25 and 35:28). The term ends with an event outside the refugee’s conduct. Nobody re-tests the claim at intervals. That is a reading of the text, not a ruling of any tradition.

Jeremiah 29 is a letter to exiles in Babylon. It tells them to build houses and plant gardens (29:5) and to “seek the peace of the city” where they live (29:7), because their welfare depends on the city’s. The instruction to displaced people is an instruction to contribute. It assumes they will stay.

The Book of Ruth opens with a famine that sends a family from Bethlehem to Moab (1:1). Ruth, a Moabite widow, arrives in Judah with her mother-in-law and becomes the great-grandmother of David (4:17). Matthew places her in the genealogy of Jesus (1:5).

Christianity. Matthew 2:13 to 15 has Joseph take Mary and the child to Egypt after a warning that Herod means to destroy him. Matthew 25:35 sets the standard for judgement: “I was a stranger, and ye took me in”. The Catechism of the Catholic Church (paragraph 2241) says more prosperous nations are obliged, “to the extent they are able,” to welcome the foreigner who seeks security. The same paragraph allows states to set juridical conditions and obliges immigrants to obey the laws of the host country. The tradition holds both. The obligation comes first.

Islam. The Islamic calendar counts years from the Hijra, the migration of Muhammad and his followers from Mecca to Medina in the year 622. Year 1448 began on 16 June 2026. Muslims number their years from a journey made under persecution. Around 615 the first Muslims to flee Mecca were sheltered in Abyssinia by a Christian king, the Najashi, who by the traditional accounts refused Meccan envoys’ demand to expel them. The Qur’an does not name the episode, though tradition connects several passages to it. Qur’an 59:9 praises the Ansar, the helpers of Medina, who love those who emigrated to them and prefer them to themselves even in need. Qur’an 9:6 concerns a person from a hostile group who seeks protection. It orders that protection be granted and that the person be escorted to a place of safety. Commentators read the host’s responsibility as lasting until that safety is reached.

Buddhism. The entry rite is the act of taking refuge in the Buddha, the teaching and the community. The 14th Dalai Lama sought refuge in India in 1959 and received the Nobel Peace Prize in 1989.

In each text the weight sits on the host. The stranger is the standard by which the host is judged. The displaced are told to contribute. The host is told to protect. None of the passages above sets a review interval for protection.

These are the texts that banners now invoke. Humanists UK reported on 8 September 2026 on protests at Dover on a Saturday and Portsmouth the following Sunday. Hundreds of men in black, in balaclavas, blocked the ports, chanting “Christ is King!” alongside “Stop the boats!” The organisers, Patriot Platform, have a founder who describes his aim as building a “Christian patriotic army”. Farage himself criticised the masked, uniformed demonstrators. In September 2025 the presidents of Churches Together in England said “The Cross and the Gospel of Christ must never be co-opted” for hostility toward others. More than thirty church leaders wrote that “the cross is the ultimate sign of sacrifice for the other”. The texts above put the stranger on the side of the cross. Whoever frames the stranger as the enemy has to leave the text behind.

The money argument, and where it stops

The money argument is not the case for protection. Protection is owed on need, and the Convention and the texts above say so without reference to profit. But a debate that speaks only of cost should hear the other figures.

The US Department of Health and Human Services (HHS) published a study in 2024 of the fiscal impact of refugees and asylees from 2005 to 2019. Refugees and asylees paid an estimated $581 billion in taxes. Governments spent an estimated $457.2 billion on them. The net impact was positive: $123.8 billion, of which $31.5 billion went to the federal government and $92.3 billion to state and local governments. Per head, the net impact was comparable to that of the total US population.

The report gives a second figure. Counting spouses and children under 18, most of them US citizens, spending was $723.4 billion and revenue $739.4 billion. The net impact falls to $16.0 billion. Both figures are positive. The difference is the boundary of the count. Schooling for the children falls inside the window, and many of their earning years fall after it. That is a reading of the accounting, not a finding of the study.

The study’s host page now states that the report does not represent the views of HHS or the Trump administration and is “provided only for historical purposes”. It is still published.

The first years cost money. A National Bureau of Economic Research study, as summarised by the Adam Smith Institute, found that the initial fiscal impact of refugees is negative. It also found that adult refugees aged 18 to 45 pay about $21,000 more in taxes than they receive in benefits over their first 20 years. The International Rescue Committee reports a median household income of $30,500 for US refugees in the country five years or less. For those in the country 20 years or more it is $71,400, above the national median of $67,100.

The European evidence is less favourable. Research on Sweden by Ruist, published through the Centre for Economic Policy Research, found refugees’ net fiscal contribution lower than that of other immigrants. It also concluded that the burden of a generous refugee policy is not heavy if refugees enter the labour market quickly.

The money argument turns on one variable: how fast refugees reach work. Short leave, repeated review and withdrawn health and food support all act on that variable. The Refugee Council warns that renewed uncertainty every 30 months damages integration. In the US the One Big Beautiful Bill Act ended Medicaid eligibility for most refugees on 1 October and also ended their eligibility for the Supplemental Nutrition Assistance Program (SNAP), the federal food assistance programme. No source reviewed tests whether removing early support raises or lowers the 20-year figure. The reading offered here is that it lengthens the loss-making phase.

The US figures cover resettled refugees and granted asylees. They do not transfer to the UK by assertion. UK evidence of equal scope was not reviewed for this piece. The UK has an older record. The Society for the Protection of Science and Learning was founded in 1933 as the Academic Assistance Council and is now the Council for Assisting Refugee Academics. A history marking its 75th anniversary records that eighteen of its early grantees became Nobel laureates and 120 were elected Fellows of the British Academy or the Royal Society.

One of that generation is easy to follow. Hans Krebs was dismissed from his post at the University of Freiburg in April 1933 and came to Cambridge that summer, after Frederick Gowland Hopkins wrote offering him sanctuary. In 1953 he shared the Nobel Prize in Physiology or Medicine for the discovery of the citric acid cycle. That is 20 years after he arrived: eight cycles of the length now proposed for refugees.

Put the human back

When the person is removed from the description of any group, the group becomes a threat or it becomes a cause. Both are abstractions. One side speaks of invasion and military-aged men, the other of victims and heroes, and each needs the other’s caricature to be heard. The extremes are what remains when the human is gone.

A refugee is a person who has lost the state that was meant to hold their rights. What remains to stand on is that they are a person. That is the premise of the Refugee Convention, and it is the premise the vocabulary of totals cannot carry.

What the debate needs is modest. Use the word that names the person: refugee, asylum seeker, child, woman, man. Count, but say who is counted. Ask of every measure, the thirty-month clock included, whom it is calculated for and what it does to the one it is calculated on. Then put to the host the question every text above puts to the host.

The person the numbers miss

Alf Dubs was born in Prague in December 1932. In June 1939, aged six, he travelled to London on the Kindertransport. His father met him at Liverpool Street station.

Unaccompanied children still receive five years’ leave under the March 2026 rules, pending a decision on long-term policy. A child arriving as Dubs did would not be on the 30-month rule. No review window of either length reaches what followed. He was elected Member of Parliament for Battersea South in May 1979, 40 years after he arrived. In 1988 he became director of the Refugee Council, the first refugee to lead it: 49 years. In 2016 he tabled the amendment asking the government to accept 3,000 unaccompanied refugee children. It passed. The amendment held the host to its own standard, which is the test the texts above set. His contribution differs in kind from the fiscal and scientific figures earlier in this piece. It is moral and political.

Dubs is 93. In 2017 the country closed the scheme his amendment created, after accepting 480 children. That is 16 percent of 3,000, and it came 78 years after he arrived. Chris Mullin’s review of his memoir, The Boy with the Knapsack, ran in The Spectator Australia on 3 October 2026. It quotes Dubs: “I shall always want to say thank you to this country.”

The final report was due next week. It will now appear in November at the earliest. The Times reported that ministers did not want it to overshadow the Budget on 28 October. The Department for Work and Pensions (DWP) says it will be published this autumn as planned.

The plan has moved before. Milburn’s terms of reference promised a final report by summer 2026. It was then expected in early autumn, then mid-October, and now after the Budget.

The next official NEET count from the Office for National Statistics (ONS) is scheduled for 26 November. The report on how to reduce that number may arrive after it.

In May the response came before the diagnosis. In October the diagnosis waits on the Budget.

The other timetable

The Department for Education (DfE) has its own review of the same children at an earlier age. The schools white paper, Every Child Achieving and Thriving, was published on 23 February 2026. It proposes the largest change to support for special educational needs and disabilities (SEND) since the current system began in 2014. Every school would write digital Individual Support Plans. More children would be taught in mainstream schools.

The timetable is long. Assessments under the new system would begin from September 2029. Changes to support under Education, Health and Care Plans (EHCPs) would not take place before September 2030.

Count forward from this term. A child who started secondary school last month will begin Year 11 in September 2030. Every pupil now in Year 8 or above will finish compulsory schooling under the old system. The reform is real. It is aimed at children who have not yet arrived.

What happens in the meantime

The old system keeps growing. In January 2026, 718,838 children and young people in England had an EHCP, 12.5% more than a year earlier. That is the largest annual rise since plans were introduced. More of them are in mainstream schools than ever: 45.5%, up from 39.2% in 2019.

Then they turn sixteen and begin to move from one department’s figures to another’s. The DfE’s own data show 22,000 young people with an EHCP were NEET in January 2026, 22% more than a year earlier. Of the 50,346 plans that ended during 2025, 17,500 ended with the young person not in education or training. The most common reason recorded for a plan ending was that the young person no longer wished to engage.

That phrase records a decision as the young person’s own. It does not record what came before it.

On the DWP side, the ONS estimates 981,000 young people aged 16 to 24 were NEET between April and June 2026, 13% of the age group, down slightly from just over a million earlier in the year. Milburn’s interim report put the cost at £125bn a year. It projected 1.25 million young people NEET within five years without change. The government’s Jobs Guarantee is not open to the 290,000 young people out of work on health and disability benefits.

The clock nobody schedules

The Remit Gap, in July, named four reviews running on four clocks with no mandate connecting them. Seen Five Times, Held by None, in August, followed care leavers through five systems that each see the same young person and none of which holds them. The white paper and the Milburn review add two more owners to that map. What none of them is timed against is the one clock that cannot be rescheduled: the age of the child.

The DfE describes the child at eleven. The DWP describes them at nineteen. Each department has its own consultation, its own legislation and its own budget round. Milburn’s interim report named the gap: no single system is responsible for stopping a young person from becoming NEET.

What follows is a reading, not a finding. Neither timetable is set by the age of the child. Each is set by the department’s own cycle and, this month, by the Budget. The child’s cycle runs regardless. One clock moves at the speed of government and the other at the speed of growing up. The second clock does not wait.

In February, Inclusion That Excludes argued the white paper’s gaps were the same ones schools had carried for thirty years. The timetable now shows how long they will stay open.

Writing in The Spectator, Isabel Hardman suggested the review’s main use to ministers may be as a future prospect: something to cite as evidence of action without yet requiring any. A report that is always about to arrive costs nothing.

A further conflict sits in inspection. Ofsted rates schools with more pupils on SEND support lower, while the white paper asks those same schools to take more of them. Last week’s piece, A Diagnosis With No Treatment, set out that problem.

The pool

In John’s gospel, a man who has been ill for thirty-eight years lies beside the pool of Bethesda. The water is stirred at intervals, and healing goes to whoever reaches it first. He tells Jesus he has no one to carry him down, and someone else always gets there before him. The story does not end with a better queue. He is healed where he lies.

The reform as written is the pool. The water moves in 2029 and again in 2030. Children who can reach it then will benefit. The cohort already waiting has no one to carry them down.

Patrick

Patrick McCabe, whose GCSEs featured briefly in last week’s piece, is sixteen. He has cerebral palsy and has attended mainstream schools in England since he was six. This summer he passed seven General Certificate of Secondary Education (GCSE) exams at Fallibroome Academy in Macclesfield, giving every answer through eye-tracking technology. His exams began on 2 February and ran every morning until 5 June. A single 30-mark question could take him six hours. “I felt I had to work harder to express what I know,” he said.

