This is a chapter about a bayonet. I am paying for it. So are you. It is pointed at people who are already down, and it makes everything worse, and it feels good to some people, and that is why we keep doing it.
Nobody chose to be born in Syria. Nobody, so far as anyone can establish, decides to become a sex offender — the evidence on this is substantial and largely ignored, because it is inconvenient: patterns of sexual compulsion are laid down early, are neurological in origin, and are not straightforwardly a matter of will. Nobody chooses poverty. The Syrian child did not select their country of birth from a list. The poor person did not opt for the circumstances that produced their poverty. And yet we have built an entire moral and legal architecture on the premise that people deserve what happens to them — that punishment is justice, that suffering is consequence, that the infliction of pain is the appropriate response to the pain that circumstances, neurology, or geography have already inflicted.
Martha Nussbaum, the philosopher who has written most carefully about this, calls it backward-looking anger. The desire to see someone suffer in return for harm done. She is not sentimental about it — she acknowledges that the desire is deeply human, almost universal, and very hard to give up even when you know it achieves nothing. The harm is already done. The suffering of the person you punish does not undo it. We know this. We want the punishment anyway. And the wanting, she argues, is what we need to examine — because it is not really about justice. It is about the feeling of justice. The hit.
In June 2026, bodycam footage was released showing Henry Nowak, eighteen, handcuffed and dying on a Southampton pavement. He had been stabbed by Vickrum Digwa, who then told police he was the one who had been racially attacked. Officers believed him. They arrested the white teenager bleeding to death and let his killer walk free of the scene. Nowak kept telling them he had been stabbed. Nobody listened until it was too late to matter. When the footage came out, several hundred people gathered outside the police station, then marched toward the murder scene and the street where Digwa’s family lived. Twelve officers were injured. Farage called it two-tier policing, which, for once, in this one narrow instance, was not entirely inaccurate — though not in the direction he meant. The system did fail Nowak, and did so in a way his death makes impossible to argue with. That failure deserved investigation, accountability, reform. It got a riot instead.
Hampshire police released the data under freedom of information law. Of the thirty-seven people arrested during the disorder, twenty-two — fifty-nine percent — had previously been reported to police for domestic abuse. Harry Varney, thirty-four, was filmed punching riot shields; he got three years. In 2023 he pinned his partner against a bedroom wall and bruised her neck, then told her she had deserved it from a previous abusive partner, and kept threatening her afterward. She told the court she now cannot be left alone with a man without feeling scared. Dillon Crawford, twenty-nine, threw a bin and a metal chair at police; also three years. He has nineteen previous convictions, including one incident in which he punched a partner unconscious, broke her teeth, and told her, when she woke, that he had put bleach in her hair. Andrew Summerhayes, thirty-eight, threw a wheelie bin at officers; three years and two months.
This is not new, and it is not confined to Southampton. After the riots that followed the Southport murders in 2024, the Guardian obtained data from twenty-one police forces: forty-one percent of the nine hundred people arrested had a prior domestic abuse report. In Bristol it was more than two in three. A separate, later count found that one in five of the original Southport cohort had gone on to be reported for intimate partner violence since — in Cumbria, more than half. Two incidents, two years apart, triggered by entirely different grievances — one an act of mass murder blamed on the wrong people, the other a genuine and specific failure by the police who were meant to protect a dying boy — and in both cases, the people who turned out in the street to demand a reckoning were, in large numbers, delivering an identical reckoning at home, to people who had done nothing at all except be nearby. Farah Nazeer of Women’s Aid put it plainly: it is not surprising that someone violent in public is also violent in private. There is a proven correlation. What she did not need to add is that this is the same population, showing the same face, in both settings. The bayonet does not know it is pointed at a stranger or a partner. It only knows it wants to be used.
Kevin Carlsmith ran an experiment that should trouble anyone who has ever felt certain that punishing a wrongdoer would make them feel better. It gave people the chance to punish someone who had genuinely wronged them, in a controlled setting, and it asked them, before and after, to predict and then report how they felt. They predicted relief. They got the opposite — participants who punished felt worse afterward than those who did nothing, and kept dwelling on the offence for longer. The anticipation is where the pleasure lives, not the act. Brain imaging backs this up: the fantasy of revenge lights up the same dopamine circuitry as any other reward, well before anything has actually been done to anyone. The riot, the punch thrown at a police shield, the bin hurled at an officer — these deliver the hit Nussbaum described. They do not deliver justice, and they do not even reliably deliver the relief they promise the person committing them. Gabrielle Adams and Elizabeth Mullen found something worse: once a wrongdoer has been punished, bystanders become less willing to help the actual victim, because the punishment itself is mistaken for the repair. The suffering of the offender stands in for the restoration of the person who was harmed. Nobody has actually been restored. Everyone has agreed to feel that they have.