He is the child the white paper describes: local, mainstream, included. He got there under the old system, through his own endurance, a school and a mother who fought the exam boards, and a charity that supports pupils using communication technology. His headteacher backs the reforms and worries they will not be properly funded.

Patrick has now started his A-levels. He will have finished them before the first assessment under the new system takes place.

The reform is being written for children like Patrick, and Patrick has already had to manage without it.

It will be the second major change in just over a year. Single-word judgements were scrapped in September 2024. Report cards replaced them in November 2025. In a survey by the National Education Union (NEU), two-thirds of staff inspected under the new system said it was as bad as the old one or worse.

The evidence behind Powell’s concern is real. An analysis by the former inspector Steve Wren found that schools with more pupils on free school meals (FSM), or more pupils with special educational needs and disabilities (SEND), received lower ratings. Secondary ratings tracked closely with how well pupils had done before they arrived.

The problem is real. The proposed fix is wrong.

What context means

Nobody would judge a hospital by its patients’ postcodes. The question is how ill each patient was on arrival and how far treatment took them. That is a measure of progress.

Education has a version of this. Progress 8 compares each pupil’s General Certificate of Secondary Education (GCSE) results with their Key Stage 2 (KS2) test results at age eleven. It measures what a secondary school added to the child in front of it.

It has not been published for 2025 or 2026 results. The KS2 tests for those cohorts were cancelled in 2020 and 2021, so there is no baseline, and no replacement was introduced. The new framework arrived in the two years when secondary schools have no progress measure. Inspectors were left with raw results and pupil demographics. That is the likeliest explanation for Wren’s secondary finding. Without a measure of progress, ratings follow intake.

The fix is to restore accurate measurement of progress, not to adjust for poverty. Free school meals is a fact about a family’s income. It is not a fact about a child’s learning.

England has already tried the other route. Contextual Value Added (CVA) appeared in performance tables in 2005/06. It adjusted for free school meals, ethnicity, mobility and first language. It was dropped in 2010/11, on the grounds that it expected different progress from different groups of pupils. Powell’s proposal brings the same logic back through inspection instead of league tables. The question was decided once and is being reopened.

Wren’s SEND finding points to a further conflict. The schools white paper, Every Child Achieving and Thriving, published in February 2026, asks mainstream schools to include more children with special needs. Schools with more of those children are being rated lower. One policy pushes children in, and the other marks down the schools that take them.

Considering

The word that matters is “considering.” Doing well, considering. Good results, considering the intake. What else would you expect, with so many on free school meals.

I grew up deaf and heard that word often. It was meant kindly. It was a ceiling presented as praise.

Contextual judgement builds “considering” into the measurement. The lower expectation stops being one adult’s remark and becomes the system’s arithmetic.

Four jobs, one instrument

Ofsted was created by the Education (Schools) Act 1992. It is asked to police compliance, deliver public judgements, drive improvement and diagnose what is wrong. Those four jobs pull in different directions. A public judgement must be comparable and final. A diagnosis must be candid, and candour does not survive when the stakes are a public rating and a career.

One instrument cannot be optimised for four purposes at once. Ofsted had ten inspection frameworks in its first twenty years. That is not the record of an institution that knows what it is for. It is the record of one that keeps rebalancing jobs it was never allowed to choose between.

A hospital diagnosis exists to select a treatment. An inspection diagnosis ends with the report. Funding, support and intervention sit with other bodies, on other timetables, if they arrive at all. A diagnosis with no treatment attached is a verdict, and teachers experience it as one.

Whose stress counts

The stress is real. Leading a school through inspection is exhausting, and the judgement is public. Nothing here dismisses that.

But look at what both reforms responded to. The 2024 change followed the death of the headteacher Ruth Perry in January 2023, after an inspection downgraded her school. This one follows complaints from school leaders. Both answer the experience of adults. Neither starts from what pupils need. Most pupils are too young to vote. They do not join unions or write to the Prime Minister.

The debate has become an argument among adults about who is working hardest. In Luke’s gospel, the Pharisee at prayer lists his efforts: he fasts twice a week and gives a tenth of all he gets. He is telling the truth. It is still the wrong account to give, because he is measuring himself rather than attending to anyone else.

Every regulator becomes political in the end. A minister can stand beside it or stand up to it, whichever suits the season. The 2024 change was presented as standing with teachers. This one is presented as fairness. In both, the regulator served as the minister’s instrument, and its purpose remained undefined.

What a student-focused inspection would measure

Accurate measurement means measuring what a child knows, not what the system allows them to show. This summer Patrick McCabe, who has cerebral palsy, sat his GCSEs using eye-tracking technology. The exams ran from February to June. They measured his stamina as much as his knowledge.

Remove the judgements. Measure accurately, identify the gap, resource the improvement.

Start with what pupils need from school. Are they there? Attendance. Can they read and write well enough to learn everything else? Literacy. Can they weigh an argument, test a claim and reach their own view? Critical thinking. Measure each against where the child started, not against where the school is. Report the gap. Attach the resources to the gap.

That is diagnosis matched to treatment. It is also a rethink that should have happened thirty years ago.

A deaf boy was told he was doing well, considering. Nobody measured where he started or named what he was missing. Nobody resourced the difference. What he needed was an accurate measure, a named gap and the support to close it. Most students want the same thing. They do not care about judgements. They want to learn.

I have spent the better part of a year moving a single, narrow set of facts between institutions that each, in turn, told me this. A court process that would not tell me what had actually been filed on my own case. A professional oversight body whose decision rested on a hearing that had never taken place, of a type the case never reached, at a court the case was never heard in. A complaints process at a firm whose own head of governance directed me, in writing, to the two people who had already told me in writing that they would not respond.

I assumed this was my own bad luck, one case, one set of overstretched people, an unfortunate run. Then the public record caught up with it.

There are three tiers of oversight in English legal services: a firm is regulated by the Solicitors Regulation Authority (SRA), and the SRA itself is overseen by the Legal Services Board (LSB). In July 2026, an independent review commissioned by the Ministry of Justice concluded that the LSB, the body at the top of that tower, had “lost its way.” Its own assessments of weaknesses building up at the SRA beneath it hadn’t kept pace with what was actually going wrong, the review found, and when the LSB finally acted, over two firm collapses in as many years, its own response had been slow and overly prescriptive.

Those two collapses were Axiom Ince, in 2023, which cost clients £60 million, and SSB Law, the following year, after which the LSB issued directions and then a public censure. Then, in early 2026, a third firm collapsed: PM Law Group, owing clients a further £40 million or so. A separate review, this one commissioned by the SRA into its own conduct, found “a failure to aggregate information” the regulator already held, and concluded its decisions had been taken with incomplete information, against a risk nobody involved had been able to see the full size of. In September, the LSB escalated again, for the third time in sixteen months.

Read those two findings together and you get a tower, not a single weak floor. The firm’s own governance function could not coordinate a response to straightforward correspondence. The regulator above the firm could not bring together information it already held. The body above the regulator could not see the regulator’s failures while they were happening, and needed an external, government-ordered review to be told so. Each level’s defence is the same sentence, restated for its own scale: we did not have the full picture in time. Nobody at any level was lying. Nobody had to. The picture simply never arrived assembled, at any height in the structure, and by the time an outside reviewer forced anyone to assemble it, the year in question had already happened to whoever was standing underneath it.

I had written to two of these bodies, in roughly this shape, more than a year before either report was published. Not as prediction, and not because I had any special insight, only because the pattern was already visible from inside a single case, if anyone had been positioned to look at it as a pattern rather than as a file. Nobody was. That’s not a complaint that I was ignored. It’s the actual finding, arrived at independently and later, by people with access to everything I didn’t have: the same picture is hard to assemble from any single seat in the structure, including, apparently, from the seats built specifically to assemble it.

None of this required malice. I don’t think any single person in this chain set out to obstruct anything. What none of them had to do, at any point, was own the outcome. That’s the actual mechanism: not obstruction, but distributed unaccountability, a system in which responsibility can always be discharged by pointing at the next box, so no single actor, and evidently no single tier, is ever the one who has to explain why nothing happened in time. Every institution’s alibi is the next institution’s jurisdiction, all the way up, until you reach the one body with no one above it to defer to, and it turns out an outside reviewer had to be brought in even there.

There’s a second mechanism underneath the first, and it does the real work on the ground, whatever is happening at the top of the tower. Distributed unaccountability explains why nobody is responsible. It doesn’t explain why the process is slow. Every one of these bodies runs on its own clock: twenty-one days here, twenty-five working days there, four-week backlogs, out-of-office replies stacked one after another so that by the time one person returns, another has just left. None of these delays looks unreasonable on its own. Together they produce something that functions exactly like a filter, whether or not anyone designed it as one: a claim survives only if the person behind it can outlast the accumulated waiting. Call it a patience tax, levied not on the strength of the claim but on the stamina of the claimant.

The people this serves are the ones already fluent in it, and who already have someone else to absorb the waiting on their behalf. A represented party has staff whose job is to sit in the queue. An unrepresented one sits in it personally, with their own name on every letter and their own life running out on the same clock the institution is running, at every tier, simultaneously.

None of this is a claim that everyone behaved badly. It’s a claim that you can build a system out of entirely reasonable individual decisions, stack several such systems on top of one another as oversight, and still end up, at every level, with the same structural blindness: information not assembled in time, and nobody whose job it was to notice until someone from outside the tower was paid to look at the whole of it at once.

The fix isn’t a friendlier auto-reply or a shorter form. My own case sits, as far as I can tell, exactly where it did before either report was published: on someone else’s desk. Nobody in this chain has lied to me. Each one has simply been the wrong person to ask, and each has told me so, in writing, courteously, before pointing me at the next desk along. The Ministry of Justice can now cite an independent review confirming the Legal Services Board agrees with me in principle. It came eighteen months late, and about someone else’s case entirely. It still hasn’t found a desk with my name on it.

It came to light because a councillor, my brother-in-law, met the man and his partner twice in February 2026, five days apart, and raised concerns about their living conditions. Their bank cards were found drained by cash withdrawals. Harrison was arrested within a week. He pleaded guilty to requiring a person to perform forced labour, an offence under the Modern Slavery Act 2015. The investigating officer said the treatment had been "stripping him of his dignity, his finances, and seriously impacting his health." Alongside the prison sentence, the court issued a Slavery and Trafficking Reparation Order: Harrison must pay his victim ten thousand pounds.

The word doing the work in that offence is "modern." It tells you the law considers this a different thing from slavery itself: a contemporary variant, with its own statute, its own vocabulary, its own reparation order, sized to one man's crime against one victim. A piece of research published two weeks ago gives that word a fixed point to sit against.

The Register of British Slave Traders, compiled by the historian Nicholas Radburn at Lancaster University, has found that enslaved Africans trafficked by the crown-chartered South Sea Company were branded with a hot iron bearing the royal crown. A doctor warned the mark could endanger their lives. The company's board, which included Bank of England directors and members of Parliament, told its agents to proceed anyway. The register lists more than a hundred MPs, dozens of the Bank of England's founding directors, over a thousand women, and forty-five Church of England clergy as investors in the trade. A line of monarchs backed it from Elizabeth I onward. Under James II, the Royal African Company trafficked at least a hundred and fifty thousand people.

Asked to respond, a spokesperson for King Charles said he was grateful to the historians "for contributing to the debate about how such issues might best be addressed." No apology. No reparation order. No figure attached to what might be addressed, or by whom, or for how much. It's what I've called elsewhere a conscience signal: a statement that answers a moral question in public while changing nothing about who holds the money or the decision. Nigel Farage did the same thing calling Jaywick's poverty sad while his party cut the payments that fund it.

Ten thousand pounds is what the law now asks one man to pay one victim of one relationship gone wrong. Nothing is what three centuries of a state enterprise, backed by the Crown, Parliament, the Church, and the Bank of England, trafficking more than three million people, has ever been required to pay anyone. The word "modern" sits exactly at that gap. It lets the newer crime carry a name, a statute, and a number, while the older one carries only the word "debate."

None of the institutions involved in 2026 are new. Harrison's crime depended on the one thing the original trade never needed: for a vulnerable man to go unnoticed. What hasn't changed is who still holds the institutions built on the version that didn't need to hide. The Crown that once profited from branding one on a human being now issues a statement about a debate.

I held that council seat myself, thirty years ago. My brother-in-law did not issue a statement. He had two conversations, five days apart, got the couple into emergency housing, and raised it with police. That became an investigation, a conviction, and ten thousand pounds a homeless man would otherwise never have seen. Nobody wrote a column about his broad shoulders. Nobody had to resign and win a by-election to keep the story from costing him anything. That is the difference between politics and its performance: one moves the money and the decision, the other only moves words around both, leaving them exactly where they started. A king was grateful for a debate. Farage found a village sad. My brother-in-law found a man in a shed, and got him out of it.

The two donors were Ben Delo, a co-founder of the cryptocurrency exchange BitMEX, and Christopher Harborne, a Tether investor, who each gave thirty-six million pounds on 12 September: the largest single donation to a British political party on record, more than seven times what Lord Sainsbury left the Conservatives in 2022. Delo explained the gift in a newspaper column: “the duty of people with broad shoulders and deep pockets to step up.”

Harborne is not a stranger to this story. In July 2026, under investigation by the Parliamentary Commissioner for Standards over an earlier, undeclared five-million-pound gift from Harborne that Farage said had funded his private security, Farage resigned as MP for Clacton. “The people of Clacton should be the judges of my actions,” he said, triggering a by-election the other major parties declined to contest. He won it with 63.3 per cent of the vote that August. The investigation was interrupted, not concluded, and resumes only if he returns to face it. Four weeks later, Harborne gave the party thirty-six million pounds more.

Clacton also contains Jaywick, on the Essex coast, named the most deprived neighbourhood in England for the fourth consecutive time since 2010 by the most recent official Index of Multiple Deprivation, published in October 2025. Farage called it sad, without connecting it to the welfare cut his own party had proposed a month earlier, or to what a fraction of seventy-two million pounds might do for five thousand people in a village he represents.

Decades ago, training for the priesthood, I knew a man preparing for a different order who wore no shoes, only open sandals, and called it a mark of humility. He lived, like the rest of his community, in a well-appointed, carpeted cell. I told him it’s easy to walk barefoot when you’ve got a shag pile carpet in your room. The gesture was real. So was the comfort it sat inside. Neither cancelled the other out.

The sandals were a conscience signal: a visible act of restraint that answers a moral question about the wearer without changing his position. Charitable trusts and foundations perform the same function, with real money attached. Named after their founders, staffed by people the founders chose, directed at priorities the founders set, they do not transfer control. Someone still decides who counts as deserving, and that decision stays exactly where it started, with the person who had the money. A foundation is proof of conscience. It is rarely proof of relinquished control, and it brings almost no one from outside the system into the room where that control is exercised. A donation survives an investigation the same way: it answers the question of conscience in public, while the money and the man who gave it keep the position they started with.

I know several families like this, not one. In each, the pattern repeats: adult children whose conversation is almost entirely about protecting what they already have, inheritance tax, investment returns, in one case a cellar of twenty thousand bottles of wine that has never been opened for anyone outside the family. Different families, same shape. None of it is illegal, or even unusual. It is simply what comfort does to attention, whether the collection runs to twenty bottles or twenty thousand.

There is a story about this, older than any of the men in it. A rich man, dressed in purple and fine linen, feasts every day. At his gate lies a poor man named Lazarus, wanting only the scraps from the table. The detail worth noticing is which of the two has a name. The parable withholds it from the rich man and gives it to the beggar, the reverse of how the story usually gets told. Today the reverse has reversed again. We know Harborne’s five million to Farage, his thirty-six million to the party, Delo’s donation to the pound, Jaywick’s exact ranking, Clacton’s exact concentration of PIP claims. The people living inside those numbers are named only when a journalist goes and finds one.

The Letter of James goes further and gets angrier. It condemns wealth built on wages kept back by fraud, on luxury and self-indulgence while others go without. What it never does, and nothing since has either, is put a number on what a fair wage or profit would be. Cap a political donation, and no one names the figure that would still count as fair influence. Say a cut protects people in genuine need, and no one says where the need ends, or who gets to decide it. That is the same choice sitting behind every foundation and every welfare threshold: who counts, and who holds the pen.

Jaywick was built in the 1930s as chalets for a summer never meant to be lived in year-round. People live there year-round now. One of them, Christopher Thompson, 60, put it more plainly than any report has managed: “there is nothing here for the kids.” His MP found that sad, resigned once over undeclared money from one of his own donors, was returned to Parliament, and then took thirty-six million pounds more from the same man. Both things are true at once, and that is the whole point: money at that scale doesn’t have to choose between sympathy and indifference. It can afford both, survive an inquiry into itself, and still not spend a pound on the difference.

The cruelty wasn’t hidden. It was the credential. And more than once in my career, I heard the justification for it offered as though it were self-evident: the blame placed not on the man administering it, but on the female colleagues who had referred the child to him in the first place. Their act of doing their job, escalating a discipline problem to the person whose role was to deal with it, became recast as the provocation. The man doing the hitting became, in his own account, merely responding to a situation women around him had created and lacked the stomach to handle themselves.

I’ve been watching for that exact inversion ever since, in whichever institution happens to be performing it. This weekend, it showed up twice, in the same forty-eight hours, in plain sight.

Birmingham, 5 September

A protester from the group Climate Resistance interrupted Nigel Farage’s keynote at Reform UK’s party conference. Footage of her removal shows a man in a “Farage 10” shirt pulling her hair and grabbing her jaw and head. A second protester says she was yanked by the hair repeatedly and had security stand on her hair while dragging her out. A third piece of footage shows an elderly attendee striking one of the women with his walking stick.

Asked about it on Sky News, Reform’s Treasury spokesperson Robert Jenrick said: “Oh, boo hoo... you take your chances if you go to a party conference and try and mess it up... come on, get a life.” Farage, shown the footage, laughed. Reform chairman Lee Anderson reportedly joked about crowdfunding the assailant a replacement walking stick. Home affairs spokesperson Zia Yusuf did condemn the hair-pulling specifically, a real distinction worth keeping, but a single condemnation sitting next to a spokesperson’s mockery and a leader’s laughter isn’t what a party actually taking the incident seriously would look like.

The same move, one level down

This isn’t confined to conference-stage leadership. In May 2026, a Reform council candidate, Harry Phillips, was reported to have sent messages calling disabled residents “benefit scroungers” and “skanks,” personally targeting one resident’s wife over her ADHD and autism diagnoses, telling her they weren’t real. The branch’s response wasn’t to deny what was said. It was to recontextualise it: the conduct was “not excusable,” but “a significant departure from his character, triggered by” harassment Phillips had separately received. That’s the identical structure as Farage’s response to the walking-stick assault, running one level further down the organisation. The cruelty isn’t denied. It’s presented as retaliation, which lets the person expressing regret avoid actually condemning what was done.

Dover and Portsmouth, the same weekend

Hundreds of men, dressed entirely in black and wearing balaclavas, blockaded the Port of Dover on the Saturday, then moved to Portsmouth on the Sunday to try to stop a group of migrants being processed, chanting “I’m England ‘til I die.” The demonstration was organised under a group called Patriot Platform. Tommy Robinson, urging his followers to join them, called the masked men “British patriots.” The Home Office called it “intimidating and thuggish behaviour.” Even a sympathetic Reform councillor conceded that local women found the sight of a large group of masked men marching through their streets frightening, while still saying he understood the anger underneath it.

The move that makes it possible

None of this depends on anyone denying the cruelty happened. It depends on a specific rhetorical move, distinct from others I’ve traced in different institutions performing similar work. Asked about the conference footage, Farage didn’t deny it or apologise. He said: “I can have milkshakes chucked in my face, be bashed over the head with placards, and that’s all fine... this is just one-sided.” That’s neither Chris Bryant’s trick, a minister asking aloud how generations of politicians had let a problem happen in a town he was currently responsible for fixing, so the person asking the question is never the one who has to answer it, nor Josh MacAlister’s, calling a crisis “a stain on society because all of us have failed,” diffusing it across an entire population so a shared stain belongs to no one. Farage’s version claims symmetry: your own side’s violence doesn’t need defending because the other side does it too, so nothing that just happened is actually violence, it’s simply the ledger evening out. The version I heard in schools worked differently again: it didn’t diffuse the blame across time or population, it relocated it directly onto the colleagues who had raised the problem in the first place. All three let the person responsible avoid ever being the answer to their own question.

It isn’t new to this party or this year. In late 2021, Boris Johnson falsely told Parliament that Keir Starmer, as Director of Public Prosecutions, had failed to prosecute Jimmy Savile. Days later, a crowd surrounded and chased Starmer through Westminster, shouting “paedophile” and “traitor.” The claim was false. Nobody in government meaningfully retracted it before the mob assembled. The mechanism, a leader’s language converting directly into a crowd’s physical intimidation of a target, is not a Reform invention. It’s a standing feature of British public life that this particular moment happens to be displaying with unusual clarity.

Masking has its own version of the same logic. A balaclava isn’t just concealment. It’s a uniform, and a uniform borrows the authority of order for an act that is, by any ordinary definition, intimidation: hundreds of anonymous men physically preventing other human beings from being processed through a lawful system. Tommy Robinson’s word for it, “patriots,” does for the balaclava exactly what “proper disciplinarian” once did for the cane. It relabels the thing being feared as the thing that’s protecting you.

What the laugh actually measures

A statement can be drafted by an aide. A laugh, caught on camera before anyone’s had time to think about how it will look, can’t. It’s the most honest data point in the whole weekend, more honest than Yusuf’s careful condemnation or Jenrick’s rehearsed dismissal, because it wasn’t produced for an audience, it was simply what the footage triggered before anyone remembered they were being filmed. What it measured was which harmed body counted as real to the person watching, and which one registered instead as material for a joke. That’s the oldest sorting mechanism there is, going back well past any classroom cane: who gets to be a person whose pain is a fact, and who gets treated as a punchline standing in for a person.

The through-line

The teacher who hit hardest was called strong because the culture around him had already decided that strength and cruelty were the same thing, and needed only someone brave enough to demonstrate it, and someone else to blame for having raised the problem at all. Nothing about that logic has gone anywhere. It has simply changed uniform: from a cane in a headmaster’s study, to a walking stick swung at a conference, to a hundred identical black balaclavas standing in a road, each one daring you to say out loud that what you’re looking at is fear wearing the costume of order.

Ian McNiff

Systems on Trial (ianmcniff.substack.com)

Examining the machinery beneath the narrative.

The jobs that were already going undone

In 2013, in 14% of private family law cases both parties had no legal representation. By the final quarter of 2023, that figure had risen to 38%, very nearly trebling, according to the Law Society’s own analysis of Ministry of Justice figures. By early 2014, within a year of the Legal Aid, Sentencing and Punishment of Offenders Act (LASPO) taking effect, 80% of family court cases involved at least one unrepresented party. Between 75% and 80% of litigants in person are in that position because they cannot afford a solicitor, not because they chose to go it alone.

None of that has anything to do with artificial intelligence. It happened because the profession’s economics stopped extending to most of the people who needed it, over a decade before any generative tool existed to blame. A solicitor complaining that AI threatens their livelihood is describing a threat to the work they were doing for the clients who could pay. The much larger population the profession had already stopped serving isn’t in that complaint anywhere, because that population was never the job AI is supposedly taking.

The regulator paid for by the people it regulates

The Solicitors Regulation Authority (SRA), the body meant to hold the profession accountable for exactly this kind of failure, is funded 60% by law firm fees, calculated on the firm’s own turnover, and 40% by individual solicitors’ practising certificate fees. Total funding for 2026/27 is set to rise to £111.5 million, paid entirely by the profession it regulates. There is no independent funding stream, nothing that doesn’t ultimately run back through the same firms whose conduct the SRA is supposed to police.