Henry Nowak is dead. Nothing that happened in Southampton that evening changed that, or made it more likely that the next person in his position gets believed. The domestic partners of Harry Varney, Dillon Crawford, and Andrew Summerhayes were not made safer by three men going to prison for throwing bins at police officers — they were made unsafe, repeatedly, by the same men, before any of this happened, and the state had already been told. Percy does not need people to be evil for this to work. He only needs them to feel, for one evening, that something has been done — and to go home afterward unchanged, to the person who already knows exactly what he is.
If punishment made anyone whole, this would be a different argument. It does not. Punishing Harry Varney does not unbreak his partner’s neck where he bruised it. Punishing Vickrum Digwa does not unbreak Henry Nowak, who is dead, and stays dead, regardless of what happens to the man who killed him. Punishment doesn’t unbreak the jaw. It doesn’t unbreak the window. It doesn’t bring anyone back. What it does is cost money — forty-seven thousand pounds a year, per person, paid by everyone, including the people who never wanted it and don’t believe it works. There is a case for honesty here. If the desire to punish is real, and deeply felt, and Nussbaum is right that it cannot simply be argued away — then let the people who feel it pay for it directly. A punitive supplement, opted into, funding the prison places of those who want them filled. And a rebate, for the rest of us, redirected to the things that the evidence says actually work — housing, treatment, restoration — so that our taxes stop being spent on a feeling we don’t share and never asked for. Let the desire for retribution show up on a bill, itemised, chosen, paid for by the people who want it. See how many people still want it, once it’s theirs to fund.
There should be a form attached. Before anyone is permitted to pay their punitive supplement — before their money is allowed to fund another year of someone else’s incarceration — their wife, husband, or domestic partner should be required to sign a sworn affidavit stating that there is no history of violence between them. Not a formality. A real declaration, with real consequences for perjury, filed alongside the tax return. It should feel invasive, because it is exactly as invasive as the thing it is checking for. Given what the Hampshire and Southport data actually show — that the people most inclined to demand punishment in public are, in significant and repeated numbers, the same people delivering it in private — this is not a bureaucratic flourish. It is the only honest gatekeeping the scheme could have. Let the person who wants someone else’s jaw broken by the state first prove they haven’t broken anyone’s jaw themselves. See how many affidavits get signed. See how many don’t.
In 1834, the Poor Law Amendment Act remade the system of relief for the destitute in England and Wales. Before it, the dominant form of support was outdoor relief — money, food, or goods given to poor people to survive in their own homes and communities. It was imperfect. It was sometimes abused. It was also, by the measures that mattered, cheaper than the alternative.
The alternative was the workhouse. The new Poor Law commissioners chose it deliberately, explicitly, and in full knowledge of the costs. Their reasoning was not economic. It was moral. Making poverty comfortable — allowing the poor to remain in their communities, supported, dignified — was considered dangerous. It rewarded failure. It encouraged dependency. It did not punish the poor sufficiently for the crime of being poor.
The workhouse was more expensive to build, more expensive to run, more expensive to administer than outdoor relief. It separated families. It broke communities. It produced misery on an industrial scale. Dickens wrote about it. Marx wrote about it. The commissioners knew what they were building. They built it anyway, because the suffering was the point. Not reform. Not efficiency. Punishment. The poor were to be made to feel their poverty as a judgement, not merely as a misfortune.
Percy did not manufacture this cruelty. He found the instinct for it — real, widespread, deeply human — and he institutionalised it. He gave it an Act of Parliament and a set of architects’ plans and a budget line. And then he charged it to the taxpayer, who paid more for the punishment than the compassion would have cost, and called it sound governance.
The prison population of England and Wales stands at around eighty-eight thousand. The average cost of a prison place is approximately forty-seven thousand pounds a year. Reoffending rates hover around fifty percent within a year of release and higher over longer periods. The evidence on what reduces reoffending is not obscure — stable housing, employment, maintained family relationships, treatment for addiction and mental illness. None of these things are primarily delivered by prisons, and some are actively prevented by them.
We know this. We have known it for decades. The alternatives — community sentences, restorative justice, drug treatment programmes, housing-first approaches for the homeless offender — are cheaper, and where they have been tried seriously, they work better. They do not feel like punishment. That is why we do not choose them. The person who has been wronged, or who fears being wronged, does not want the offender housed and employed and reconciled with their family. They want them to suffer. The suffering is the point. And Percy, ever attentive to what people want to feel, builds the system around the feeling rather than the outcome.