That structure sits uneasily alongside what’s happened with AI-generated fake citations in UK courts this year. In R (Ayinde) v Haringey, a pupil barrister settled grounds for a homelessness judicial review citing five authorities that did not exist. In a companion case, Al-Haroun v Qatar National Bank, a solicitor’s fabricated citations were challenged, and she either invented more or denied using AI at all. Dame Victoria Sharp, ruling on both, found generative tools “are not capable of conducting reliable legal research” and warned lawyers who don’t verify their output face “severe sanction.” The pattern recurred: a barrister who produced further unverified material during a hearing’s own adjournment, a solicitor who signed a statement of truth without checking a paralegal’s AI-drafted work at all. In every one of these cases, it was a judge who caught it. I haven’t found an instance this year of the SRA surfacing one first, which may simply reflect how few cases there are to check against, but it’s consistent with what you’d expect from a regulator whose income depends on the continued goodwill of the firms it regulates: a structural incentive toward reactive rather than probing oversight, my reading of the arrangement rather than something the SRA has said about itself.

The excuse and the symptom

“AI is taking our jobs” is a story about competition and disruption, which is a comfortable story for a profession to tell about itself. It’s a story about the future. The actual record shows a profession that had already stopped reaching most of the public well before AI arrived, and a regulator structurally positioned to catch failures late rather than early. AI didn’t create either problem. It has simply made one of them, the citation fabrication, visible and countable in a way the other, a decade of LASPO-driven exclusion, never has been, because nobody built a public database to track how many people gave up trying to find a solicitor at all.

Two figures nobody reads together

A trebling of unrepresented family cases in a decade. £111.5 million in regulatory funding, all of it sourced from the profession being regulated. Two figures from the same industry, and no single report puts them side by side.

A covenant that was already broken

The duty of candour to the court and the promise of equality of arms regardless of means are covenantal claims the profession makes about itself. Neither requires AI to be broken. Both were already compromised, one by cost, the other by a regulator with no independent standing to enforce it early.

A structural fix nobody’s proposed

Parliament created a regulator funded by the industry it oversees and calls that arrangement independent regulation. Nobody has proposed changing the funding model, because that would be a structural fix, and structural fixes cost the people who’d have to pay for them.

The anxious conference and the decade of silence

A profession can be generously self-critical about a new technology’s risks, holding conferences, issuing warning notices, drafting guidance, while remaining almost entirely silent about the much larger and older population its own economics quietly wrote out of the client category before any of this began. The visible failure gets the anxious conference. The invisible one gets a decade of silence.

Beth

Beth’s mother paid for a solicitor to help with her custody case, until the money ran out. Beth carried on alone. The solicitor had given her some advice before withdrawing, but none of it told her what the later hearings would actually involve, documented in research into litigants in person in the post-LASPO family court. She lost custody of her daughter. Nobody fabricated a citation in Beth’s case. Nobody needed to. The representation she needed simply stopped being affordable partway through, which is a quieter failure than a fake case law reference and a far more common one, and it happened without a single headline, a single sanction, or a single regulator noticing at all.

Ian McNiff

Systems on Trial (ianmcniff.substack.com)

Examining the machinery beneath the narrative.

These six are drawn from 112 notifications the Department for Education (DfE) received in 2025 for the deaths of care leavers aged 18 to 24 in England. The review that examines them, ‘You really are on your own’: a study of the early deaths of young people who were in care, published 25 August 2026 by Ashley John-Baptiste and Clare Chamberlain, puts the death rate for this group at roughly three to four times the general population’s. An earlier, Nuffield-linked estimate put the same gap at seven times. Two credible figures, produced within a few years of each other, sit more than double apart. That isn’t measurement error. It’s a sign that no single body owns the count.

The pipeline nobody assembles

Just 7.2% of looked-after children achieve grade 5 or above in English and maths GCSE, against 40.1% of everyone else. Some 41% of care leavers aged 19 to 21 are Not in Education, Employment or Training (NEET), against 12 to 13% of their peers. Over a quarter of the adult prison population has been in care at some point, and care-experienced people are seven times more likely to be imprisoned than those who haven’t, a separate multiplier from the death-rate figure above, arrived at by different researchers measuring a different outcome. Then the death notifications: three to four times the general rate, one fifth of them landing at exactly age 18.

Five figures. Five different owners: DfE exam data, DfE and Office for National Statistics (ONS) labour market data, Ministry of Justice custody data, and DfE’s own safeguarding notifications. Each dataset is internally coherent. None of them speaks to the others. There is no single document in England that shows what happens to the same child moving through all five, because no single body has ever been asked to build one.

What breaks before eighteen

Michael had more than ten foster placements before he turned eighteen, several of them positive, all of them ended by circumstances outside his control. A 2021 study led by University College London’s Amanda Sacker, tracking long-term outcomes for people who had been in care, found the type of placement mattered more than almost anything else: those raised by relatives did best across health, income and family life, those in residential care did worst, foster care sat between the two. The pattern isn’t about placement type. It’s about what each one does or doesn’t let a child build: a small number of people who stay, long enough and steadily enough, to become someone the child can trust.

Janey had one. A foster carer she called Mum, who encouraged her singing, who Janey stayed in touch with past eighteen. Her sister’s view, reported in the review, was that a formal ongoing role for that foster carer in Janey’s adult life might have changed what came after. The relationship existed. What didn’t exist was any structure to carry it across the threshold with her.

The eighteenth birthday matters. But a cliff is only fatal to someone already unsteady on their feet, and the steadiness or its absence gets built in the years before eighteen, placement by placement, relationship kept or relationship severed. Most of the young people in this review had ten or fifteen years of exactly this instability behind them before the state’s support fell away in front of them. The age-18 rupture is the trigger. The fragmented years preceding it are the loaded chamber.

One clock, five doors

What the trigger does is fire five separate mechanisms at once, none of them synchronised with each other. Children’s education stops being a corporate parent’s statutory concern. Paediatric and children’s mental health services stop monitoring. Children’s social care becomes adult social care, where the language of “capacity” allows a young person to decline a service that safeguarding duties would have overridden the day before. Sentencing stops carrying any flag for care history. Employment support switches from a virtual school head who has known a child for years to a jobcentre that has never met them.

One date. Five doors, all closing within the same fortnight, none of them checking whether anyone was standing on the other side. Every young person named in the review walked into one of these doors alone. The review notes, almost as an aside, that a young person is legally the same person at seventeen years and eleven months as at eighteen and one month, and administratively an entirely different one. That single sentence carries the argument. A threshold set by a calendar cannot answer a question that only relationship can answer, which is whether someone is ready. It answers a different question instead: whether the institution can close the file.

Confession without a confessor

Children’s minister Josh MacAlister called the findings “a stain on society because all of us have failed.” Set that next to a Westminster minister, covered here previously, who asked aloud how generations of politicians had let entrenched deprivation happen in a town he was currently, personally, responsible for addressing. The two sentences run in opposite directions and arrive at the same place. One spreads blame backward through time, so the person asking the question is never the one who has to answer it. The other spreads it sideways through the population, so a stain shared by everyone belongs specifically to no one. Both let a serving minister, holding the actual levers of policy, convert his own accountability into a shared condition instead of a personal one. The review itself states plainly that it was not within its scope to develop detailed policy recommendations. The document commissioned to name the crisis was built, from the outset, without the power to prescribe the fix.

Three ways to disappear

There’s a woman, ninety-two, in a former mining community, who once couldn’t pay her poll tax. No telephone, no bus route to the office that might have helped, immobile. She left no data trail because the infrastructure never reached her in the first place. She wasn’t rejected by any system. She was never a category a system was built to notice. That is the first kind of exclusion: absence.

There’s a Harvard researcher whose grant, sound enough to clear expert peer review, was stopped by an algorithm screening for disfavoured words. He was visible, reached, assessed, and then filtered out on a technicality that had nothing to do with the quality of his work. That is the second kind: rejection.

Care leavers are neither of these, which is what makes the pattern worth naming. They are, if anything, the most documented young people in the country. Five separate systems see them clearly, completely, and accurately, each within its own narrow remit. What doesn’t exist anywhere is the infrastructure that would let those five sightings add up to a person. That is a third kind of exclusion, harder to see precisely because it looks like the opposite of neglect. Nobody failed to notice these young people. Everybody noticed them, separately, and nobody was ever responsible for putting the sightings together.

The quiet that follows

No comment came from the prime minister. No senior cabinet voice beyond the children’s minister spoke at all. Coverage existed, LBC, Morning Star, the specialist press that follows social care closely, but it moved at wire-service volume, not front-page volume. The day the review was published, LBC’s own site carried the story a few paragraphs above coverage of Andy Burnham signing an AI defence data-sharing agreement with Volodymyr Zelenskyy in Kyiv. That is simply where the attention went. When the Covid inquiry produced a comparable “national tragedy” framing with a death toll attached, it led every front page in the country. This one didn’t lead any. Whether that gap was engineered or simply allowed to happen is a separate question from whether it exists. It exists.

David

David spent three days a week at the care leavers’ hub, chatting with staff, managing his diabetes well enough that everyone who knew him assumed he had it handled. In the summer of 2025 the hub closed for three weeks of refurbishment. His personal adviser was on leave. Sometime in that gap, his insulin ran out, or his levels turned, and there was no one left whose job it was to notice. His body wasn’t found for close to a fortnight. A neighbour called the police, not the local authority that had been legally responsible for him seven years earlier and administratively unconnected to him for the last two.

He had, at various points, a children’s diabetic nurse who knew his readings, a progression adviser from the virtual school, a personal adviser who helped him keep his flat in order when things got hard, and a hub three days a week where somebody would have clocked his absence within a day. Every one of those people did their job. None of their jobs included the other three. He was seen, completely, by four different systems, and held by none of them at once. That isn’t a gap in the record. It’s what the record was built to produce.

Ian McNiff

Systems on Trial (ianmcniff.substack.com)

Examining the machinery beneath the narrative.

Movement is the only thing the language of results day knows how to describe. And movement can only be measured for a child who is already inside the frame. This is a piece about the ones who aren’t, and about why their absence is not an accident of the system but a function of it.

The grade fixed before the pupil sits down

The distribution of grades is not discovered from performance. It is set in advance, at least in part. The regulator, Ofqual, works to a principle called comparable outcomes: it monitors whether a cohort’s underlying attainment has genuinely shifted, using an annual National Reference Test, and adjusts grade boundaries accordingly, so that a grade means roughly the same standard from one year to the next. This year Ofqual lowered the maths grade boundaries after the Test suggested a small, sustained rise in underlying attainment, and its executive director of delivery, Amanda Swann, called the change a “direct reflection of stronger student attainment.”

The verdict, not the dialogue

I know the wait from the inside. I am deaf. Results day was never a dialogue for me. It was a verdict, delivered once, with no hearing beforehand and no right of reply after. The only input a student has into that judgment is the exam itself: a single piece of evidence, produced under exam-hall pressure, closer to a witness statement extracted in the worst possible conditions than to any fair account of what a person knows.

If you were recorded as having special educational needs, as I was, there was no channel to say the record didn’t capture how you actually understood something. The category spoke for you before you were in the room. Nobody explained the verdict when it arrived. What explanation there was came earlier, and it was admonishment rather than diagnosis: attendance figures, effort grades, a deficit read back to you, with no account of what had produced it or what to do differently before the next judgment.

That absence of dialogue is not historical. I wrote a set of revision guides this year for my grandson, who sits his own GCSEs next year. After his mocks, he came away knowing certain grades weren’t good enough. Nobody gave him anything more specific than that. No breakdown of what the gap was made of, no next step beyond working harder. He has two parents at home and a grandfather writing him bespoke material, and the system’s only feedback was still a closed verdict. If that is what the silence looks like for a boy with every support available to him, it is worth sitting with what it looks like for a child with additional needs, or with nobody at home able to fill in what the school didn’t say.

Two accounts of the same mechanism

The day after this year’s results, Education Secretary Lucy Powell told BBC Radio 4 that under comparable outcomes, “around a third of students... every year can’t get a grade 4.” Amanda Spielman, the former chair and chief regulator of Ofqual, publicly rejected the claim, saying Powell had “either didn’t understand her briefing or intentionally misled the public.” Ofqual’s own published guidance says there is no fixed quota on how many pupils can pass.

Powell has also described the effect of that same third failing English or maths repeatedly as “we are branding them a failure.” Both statements are hers, made within days of each other. One insists a third of pupils are structurally destined to fail, which would make the outcome nobody’s fault. The other calls the branding of that same group a harm the system does to them, which would make it somebody’s responsibility. Neither account was offered to a pupil directly. Both were offered to justify a policy, in one direction or the other, after the grade was already given.

Fourteen years is not a delay

Underneath that dispute sits a plainer fact. The Education Policy Institute has tracked the secondary school disadvantage gap in England since 2011, the year the pupil premium was introduced: extra funding attached to every pupil eligible for free school meals, built for one stated purpose, to close that gap. The Institute’s series shows it narrowing from 19.7 months in 2011 to 18.2 months by 2014, then stalling. By 2022 it had widened past its own starting point, to 18.8 months. By 2023 it stood at 19.2 months, close to where it began fourteen years earlier, and the Institute records it as the widest since the series started. Fourteen years, one named policy instrument, four changes of secretary of state, and the distance it was built to close is now roughly the distance it started at.

That is a fact. What it means is a reading, and it is the same reading I made in a different piece, about a different street. What the Streets Carried named a pattern that held for 238 years: soldiers honoured, weavers left in an unmarked grave, the record kept clean by simply not keeping one. The conclusion there was three words. That is not neglect. That is policy. Fourteen years is a shorter span. The shape is the same.

The ones the register never counted

A second group falls out before the exam is even sat. FFT Education Datalab, which tracks attendance across roughly 10,000 schools, reports that across the 2025/26 academic year, the year this week’s cohort actually sat in Year 11, secondary persistent absence ranged from 23% in London to 27% in the North East, little changed from the year before. In 2018/19, before the pandemic, the secondary rate was around 14%.

These pupils have a name in the commentary that surrounds the statistics, though not in the statistics themselves. Robert Halfon, at the time chair of the Education Select Committee, warned in 2022 of children at risk of “slipping through the cracks and lost to the system forever.” A pupil who was off-rolled, or never returned to school after 2020, produces no grade at all. Results day has a category for every grade from 9 to 1. It has no category for absence from the register itself, and no dialogue with the child about why they left it.

The place the story runs on a different day

The fourth concealment is geographic. Blackpool’s local authority recorded a General Certificate of Secondary Education (GCSE) Attainment 8 average of 37 for its most recent cohort, against an England average of 43.8. That figure sits in a Department for Education release most results-day coverage does not open.

Ekhlas Rahman, headteacher of Cumberland Community School in Newham, one of the poorest communities in the country, said this year that 60% of the school’s disadvantaged pupils reached a grade 5 in English and maths, against a national figure of 26% the previous year, and that “there will be no barrier to success” at Cumberland. The achievement is real. But it is spoken in the same week the national data shows the disadvantage gap widening, and a single local success, offered in the register of universal triumph, does the work of concealing the trend it sits inside.

What is actually being measured

None of this means the aggregate is wrong. Grade 4 and above at 67% is an accurate description of the pupils who were entered, sat the exam, and were marked. The sentence describing that aggregate is written as if it describes everyone. It was only ever built to describe the ones still in the room, and even they got no hearing before the verdict was passed.

My grandson has a grandfather who can write him a revision guide. Most children in the persistently absent figures do not. Most children behind Blackpool’s Attainment 8 average do not. Results day gives all of them the same thing regardless: a verdict, no dialogue, and no appeal that does not already assume the record was right in the first place.

Systems on Trial examines the machinery beneath the narrative. ianmcniff.substack.com

That was one day. Two more are booked for 21 and 22 September, and this time Cardinal Allen won’t be alone. Staff at Corpus Christi High School in Preston have voted to join them. The dispute isn’t about pay. It’s about a deadline the Bishop of Lancaster, Paul Swarbrick, set in 2020: every Catholic school in the diocese, from the Fylde coast to Cumbria, folded into one of three diocesan trusts by the end of this academic year.

Behind a single strike day sits six years of a clock nobody outside the diocese was watching. Twenty of the twenty-two Lancaster schools in Blackpool, Fylde and Wyre have already joined the Blessed Edward Bamber Multi-Academy Trust (MAT). Only Cardinal Allen and St Bede’s, Lytham, are still outside it. The National Education Union (NEU) says up to five teachers at Cardinal Allen could lose their jobs as a direct result of conversion.

None of that is unusual for Lancaster specifically. Schools Week’s analysis, last updated in 2024 and still cited without revision in its 2026 coverage, found that eighteen of England’s nineteen Catholic dioceses are pursuing the same kind of trust growth. Every diocese named in reporting since is consistent with that count, not a departure from it. The Archdiocese of Liverpool has torn up an earlier plan for twelve smaller trusts and is now moving over 200 schools into three. The Archdiocese of Birmingham is folding all 236 of its schools into five mega-trusts, one of them a 63-school merger that will become the second-largest academy trust in England. Westminster has abandoned its own earlier plan for smaller, local trusts and now wants four chains each serving over 10,000 pupils. Salford wants three trusts of at least fifty primaries and ten secondaries each by 2030. Lancaster isn’t an outlier. It’s on schedule, in a system that is consolidating faster, not slower, than it was two years ago.

None of those totals has a face attached to it. The next document does, or is supposed to.

Cardinal Allen’s own admissions policy says the school exists to offer opportunities for the formation of each person, integrating human development and the values of Christ. Ofsted inspected the school in April and found it meeting expected standards in every area, writing that pupils thrive in a culture of warm relationships rooted in Catholic values. Read those two documents together and they’re describing the same thing: a school whose whole stated purpose is what happens between one person and another, not what happens on a consolidated balance sheet.

But Ofsted isn’t the inspection that’s supposed to test that purpose. Every Catholic school in England is also inspected separately, under Section 48 of the Education Act 2005 and Canon 806, by inspectors trained by the Catholic Schools Inspectorate and reporting to the bishop rather than to the state. That inspection grades a school on the extent to which pupils contribute to and benefit from the Catholic life of the school, on the quality of provision for that life and mission, and on how well leaders and governors promote, monitor and evaluate it.

I made the underlying argument once already, in Nottingham, before Lancaster’s timetable had reached this stage: the more autonomy a diocese claims from the local authority, the more of its own capacity it must spend running a local authority in miniature. HR structures, finance teams, compliance functions, all secular by design, all built at trust level instead of diocesan or council level. Every hour and every pound that machinery consumes is an hour and a pound not spent on the thing Voluntary Aided status was supposed to protect in the first place. Lancaster’s three trusts are that argument playing out at diocesan scale. Liverpool’s 200 schools, Birmingham’s 236, Westminster’s 187, are the same argument playing out nationally, at a pace that has clearly quickened rather than slowed since the last time anyone counted.

Mass attendance is not the measure I’d use to judge whether that trade was worth it, and I don’t think its decline tells us anything except that the Church’s relevance to the people it was built to serve has been shrinking for decades. What matters is what an institution does with that shrinkage. Here, as everywhere else in this project, the answer is not to go back to the people and ask what they need. It’s to shore up the power base, protect the finances, reduce the number of people in the room who might object, and increase the distance between the institution and the people it exists for. Lancaster’s timetable, Salford’s target of fifty-plus-school trusts, Birmingham’s five mega-trusts: none of that is a plan to put more priests in more parishes or more chaplains in more corridors. It’s consolidation, and consolidation is what a shrinking institution does to protect itself, not what it does to serve the people leaving it. That would be true of a secular institution facing the same numbers. It happens to be true of this one too.

That answer arrives in the same year the Church’s own leadership said the opposite, loudly, from Rome. In May, Pope Leo XIV released his first encyclical, Magnifica Humanitas, timed deliberately to the 135th anniversary of Rerum Novarum, Leo XIII’s founding document of Catholic social teaching on the rights of workers. Without a corresponding growth in humanity, the new encyclical warns, technological and administrative progress risks producing “having more” without “being more,” a world where a person’s value gets measured by what they produce rather than who they are. That is the Pope, this year, naming exactly the failure mode a diocesan academy trust represents in miniature. A school folded into a fifty-school trust to protect a balance sheet is that warning made local.

It is not hard to see where a diocese with that same energy could make a difference instead. England has a live NEET crisis, a vocational education agenda still waiting for anyone to specify what a young person is meant to be able to do at the end of it, and a curriculum debate that keeps measuring exposure rather than application. It has an inclusion agenda and a mental health crisis in schools that neither Ofsted nor Section 48 currently treats as central to what a Catholic education is for. A diocese with the standing of Rome behind it, in the same year Rome published a manifesto for human-centred systems, could have spent its authority arguing for curriculum built around humanity rather than compliance. Instead the energy has gone into governance architecture: which trust, how many schools, whose board, whose funding agreement. The choice was available. It was not the one made.

Lancaster adds something Nottingham hadn’t yet shown me. Education Uncovered reports that the diocese is not renewing the appointments of school governors who haven’t supported academisation, a practice also alleged of the Diocese of Westminster. The National Association of Head Teachers has warned that the resulting vacancies could put school safeguarding at risk. A governing body exists, among other things, to ask questions on behalf of the people a decision will affect. Not reappointing the people who ask them is not a scandal. It’s tidier than that. It’s a mechanism for making sure the question doesn’t get asked twice, and it means the power to decide what “Catholic life” requires sits with fewer people at every stage of the process, not more.

Nat Kerray asked it once already. He’s a curriculum leader at Cardinal Allen and the school’s NEU representative, and before any of this became a strike, he was part of the group that looked seriously at joining a trust and turned it down. He looked at the sums and decided joining a trust made no financial sense, he told the local paper: it would have cost the school money it didn’t have to spare, on top of losing up to five teachers. That’s not a man taking a position on principle. That’s a man who did the sum, on his own school’s numbers, before the diocese did the sum for him and reached a different answer. In July he stood outside the same building holding a placard instead of running a lesson. In September he’ll do it again, for two days instead of one.

An inspection report can write a sentence like pupils thrive in a culture of warm relationships. It can’t put a face to that sentence. It can’t promise the face will still be there in September.

The face belongs to Nat Kerray, on Melbourne Avenue. It belongs to whichever governor in Lancaster or Westminster finds their term wasn’t renewed. No vote taken against them. No reason given.

A trust board can meet every statutory obligation it has without looking at either of them. Neither one shows up as a line in the document that proves a trust is viable, and neither shows up in the Section 48 report that’s supposed to prove something else entirely: that the point of all this was ever the people in the building.

None of that appears anywhere in a trust conversion timetable either. A Multi-Academy Trust’s founding documents answer questions about governance, finance and compliance. They do not ask whether the sixteen-year-old six streets from Cardinal Allen, who left school with nothing to walk into, was ever going to be visible to the structure being built around him. Blackpool’s NEET rate among sixteen and seventeen-year-olds is more than double the national average, and on the broader sixteen-to-twenty-four measure roughly one in five of the town’s young people are not in education, employment or training, a rate that could be seen as catastrophic for the young people concerned. Fylde and Wyre, the districts either side of Fleetwood, are lower but rising. Lancaster itself has seen one of the steepest year-on-year increases in the county. That is not a coincidence of geography. It is the same map. The diocese spending six years engineering three trusts across the Fylde coast, central and north Lancashire and Cumbria is the same diocese whose largest town has a NEET rate this severe. It was true before Bishop Swarbrick set his deadline in 2020. It will still be true after every school in the diocese has a trust’s name on the letterhead. The conversion was never going to touch it, because it was never designed to.