Between 2016 and 2020, women were imprisoned in England and Wales for not paying their television licence fee. Not many — three at any one time, typically — but enough to make the point. Picture who those women probably were. Not the imagined scrounger of the tabloid imagination, but the more likely version: young, not much money, a television on in the corner because there are children who need occupying while she tries to get something done — some work, some homework, something. We took her away from those children. We damaged those children — in ways that will not appear in any single budget, but will reappear, eventually, in health, in education, in the criminal justice system, in the next generation’s incarceration rate. The licence fee she didn’t pay was under £170. The prison place cost forty-seven thousand pounds a year. The question of what we were purchasing does not have a satisfying answer. And then there is the talent. The thing we don’t count because we never see it — what she might have done, made, contributed, become, if we hadn’t spent forty-seven thousand pounds a year destroying the conditions in which any of that was possible. The bayonet has a worker on each end. It hurts the person on the sharp end. And it hurts the rest of us too — not just in the tax bill, but in everything we never got, from people we never gave a chance to.
Joint enterprise law allowed — and in some cases still allows — a person to be convicted of murder for being present when someone else committed it. Charlotte’s brother received nineteen years. He was a member of the same gang as the person who did the stabbing. On the day in question, he threw a mobile phone at, and punched, another member of the gang — not the person who was stabbed and killed. That was the extent of his act. The rest is supposition: he should have known his friend had a knife, because young people in gangs tell each other everything, or so the prosecution supposed. He should have known. He is serving nineteen years for what he should have known. The Supreme Court ruled in 2016 that the doctrine had been misapplied — that mere foresight was not enough, that the law had gone wrong. Wrong, in the Court’s own word. That ruling did not empty the prisons of the people convicted under the wrong law. They are still there. The question — what is this achieving, what are we purchasing, at this price, for these people — does not have a satisfying answer. Only the feeling. Something has been done.
There is a particular cruelty in the disenfranchisement of prisoners — the removal of the right to vote from the people with the most direct experience of the criminal justice system, and therefore the strongest interest in its reform. Britain is one of very few democracies that does this as a blanket policy rather than as a specific judicial penalty. The European Court of Human Rights has said, more than once, that this blanket ban is unlawful. Britain has ignored it. The people inside the system are not permitted to participate in the system that put them there. Percy wrote those rules too.
One Saturday evening I was the only passenger on a forty-seat bus returning to Sojuela, a small village in La Rioja, near Logroño, in northern Spain. The bus ran anyway. The driver, who had been born in Romania, told me it was paid for by the EU.
Sojuela is largely populated by old people. It has a town hall and a bar. During the week it is served by public transport. Bread comes every day. Fish comes on Fridays — to the square, not to your door — and people come out to meet it, and while they wait they talk. They are in the same place at the same time as people they did not specifically choose to be with. That unchosen togetherness is not inefficiency. It is the thing itself.
In Britain, the supermarket delivers to your door. A man comes. He goes. You sign for your groceries on a screen. Nobody waits in a square. Nobody chats. The neighbour you didn’t know was struggling remains unknown and struggling, until the situation becomes serious enough to generate a care bill or a hospital admission or a social services referral, at which point it is logged in a different budget from the one that cancelled the bus.
Danny Dorling, the geographer who has spent a career following money to where it actually goes rather than where it is said to go, points out that it is cheaper to keep the village alive than to let it die and pay for the consequences. The bus that runs to Sojuela whether or not anyone is on it is preventive spending. It keeps the old woman in her home, connected, purposeful, part of something. Remove it and she eventually needs the care home, the ambulance, the GP appointments that are really loneliness presenting as symptoms. The saving from cutting the bus appears in the transport budget. The cost appears in health, in social care, in housing, in the criminal justice system when the isolation produces the crisis that produces the incident. Different departments. Different ministers. Different accounting periods. The bus gets cut and everyone congratulates themselves, and the costs accumulate quietly elsewhere, and nobody joins them up.
Spain is not more equal than Britain. On some measures it is less so — the concentration of land in the south, the old latifundia estates, a one percent that is deeply entrenched. I did not know what latifundia meant until I looked it up. The word is Latin — broad farm, broad estate. It refers to the enormous landholdings of Andalusia and Extremadura, distributed in vast grants to the nobles and military orders who fought the Reconquista, the centuries-long campaign to retake Spain from the Moors. The people who worked the land owned none of it. They were day labourers — jornaleros — hired when needed, discarded when not. It produced some of the worst rural poverty in western Europe and was one of the reasons the Spanish Civil War was so savage in the south. Franco’s victory preserved the system intact. It is the Spanish enclosures, achieved by a different history but the same logic: land that could support many, concentrated in the hands of few, by an act of power dressed as an act of God.