This is the lesson underneath the whole story, and it isn’t unique to Lancaster, or to the Catholic Church, or even to education. Liverpool’s 200 schools, Birmingham’s 236, Westminster’s 187, a diocese not renewing a governor’s term: these are all announcements about structure. Bigger units, cleaner compliance, a board that can absorb a deficit. Every one of them is defensible on its own terms, and every one of them is easier to make, easier to measure and easier to announce than the harder work of reaching the pupil who was already falling through the gaps before the reorganisation began. Structure is legible. A trust board can show its viability in a spreadsheet. It cannot show, in the same spreadsheet, the NEET teenager in Blackpool, the governor whose term wasn’t renewed, or Nat Kerray standing on Melbourne Avenue. None of them run through the ledger the announcement gets made from.

That is the pattern this project keeps finding, in a diocese, a regulator, a report, a research funding algorithm: the people who most need an institution to work for them are the ones least likely to appear inside the document that proves it’s working. A system can hit every structural target it sets itself and still have made no contact at all with the person it was supposedly built to reach. Bishop Swarbrick’s trusts will likely all be delivered on schedule. Blackpool’s NEET rate is not on that schedule, and nobody has yet had to explain why.

Ian McNiff

Systems on Trial (ianmcniff.substack.com)

Examining the machinery beneath the narrative.

Jason Arday was found dead in London on 14 August. He was forty one. A week earlier he had resigned as Professor of Sociology of Education at Cambridge, a post he took up in 2023 as the university’s youngest ever Black professor. The resignation followed weeks of scrutiny into his academic qualifications, his honorary appointments, and allegations of plagiarism in the doctorate he completed at Liverpool John Moores University (LJMU).

LJMU is the institution with the authority to rule on that doctorate, because it is the institution that awarded it. It convened an Academic Misconduct Panel. In February, the panel examined the allegations of plagiarism, cheating in the viva voce, and exceeding the permitted word length, and reached a unanimous conclusion on each: not proven. That did not stop the story. Other investigators, working from outside any formal process, kept publishing their own counts of matched passages, and much of the press kept treating those counts as more credible than the panel’s finding.

His family released a statement after his death. They did not defend the doctorate. They did not answer the critics. They described him as an amazing father, partner, brother, uncle and son, and asked to be left alone. Amazing is doing real work in that sentence. It is a judgment offered with no burden of proof attached, the opposite of everything argued about him for three weeks. Five words, one adjective, and none of it a case that needed making. Almost every word used against him elsewhere was a case. What follows takes their five words in turn, and asks what each one actually requires of us, set against what the argument actually did.

I. Father

Read one way, Cambridge’s own 2023 announcement could be a kind of paternal posture. It did not simply say he was one of the youngest professors it had appointed, which was also true. It specifically said he was the youngest Black person ever appointed to a professorship there, in an institution eight hundred years old, and it said so with evident pride, the way a parent claims credit for a child’s arrival. That is a reading, not a finding, and I hold it loosely. What is not a reading is what came after. When the trouble arrived, the same institution reached for the other half of that posture. Not protection. Correction. An investigation into the very appointment it had been proud to announce.

Neither posture is fatherhood, whichever way the first one is read. A father does not require his child to have earned the relationship through flawless conduct, and does not revoke it the moment conduct is questioned. His family did not simply call him a father. They called him an amazing one, and needed no panel to say so. Cambridge spent three years presenting his arrival as evidence of its own progress, then weeks investigating whether the arrival had been earned. His family required neither measurement. Amazing was enough, and it was theirs to give.

II. Partner

A partnership is a relationship between equals. Both parties chose it. Neither one holds a permanent right to sit in judgment over the other. Almost nothing written about Jason Arday this year was a partnership in that sense. It was adjudication, on one side pressing a verdict, on the other side defending against it, and everyone else keeping score.

Two questions are worth putting to the investigators who kept the story running after LJMU’s panel had already ruled, because nobody else was claiming the authority to overturn a university’s own disciplinary finding. Do they run the same forensic comparison against every doctorate awarded by every university, or only when a name and a story make the exercise worth the effort. And would they accept a stranger with no institutional standing publishing a competing verdict on their own methodology, and watch the press treat it as more credible than their own. If the answer to either question is no, the exercise was never a partnership with the truth. It was a hierarchy dressed as scholarship, in which one side always retained the right to keep asking until it got the answer it wanted.

Sylvia Wynter’s work, which sits behind the doctorate I am now writing, argues that the modern idea of the fully human, the rational, credentialed, self-authoring subject the West calls Man, was never a neutral description of what a person is. It was a category built to rank people against each other, admitting some fully and others only on probation, pending further proof. Partnership does not work that way. It does not require you to keep proving yourself acceptable to someone else’s standard. Arday said his autism meant he processed information partly through mimicry, and under close scrutiny this resembled the borrowing his critics wanted it to be. LJMU’s panel, examining the actual thesis, called it honest and reasonable error. Much of the response to his own account of himself was not engagement with that finding. It was ridicule of the idea that a mind could arrive at knowledge by a route the institution had not pre-approved. That is not partnership. It is hierarchy, and he was never permitted to be its equal.

He knew what real partnership looked like, because he had lived it. By his own account, the person he chose to build a life with taught him to read and write, after a childhood in which almost nobody thought he would. Not through correction. Not through a panel. Through patience, and the assumption that he was already whole and simply needed help getting somewhere he was always capable of reaching. That is what partnership does. It does not wait for you to arrive fully formed before it commits to you. He had that, freely given, for the rest of his life. No institution ever offered him anything close to it.

III. Brother

Brotherhood does not require you to represent anything. It is solidarity that does not depend on your usefulness to a cause. What Jason Arday received instead, from the moment questions about his record surfaced, was two camps that each needed him to represent something. One side needed him to prove the system still discredits Black excellence by any means available. The other needed him to prove that standards had been abandoned in the pursuit of representation. Both camps produced supporters who spoke of him warmly, as an ally, almost as kin. But brotherhood does not keep score of what you can still prove. What he received, on both sides, still asked something of him in exchange for the warmth.

He was an actual brother, raised alongside his siblings on a council estate in south London. That bond asked nothing of him except to belong.

IV. Uncle

An uncle is not required to correct you. He is not your judge and not your teacher. He is usually the one who tells you that whatever you are agonising over has happened before, to someone else, and that you will survive it too.

Ecclesiastes says the same thing, in a register the plagiarism debate never allowed itself to consider. What has been will be again. What has been done will be done again. There is nothing new under the sun. Not a cynical verse. A sober one. It says the fantasy of pure originality, a mind producing something from nothing, owing nothing to anything that came before it, was never available to anyone. The academy runs a citation system to manage that fact honestly, and the system matters. But there is a difference between the discipline of citation and the fantasy that true originality arrives uncontaminated by influence. Nobody clears that bar. Beethoven did not clear it. Neither did the theologians who trained me, nor the inspectors who trained me after them, nor, if I am honest, this piece. The question was never whether Arday’s mind touched other minds on its way to a conclusion. Every mind does that. The question the panel actually asked, and answered, was narrower and fairer. Did he do it honestly, within the standards of his time. They said yes. What followed was not a request for more evidence. It was a refusal to let the matter rest, from people standing in exactly the relationship an uncle refuses to take up. Permanent corrector. Permanent judge.

Arday spent much of his career trying to hold a door open for people whose path into a room like his did not look like the standard one either. Not by lowering what the room expected of anyone. By insisting the room was wide enough to hold more than one way of arriving in it. That is the avuncular instinct at its best. Wanting someone else to get there too, without requiring them to get there exactly the way you did.

V. Son

Cambridge’s own announcements did not only make him a professor. They made him a prodigy, the youngest, the one who overcame. That language, however well meant, keeps a person permanently a child in the public eye. A prodigy is always being measured against an age. He is not allowed the ordinary adult right to make a mistake and be unremarkable about it, because the entire story depends on him remaining exceptional. The youngest-ever framing that made him a sympathetic figure in 2023 is the same framing that made every subsequent question about him feel, to a watching public, like a fall from a height rather than an ordinary professional dispute about citation practice.

He was, underneath all of that framing, somebody’s son. The most dependent of the five words his family chose, and the last one they said. Sons are not required to have earned their place before they are loved. That is the one relationship in the family’s statement that asks nothing of a person at all, and it is also the one word that the entire argument, on every side of it, found no use for.

His family did not make a counter-argument to the plagiarism allegations or the accusations of racism on either side. They named five relationships and asked to be left alone inside them. That is a much smaller request than it should have to be, and it is the only one in this entire story that anyone still has the power to grant. The doctorate was examined. The finding was given. Whatever anyone still wants to argue about the academic question, none of it touches the five words his family chose. Father, partner, brother, uncle, son. That is who he was, underneath everything that was said about him, and it is who he remains.

Ian McNiff writes Systems on Trial and Top Right Learning on Substack, and is completing a PhD on Haiti, decolonial political theology, and the limits of the category of the human.

ianmcniff.substack.com

Read the other direction, the formula says something sharper than its authors intended. Politeness is not a virtue distributed equally across a conversation. It is a tax, and the tax rate is set by whoever holds more power in the room. The person with less power pays it in full. The person with more power can choose whether to pay it at all, and when they choose not to, the choice itself rarely gets named as a choice.

What the tax buys

Scholars writing on tone policing describe the same mechanism from the other end. The demand for civility, directed at whoever is raising a grievance, does not create fairness in a conversation. It enforces whichever arrangement existed before the grievance was raised. Martin Luther King made the same point from inside it, in a letter written from a jail cell, naming the person most invested in delay not as an opponent but as the moderate more committed to order than to justice. The obstacle was never the shouting. It was the demand that injustice be raised quietly enough not to disturb anyone already comfortable with it.

That is the mechanism running underneath every institution I have written about this year. Politeness is not the opposite of aggression. It is aggression’s most efficient form, because it does the work of refusal while remaining unanswerable as refusal.

The polite extremist

The New Statesman gave Jacob Rees-Mogg that description in 2018: old-school charm and civility, doing work his critics say the policies themselves could not do alone, functioning as cover for positions on welfare, austerity, and abortion that a blunter register would have exposed to challenge far sooner. Rees-Mogg has said publicly that heckling is good for politicians, and is on record elsewhere resisting displays of emotion as a standard he holds himself and others to. Both positions share a structure: composure, offered as the standard everyone else should be held to, from a position that costs him nothing to maintain and asks everything of whoever is expected to meet it.

The same mechanism operates at national political scale. The register a politician speaks in is itself a form of argument, not neutral packaging around one. A courteous voice describing a severe policy borrows credibility from the courtesy and lends none back. The listener has to do the work of separating tone from content, and most listeners, most of the time, do not, because the tone registers first and the content arrives already pre-approved by the manner it travelled in.

A different version of the same trade

Not every use of the mechanism is extremism in disguise. Some of it is simply emptiness, graded generously because it arrived in a pleasant register. Andy Burnham’s first weeks as Prime Minister were widely reported as a success on the strength of tone: informal, communicative, confident enough to speak without a lectern, an upbeat mood in Downing Street. Commentary at the time noted the substance running well behind the manner, few detailed policy commitments, and one assessment observed directly that his early actions had not matched the scale of his own rhetoric.

None of that makes the warmth dishonest. It makes it a currency, spent the same way in every institution in this piece, exchangeable for something a stated position would have had to earn on its own terms. A bold vision can be interrogated, argued with, held to a date. An affable manner can only be liked or not liked, and liking is not a standard anything gets measured against.

“Some of my best friends”

Sympathy is cheap to express and expensive to act on, and the gap between the two is where a great deal of prejudice hides in plain sight, dressed as kindness.

I am deaf. I have lost count of the number of times someone has told me, warmly and unprompted, how sorry they are, how hard it must be, how well I cope, in the same exchange where the accommodation that follows is to shout, louder, at me, without ever checking whether what they said has actually landed. Volume is the accommodation people reach for because it costs them nothing to perform and asks nothing of them afterward. It does not require facing me. It does not require confirming I understood. It requires raising the voice and considering the job done, whether or not anything was actually communicated.

There is an older, cruder version of the same move. Some of my best friends are black, offered as a defence against a charge of racism, works on the same logic: a visible gesture of effort presented as proof against a practical pattern, as though the two sat on the same ledger and one could cancel the other out. Nobody who shouts at me thinks they are failing to accommodate me. That is exactly why it functions as a failure that never gets corrected. The tax runs in both directions. I work harder to extract meaning from noise not built for me to parse. They consider the account settled by having visibly tried.

The asymmetry in institutions

A bishop expresses great upset about a decision he holds the power to reverse, and the upset stands in for the reversal. Two people close to a report on clerical abuse listen with real warmth and offer nothing that can be checked against later, and the warmth is received, by them, as the response rather than the absence of one. A regulator’s letter uses the word cannot twice, three months apart, about a complaint it never appears to have read the second time, and the courtesy of the letter’s tone does the work that its reasoning does not. A charity’s chief executive signs his emails with the word his own organisation was built on, then reaches for the word disappointed the moment that same organisation’s language is turned back toward him rather than deployed on his behalf.

None of these are impolite. That is the entire point. Each one would be easy to argue with if it were rude. None of them are rude, so none of them can be argued with, only felt, and feeling is not a form institutions are obliged to answer.

Where the pattern holds under pressure

I have spent years inside processes where the self-represented, or whoever holds the least resources, watches fairness used the same way, as a word placed in front of a process to describe what the process is not doing. Fairness, spoken by whichever party holds more resources, more representation, more procedural comfort, is not a claim about outcome. It is a claim about tone, offered instead of an account of consequence.

I recognise the mechanism because I have now watched it operate identically in a diocese, a charity, a regulator, a court, and in conversations with strangers who meant well. The word changes. Fair. Brave. Cannot intervene. Great upset. Disappointed. So sorry to hear that. The function does not. Each word is available to the party already holding more power, and each word makes the less powerful party’s directness look like the thing that broke the courtesy, rather than the thing that finally named what the courtesy had been covering the whole time.

What this asks of the reader

Brown and Levinson thought they were describing manners. They were describing a ledger. Every polite phrase in an unequal exchange is a payment, and the question worth asking of any courteous sentence, institutional or personal, is not whether it is kind. It is who was required to pay for it, and who got to keep the account.

Systems on Trial examines the machinery beneath the narrative. ianmcniff.substack.com

“I cannot do more than repeat what I have said previously that investigating complaints and holding members to account is not our role and we do not have the resources to do it anyway. If you have complaints against a firm, whether or not an MBC signatory, then there are of course other avenues you can pursue.”

The chief executive of the Mindful Business Charter (MBC) signs his emails Be Brave. That is not incidental phrasing. It echoes the charter’s own founding language, which commits signatories to give each other permission to be brave. He closed a message confirming his organisation will not check whether a signatory kept that commitment, with the word the commitment is built on.

What was already built

I wrote about the Mindful Business Charter in April, after receiving a letter from an MBC signatory law firm, threatening me with a penal notice as a litigant in person, carrying a footer committing the sender to reduce avoidable stress at work. The charter’s chief executive had been helpful before that. He sent me the organisation’s own guidance for litigation and dispute resolution, dated January 2024, written by a taskforce of signatory representatives. He is named among the taskforce himself.

The guidance describes itself honestly. It is, in its own words, “deliberately a permissive framework, as opposed to a prescriptive rulebook.” It does not claim to enforce anything centrally. That is not the gap.

The mechanism that exists, and who it is built for

The guidance does contain an accountability mechanism. It anticipates practitioners calling each other out when the guidance is not being followed, and it asks that this happen, and be received, courteously, on both sides. That is a real structure. It assumes two represented professionals, opposing counsel to opposing counsel, holding each other to a shared standard they both signed up to.

It has no channel for the person the letter was sent to, if that person is not a practitioner. A litigant in person is not inside the peer relationship the guidance describes. There is no represented opponent on my side of the correspondence to do the calling out. So I raised it with the organisation itself instead. That is when I was told investigating complaints is not the organisation’s role.

The guidance’s own fourth principle states that an opponent is a human being, worthy of respect, whether or not they are the one holding the pen. I am the human being the letter was addressed to. The mechanism built to protect that principle was never built to hear from me.

Outside the mechanism entirely

I have written before about a ninety-two year old woman in a former mining community who left no data trail an algorithm could find. Not because she had nothing to offer. She had fed families through a strike and held a community together through consecutive crises. The infrastructure that trains machines on what was said, posted, recorded, searched, shared never reached her, so there was nothing there for it to learn from. She was invisible to the system not because it judged her and found her wanting. It never registered that she existed as a category of input at all.

The MBC’s guidance is a smaller version of the same architecture, built for a different purpose. It was designed around represented professionals speaking to each other, and it works, as far as it goes, exactly as intended, for that population. A litigant in person is not filtered out of it the way a grant application gets filtered out by a keyword list, screened, then rejected. There is no active decision anywhere in the guidance to exclude someone who is unrepresented. There does not need to be one. The category was never built to notice them.

That is the more durable form of exclusion. A rejection can be appealed, because a rejection is a decision, and a decision has somewhere to be taken. An absence has nowhere to go. Nobody excluded me from the mechanism the guidance describes. The mechanism was simply never drawn wide enough to include the person actually holding the letter.

The same theology, secularised

Even where a mechanism does reach someone, being heard is not the same as being able to compel a result. The Cross of the Moment (Durham University’s report on clerical abuse in the Catholic Church, examined in an earlier piece) depends on an institution reaching the conclusion, on its own, that it needs forgiveness. Nobody can compel that conclusion. It can only be invited, and the invitation can be declined indefinitely without any consequence attaching to the decline.

The MBC’s guidance depends on the same structure, moved into a professional register. Calling out a peer for non-adherence works only if the peer being called out concludes, unprompted, that they were wrong. Nothing in the guidance obliges that conclusion. Nothing follows if it doesn’t arrive. The guidance can be cited, gently, and received, courteously, and rejected entirely, and the mechanism has still functioned exactly as designed, because the design never included a way to be wrong against your will.

Both systems call this restraint. A report that will not compel conversion protects the sincerity of whatever conversion eventually happens. A charter that will not compel acknowledgement protects the voluntariness of whatever acknowledgement eventually happens. In both cases what actually gets protected is the institution’s freedom to conclude nothing, indefinitely, while looking, throughout, like a body engaged in serious reflection.

Why the queue never shrinks

There is a third layer underneath the first two. Even a complaint that escapes both the mechanism’s blind spot and the requirement for voluntary acknowledgement, by reaching a body with an actual complaints process, meets a structure with no incentive to close it quickly.

The MBC is funded by its signatories. Membership costs between £500 and £2,750 a year, scaled to workforce size, paid by the firms whose conduct the charter’s guidance describes. The Solicitors Regulation Authority (SRA) is funded the same way in principle, if not in name. Sixty per cent of its income comes from law firms, charged on turnover. Forty per cent comes from individual solicitors’ practising certificate fees. The SRA has proposed a 29% increase to its total funding for 2026/27, to £111.5 million, citing the cost of investigation and intervention work. The regulator is paid for entirely by the profession it regulates, and is now asking that profession for more.

I reported a signatory firm’s conduct to the SRA in November 2025. The response found a potential breach of the SRA’s own rules, judged it “minor or inadvertent in nature,” and closed the case. I reported again on 28 November, stating explicitly that this was a new report, not a resubmission of the first. The SRA disagreed. Its response, dated 13 February 2026, states it found no new information in the second report, then closes it on the same grounds as the first, the same phrase, minor or inadvertent, used again three months later. Neither decision carries a right of appeal. Both promise to keep the report on file, in case a pattern emerges from future complaints. Nothing is required to happen before then.

A diocesan bishop said, in a statement reported by the press and quoted in “The Bishop Who Cannot Intervene,” that he could not intervene in a trust’s financial decisions. The SRA’s February letter, summarising its own earlier reasoning, told me directly it “cannot intervene in legal disagreements.” Two different institutions, two different registers, the same word, one reported to the world and one addressed to me by name. It does not matter, from where I am standing, whether each boundary is honestly drawn or not. What matters is that every boundary, honest or not, resolves the same way.

The lines are rehearsed, not reasoned

Neither dismissal engaged with what was actually in front of it. The bishop’s statement that he could not intervene in the trust’s financial decisions did not address the Catholic Schools Inspectorate framework that reports to him, which lists chaplaincy provision as evidence for judging a school’s Catholic life and mission. The reason to intervene was sitting inside his own Church’s machinery. The statement did not mention it.

I was not asking the SRA to intervene in the underlying legal dispute, the proceedings themselves. I was asking it to review a firm’s conduct against its own conduct rules, which is the reason the SRA exists. Recasting a conduct complaint as a request to adjudicate a legal disagreement is not a neutral misreading. It is the same move used throughout legal practice whenever conduct needs to go unexamined: declare the matter ongoing, or otherwise outside the regulator’s remit, and whatever happened inside it stops being something anyone has to look at. The most extreme version of this logic is the so called super-injunction, where the fact that something cannot be discussed is treated as though the thing itself does not exist. The SRA’s letters run on a milder version of the same premise. Reclassify the complaint as something the regulator cannot touch, and the conduct inside it is treated as though it was never raised.

The second letter compounds this. It said it found no new information in the 28 November report, one I had told them explicitly was new not a resubmission, and which ran to pages of material I had independently verified and evidenced. The letter’s finding did not describe what was in it. It restated the earlier finding, in the earlier words, as though the second report had not been read.

I have written before about an algorithm that screens research grant applications for 235 disfavoured terms, filtering out sound science because two words appeared on a list. A word-matching filter and a form-letter dismissal are the same mechanism wearing different clothes. Neither weighs what is actually submitted. Both retrieve a stored response and attach it to whatever arrived.

What this is not

This is not the pattern I described in “The Bishop Who Cannot Intervene,” where a bishop said he could not intervene in a decision he in fact retained the power to reverse. The MBC’s chief executive is not claiming a power he holds and declining to use it. The charter was genuinely built as permissive, and he is accurately describing that design. The failure is not a broken promise. It is three separate failures stacked on top of each other: a mechanism that never registered me, a theology that cannot compel acknowledgement even where it does, and a funding model that profits from the queue behind both.

What is left

A guidance document that tells practitioners their opponent is a human being worthy of respect, and then builds its only accountability mechanism for practitioners alone, has answered the question of who it considers a full participant in the dispute. An institution that will not compel acknowledgement, and a regulator paid for by the firms it regulates, have answered a second question between them: what happens to the complaints that arrive anyway. Nothing is required to change. Someone is required to pay for the queue to keep moving. That is not the same thing.

Be brave, the charter says, to the people who pay to belong.

Systems on Trial examines the machinery beneath the narrative.

ianmcniff.substack.com

If this piece is worth sharing, pass it on

Multi Academy Trusts (MATs) do not run on that design. St Ralph Sherwin Catholic Multi Academy Trust, founded by the Diocese of Nottingham in 2018, is a company limited by guarantee. It is £9.5 million in deficit, the largest of any of England’s 2,200 academy trusts. The Department for Education has issued it two notices to improve, the second in February 2026, for breaching the Academy Trust Handbook. Its own accounts record an unauthorised £200,000 loan from the diocese that founded it.

What was cut, and what was advertised

Fifteen lay chaplains served the trust’s 25 schools. All fifteen were told to reapply for three posts. One governor described a chaplain coming to him in tears. Chaplaincy is not funded by the Department for Education. It is a cost the trust chose to carry, and chose to cut.

While the reapplication process ran, the trust advertised a new post: Director of People, spanning all three Catholic Multi Academy Trusts (CMATs) in the diocese, salary up to £108,499. One trust employee was paid more than £150,000 in 2025. Fr David Cain, a parish priest in Leicester, called it “a reflection of the preferential option for the rich.”

A former governor put the comparison plainly. Chaplains were the adult a vulnerable child trusted most, the figure who could bring a pupil at risk of exclusion back into ordinary school life, who could reduce a child’s exposure to county lines gangs or radicalisation. No Director of People does that. The role manages staff. It does not sit with a fourteen-year-old who has nobody else to sit with.