But Spain has something that Britain hollowed out and then forgot it had lost: a social infrastructure that belongs to no one and therefore to everyone. The paseo is free. The plaza is free. The bar in Sojuela where you can sit for two hours over one coffee without being moved on costs almost nothing. The culture of being together in public — visibly, daily, across ages and classes, without a specific purpose and without spending money — distributes something that does not appear in any Gini coefficient. Dignity, perhaps. Presence. The sense of belonging to something that is not a transaction.
Richard Wilkinson and Kate Pickett, in The Spirit Level, demonstrated that more equal societies do better on almost every measure that matters — crime, mental illness, health, life expectancy, social trust, imprisonment rates. The data is robust. Britain sits near the wrong end of almost every chart. The United States sits at the bottom. The argument is compelling and largely correct, but it needs one addition: what the Spirit Level measures is income inequality, the Gini coefficient, the ratio of top to bottom earnings. What it is really describing is social distance — the felt gap between people, the degree to which society is experienced as hierarchical, status-ridden, and threatening.
Spain confounds the prediction because it compresses social distance in ways the Gini coefficient cannot see. The old woman and the young man waiting for the bread van are not equal in income. But they are equal in the square. They occupy the same space, perform the same wait, exchange the same pleasantries. The ritual equalises them in the way that matters for trust and safety and the reduction of fear. Britain has high income inequality and high social distance. Spain has high income inequality and lower social distance. The Spirit Level is right about what matters. It just needs the square added to its model.
By every assumption of the punitive worldview, Spain should be in trouble. It shares a continent with Russia. It took in more immigrants faster than Britain. It has higher measured inequality. More threat, more diversity, more inequality — the logic says more crime, more fear, more need for a strong state. Spain’s incarceration rate is roughly 120 per 100,000 people. Britain’s is around 160 — one of the highest in western Europe. Spanish recidivism figures are significantly better. Spain did not choose the bayonet. It chose the square. And it got better outcomes, at lower cost, and called it Tuesday.
Britain does publish a composite figure: the estimated cost of a single murder, when you add investigation, prosecution, trial, and incarceration over the sentence served, runs to around two million pounds. Spain does not publish an equivalent number. What it does publish, separately, is instructive. The cost of a prison place in Spain runs to roughly twenty-three thousand euros a year — just under half the English figure. The homicide rate runs to roughly 0.69 per hundred thousand people, against Britain’s 0.95 — about a third lower per head of population. There is no Spanish equivalent of the two-million-pound figure because Spain has never needed to assemble one: the political incentive to dramatise the cost of crime, to make the bayonet feel expensive and the square feel free, has never been strong enough there to generate the accounting. The number exists in England because England wanted it to. The absence of the number in Spain tells you something about what Spain wanted instead.
The Bibby Stockholm — the barge moored in Portland Harbour to house asylum seekers — cost more to operate than housing the same people in the community would have done. It was chosen anyway, because it looked like punishment, and punishment looks like seriousness, and seriousness is what politicians reach for when they have no intention of solving the problem. Spain, in the same period, was actively recruiting immigrants to fill genuine labour shortages, offering clearer pathways to citizenship, running the bus to Sojuela whether or not anyone was on it. Spain’s economy in 2024 was the fastest growing in the EU. The Economist named it the best performing in the OECD. The driver who told me about the EU funding was part of that economy. He came. He works. He drives the bus that keeps the village alive.
There is a belief, occasionally encountered in Britain, that Spain has no social security system — that its social cohesion is the product of some kind of Mediterranean self-reliance rather than public investment. This is wrong. Spain has a universal public health system, the Sistema Nacional de Salud, free at point of use and rated among the best in the world. It has a public pension system, unemployment benefit, social services, and since 2020 a minimum vital income — a basic income floor for those with nothing else. Public spending on social protection runs at around a quarter of GDP, broadly comparable to the EU average. What Spain does not have is a single centralised national bureaucracy — healthcare is devolved to the regions, each running its own system. That may be the source of the confusion. But the coverage is universal and the spending is real. The paseo and the welfare state are not in competition. They are part of the same settlement: the idea that people are owed something by the society they live in. Britain chose the Bibby Stockholm. Spain chose the minimum vital income. Spain’s economy grew faster.
In March 2026, the United States and Israel launched a war on Iran. Spain closed its military bases at Rota and Morón to American forces. Then it closed its entire airspace to any US military flights connected to the conflict. Defence Minister Margarita Robles called the war profoundly illegal and profoundly unjust. Trump threatened to cut all trade with Spain. Spain held its position. France and Germany had opposed the Iraq war in 2003 but allowed US jets to fly over their airspace anyway — there are practices between allies, said the French foreign minister, that must be respected. Spain did not respect them. The same country that runs the EU-funded bus to Sojuela, that takes in the immigrants and puts them to work, that chooses the square over the bayonet — that country looked at an illegal war and said: not through our sky.