What the bishop said

Bishop Patrick McKinney said the prospect of losing the chaplains had caused him real distress. He described the three trusts in his diocese as structurally separate from the diocese itself, and said the diocese “cannot intervene directly into operational matters, which include financial decision-making.”

He also has the power to appoint directors to the trust’s board. The diocese that cannot intervene is the diocese that founded the trust, lent it money, and appoints the people who run it.

That is not a contradiction the bishop failed to notice. It is the shape of the office now. Warmth is offered. Structure is not touched. I have watched this pattern in an archbishop, in a CEO, and now in a diocesan bishop facing a decision that sits entirely within his own gift to reverse.

Why I resisted

As Director of Schools in the Diocese of Clifton, I resisted Multi Academy Trusts as a vehicle before most dioceses had committed to them. The argument for MATs was always framed as protection: pool resources, gain scale, secure the schools’ future. What the argument left out was what got traded to buy that scale.

Local authorities held certain powers over school improvement, staffing frameworks, admissions support. A MAT absorbs those powers. To hold them, it must build the administrative machinery a local authority already had, HR structures, finance teams, compliance functions, at trust level instead. That machinery is secular by design. It has to be. Nothing in employment law or safeguarding regulation asks whether an HR director understands transubstantiation.

The paradox is this. The more autonomy a diocese claims from the local authority, the more of its own capacity it must spend running a local authority in miniature. Complete autonomy over spirituality and religious practice, the thing Voluntary Aided status was built to protect, gets less resourced, not more, the further the trust extends into secular administration. Nottingham employs one Director of People across three trusts on a salary just over £100,000. It employed three chaplains, after cuts, across twenty five schools. The administrative claim to autonomy was funded. The spiritual claim was not.

There is a second problem underneath the first. The Catholic Schools Inspectorate, which reports to the diocesan bishop on exactly this question, names chaplaincy provision as evidence for judging a school’s Catholic life and mission. The trust that cut chaplains from fifteen to three has weakened its own schools against the metric the Church uses to judge whether they are succeeding at being Catholic at all. A Director of People does not appear anywhere in that framework. The role manages staff, not mission. It cannot manufacture the thing inspectors are sent to look for, and neither can a board minute.

What I recognise in it

Patrick McKinney and I trained at the same seminary. St Mary’s College, Oscott. I arrived in 1978. He was ordained in July of that year and left. I do not know the man. I know the institution that formed him, because it is the institution that failed to form me into anything except someone who learned to navigate it rather than be shaped by it. Oscott teaches obedience to structure before it teaches anything else. That formation does not disappear when a seminarian becomes a bishop with a board to appoint and a trust he founded that cannot pay its own suppliers. It explains the statement perfectly. Distress, expressed. Structure, protected.

I warned against this trajectory before Multi Academy Trusts existed. Dioceses that resisted the structure early, mine among them, understood that consolidation trades local, personal, mission-tied governance for scale, and that scale answers to a board and a deficit line before it answers to a parish. Nottingham did not resist. Most dioceses did not. The status marker of running a MAT arrived, and arrived first, before the question of what it would cost the schools spiritually was ever seriously asked.

What the exchange has become

Voluntary Aided status was built on a specific trade: the community’s money for the community’s voice in the school’s spiritual formation. That trade has not been formally abolished. It has been hollowed out from inside. The diocese still appoints foundation governors. It still, on paper, protects religious character. What has actually happened in Nottingham is that the money moved to a company board, the chaplains moved to a reapplication list, and a new corporate role was created and funded in the same season the trust told its own priests it could not afford them.

Whether or not you think chaplaincy is the right form of pastoral support, it is a better vehicle for direct contact with a struggling child than a £108,000 HR post ever will be. One sits with the child. The other manages the adults around them.

The bishop said the cuts caused him real distress. His board minutes will show whether that distress changed anything.

A fourth track sits alongside the three already named. On 29 June, in a separate speech, Burnham proposed devolving employment support to metro mayors, run through organisations people trust rather than the Department for Education mechanisms already in place. Four tracks now. Curriculum content. Teacher training. Routing at 14. Employment support, devolved. Four clocks, four owners, no document connecting them.

Why the five before it failed

“The Remit Gap” named the mechanism in Plowden. Recommendations folded into what already existed, not used to redesign it. That is not one report’s failure. It is the pattern across all five, back to 1959.

The evidence that the pattern is still running is not theoretical. Just over 1.01 million people aged 16 to 24 were Not in Education, Employment or Training (NEET) in the first quarter of 2026, 13.5%, rising every year since 2021. Sixty seven years, five reports, one diagnosis restated five times, and the outcome the reports were meant to fix is worse now than when the reports began.

A second gap sits underneath the first one

Call it the application gap. It is not about whether the tracks talk to each other. It is about whether any of them, taken on its own terms, produces something a 15-year-old can use.

I have spent the past month writing revision guides for my grandson, who sits his General Certificate of Secondary Education (GCSE) exams next year. I read what the commercial market currently sells him. It is past papers. Nothing else. No rationale for why the topic is on the syllabus. No rooting of the subject in where it came from. No context for what the knowledge is for. No account of who tried to teach this before, what they got wrong, and why the current approach exists instead of the one it replaced.

A revision guide built entirely from past papers is not teaching a subject. It is teaching pattern recognition against a specific exam format. Get the pattern right and you pass. You have not been given the subject. You have been given the shape of the test.

“The Remit Gap” already showed this mechanism twice, in the Computer Science GCSE and the Travel and Tourism BTEC. Content sits in a specification. Nobody has been asked whether a student comes out able to do anything with it.

Burnham’s technical routes carry the same risk. Tes reported that the government has not yet specified how much technical content schools will teach directly, or whether it expects specialist teachers to lead the pathways. A policy that puts 14-year-olds into technical education without specifying content, without specifying who teaches it, without specifying what skill a young person carries out the other end, treats practical learning the way the revision guides treat academic learning. Work experience is not curriculum. An internship is not curriculum. Neither becomes curriculum unless it produces a specific, transferable skill, taught in a context the learner understands, assessed against something more than attendance.

Curriculum is not what is taught. It is what is learned and can be applied.

That distinction is the whole argument, not a refinement of it. A specification document listing topics is an input. A young person who can explain why the topic matters, where it came from, and what they can now do with it, is the output. Every track currently in motion measures the input and calls it reform.

This was not always misunderstood. The Statute of Artificers, given royal assent under Elizabeth I in April 1563, put apprenticeship on a statutory footing. Seven years, tied to a specific trade, state regulated rather than left to guild discretion. Seven years builds a skill because seven years is what the skill takes. A headline announcement with no specified contact hours is not the same undertaking, whatever the word technical is doing in the policy title.

Where this leaves the remit

Confirmed in writing on 6 August: four tracks, no mandate to connect them. The application gap says something harder underneath that. Even a government willing to connect all four tracks tomorrow would still be measuring exposure, not application, because none of the four has specified what a young person is meant to be able to do at the end of it. Sixty seven years and five reports have named the remit five times. None of them touched what learning was supposed to produce. The NEET figures are what that omission costs.

1963 The Newsom Report, “Half Our Future,” argues the country’s future depends on better education for pupils of average and less-than-average ability.

1967 The Plowden Report argues education should be built around the whole child, not fragmented into subjects. Implementation is piecemeal. Recommendations are folded into what already existed, not used to redesign it. The holistic principle survives in early years policy. It does not travel into secondary education, and it never reaches vocational education at all.

2003 Sir Mike Tomlinson proposes technical education routes from 14, tied to employer involvement and school-industry partnership.

2026 Andy Burnham announces the same structure.

Sixty-seven years. Five reports. One diagnosis, restated five times.

Research comparing people born in 1958 and people born in 1970 found intergenerational mobility fell between the two cohorts. No birth cohort since has recovered to 1958 levels. The reports started when mobility was still where it had been. It has not returned there since.

The numbers now. 1.01 million people aged 16 to 24 in the UK were Not in Education, Employment or Training (NEET) in January to March 2026, 13.5%, the first time over a million since 2013, rising every year since 2021. 39% of them are unemployed, both figures up on the year. Persistent school absence, missing 10% or more of sessions, ran at around 10% to 11% in the years before the pandemic. It is now running at 18% to 19%, down from a 2022/23 peak of 21.2% but nowhere close to where it was. Absolute income mobility is falling. Long range upward mobility is declining.

Four measures. All moving the wrong way since the reports began proposing fixes.

The country has not been short of reports. It has been short of one that connects to the others.

Three tracks, three owners

Three pieces of work are running now on vocational and academic education. None speaks to the other two.

Curriculum content. Professor Becky Francis led the first full national curriculum review in over a decade, reported November 2025. Scraps the English Baccalaureate (EBacc). Replaces the current General Certificate of Secondary Education (GCSE) in Computer Science with a broader Computing GCSE. Adds a Level 3 vocational pathway, V Levels.

Teacher training quality. The Initial Teacher Training (ITT) market review, 2021. Restructured accreditation and provider standards. Did not touch curriculum content. Not revisited by Francis’s panel.

Vocational and academic structure. Burnham, 2026. Routing, not teaching. References V Levels only in passing.

Each review is competent within its own remit. That is the mechanism. No mandate covers the whole system, so no review has to.

Not inference. In the Commons debate on the Francis review, 5 November 2025, an MP put teacher recruitment into ITT to the Secretary of State directly. The reply: numbers are rising, but “there is more to do.” The same day, the Association of School and College Leaders’ general secretary, Pepe Di’Iasio, called physics recruitment into ITT “dire.” Francis herself, asked about a separate Department for Education (DfE) plan for a Year 8 reading test announced weeks before her own report, called it “very separate to what the government is suggesting.” Minister, sector leader, and review author: three different people, the same week, describing three pieces of the same system that do not talk to each other.

Where it shows: a computer

The current Computer Science GCSE content is largely unchanged since 2016. Francis’s review calls it “narrowly focused,” recommends a broader Computing GCSE. Publication of the new curriculum: spring 2027. First teaching: September 2028.

The lesson running in classrooms now, students handwriting the parts of a computer twenty years after their predecessors built PowerPoint presentations at home without being asked, is the current, official curriculum. The agreed fix is three years away. Computing At School reports many schools lack the staff capacity to teach the subject well regardless of content. That is ITT’s territory. It was not asked.

Where it shows: a holiday

Tourism is the vocational subject meant to prove relevance is possible. The Level 1/2 qualification, Pearson’s BTEC Tech Award in Travel and Tourism, has a component called Global Influences, covering economic and political factors, natural disasters, and case studies. It names consequences. It does not examine the infrastructure producing them: water and waste capacity, housing displacement from short-term lets, the night-time economy as its own labour and safety system, who takes the margin from visitor numbers and who carries the cost.

A vocational course that taught the trade would leave a student able to ask who benefits and who pays. Naming that a place has economic and political factors is not that question. The course was never asked to teach it.

What drives it

The quality of teaching is what makes any of this work. Relevance and immediate application are what make a student want to study something, and that is also what produces curiosity, not the reverse. Parity of esteem between vocational and academic routes does not follow from saying it. It follows from both routes being taught well, by people equipped to teach them, in subjects built to connect to something the student can use. None of that is what any of the three current reviews is set up to measure.

What the student experiences

One borough’s data, not a national figure, but the gap it shows is the relevant one. Tower Hamlets’ Pupil Attitudes Survey, 2022: 41% of secondary pupils said most of their teachers made lessons “fun and interesting,” up from 28% in 2017, still under half. 69 to 70% of the same pupils said school gave them “useful skills and knowledge.” Pupils will credit the institution with usefulness while finding the lesson itself disengaging.

Twenty years, one proposal, no mechanism

In the National Strategies, twenty years ago, I raised two ideas: recognition and support for students who are also carers, and micro credits for students managing education through difficult circumstances. Both dismissed. Neither has resurfaced in Francis, in the ITT review, or in Burnham’s announcement. None of the three current tracks has a slot for a proposal that does not sit inside one of their three separate remits. That is not oversight. That is what a system without a connection between its own reviews produces, by design.