Liberals and their institutions own this failure. They built DEI, funded it, defended it, extended it into every institution they touched, and called anyone who questioned it an enemy of equality. The right despised it, so liberals doubled down. DEI became a litmus test, a line in the culture war, and that made it nearly impossible to say what needed to be said: it was never built to cost them anything.

By liberal I don’t mean the left. I mean the heir to the Rockefeller Republican: comfortable on social questions, loyal to the economic order. Dukakis, Clinton, Gore, Kerry, Obama, Clinton again, Biden, Harris. Every one a protector of the arrangement. I also mean the professional class that built the DEI apparatus and lives off it — HR, the endowed chairs, the consulting. Some believed it. Some only sold it. This essay goes easy on the first. The second earned everything that follows.

The racial wealth gap has not moved in forty years, the same forty years in which diversity training became the liberal answer to structural inequality and racism. The Chief Diversity Officer (CDO) proliferated across corporate America while union membership continued its decline. A title was added. The structure was not touched.

It wasn’t that long ago, and it wasn’t safe.

Civil rights activists faced police dogs, fire hoses, and federal indifference. SNCC organizers were murdered in Mississippi for registering voters. The FBI ran a covert program — COINTELPRO — not to monitor Black leadership but to destroy it. Fred Hampton was assassinated in his bed by the Chicago Police Department and the FBI working in coordination, before the phrase “systemic racism” had been processed into a corporate onboarding module. Four students were shot dead at Kent State by the National Guard for protesting a war abroad. The movements of that era understood empire and domestic repression as one system.

They were not monolithic. They disagreed about strategy: integration versus separation, reform versus revolution, the ballot versus the street. But among their more radical currents — SNCC, the Black Panther Party, Malcolm X’s Organization of Afro-American Unity — the dominant question was not how to prove the structure took the problem seriously. It was how to dismantle the structure and rebuild it.

Malcolm X did not want a seat at the table. He wanted to know who built the table, who owns the wood, and why his people were expected to be grateful for the invitation.

Those were strategy arguments worth having. DEI managed by a CDO isn’t one.

DEI began as a hire. Through the 1990s, corporations and universities built the machine — a Chief Diversity Officer, an annual equity report, a training everyone had to sit through — a full generation before it had a canonical text. I was a manager at Coca-Cola until 1995. Women and minorities were moving into jobs that had only ever gone to men, the lawyers could see the backlash coming, and I sat through the training like everyone else.

It didn’t work. In 1999, four years after I left, Black salaried employees sued. In November 2000 Coca-Cola settled for $192.5 million, the largest race discrimination settlement in American history at the time. The centerpiece of the deal was a court-appointed task force, and one of its jobs was making sure managers received annual diversity training.

Coke’s lawyers did not invent that structure. They took it from Texaco, which settled its own race case in 1996 for $176.1 million and funded a task force to build the training and monitor compliance. The machine spread the way boilerplate spreads — through legal departments.

The training had been running for years before the lawsuit. The remedy was more training.

The exchange was simple. Take the language of the civil rights movements and hand it to human resources. The institutions bought it because it was cheap, and because the alternative asked something of them. The CDO does three jobs at once: proof of commitment, absorber of grievance, human shield against accountability. Black CEO representation in the Fortune 500 stayed near one percent through 2023, after decades of commitments and billions spent to change it.

Executives perform empathy with the assurance of people trained to perform it. The unconscious bias workshop runs from 10 to noon. Lunch is provided. At 1 p.m. everyone returns to the same structure they left at 9:59. The process became the point, and the process never ends.

There’s an explanation for this, and it isn’t new.

For generations, America’s institutions — corporate, academic, and political — have mastered the art of using the language of justice to keep the same people on top. The Nigerian-American philosopher Olúfẹ́mi O. Táíwò gave that move a name — elite capture — and a 2022 book, Elite Capture: How the Powerful Took Over Identity Politics (And Everything Else).

The boardrooms added a face or two while the rent soared. Universities issued equity reports while food deserts spread. HR departments held bias trainings while the unhoused froze on the streets. DEI did not cause these conditions. It gave cover to the institutions that could have changed them and didn’t.

Liberals needed a hero. Enter Robin DiAngelo, a white academic who could tell white audiences the problem was them. White Fragility became the canonical text of what John McWhorter named the third wave of American antiracism. Strip the vocabulary and a religion is what remains. Original sin, confession, penance. What it never offers is absolution.

Here is what DiAngelo argues: white Americans become defensive when confronted with racial discomfort, and that defensiveness is itself part of what keeps inequality in place. Any denial of racism is proof of guilt. Disagreement is evidence the framework is correct. You cannot rebut it. You cannot opt out. The only available response is contrition.

An argument you cannot lose does something else, too. It flattens more than 40 million Black Americans — urban and rural, immigrant and slave descendant, working-class and Ivy-educated, religious and secular, conservative and socialist — all of it reduced to a single job: producing white discomfort. Black people enter as occasions for white guilt and are asked for nothing else.

Disagreement became evidence, and evidence went in a file. People learned to perform assent in rooms where the alternative was a conversation with someone whose title contained the word inclusion. The framework asked everyone to interrogate their complicity except the people selling it.

An institution can buy the entire machinery of what it called racial reckoning and wake up the next morning with its org chart, its pay bands, and its endowment exactly where it left them. DiAngelo’s listed fee runs $14,000 an event; her agency puts the keynote range at $30,000 to $40,000. The workshops ran.

For the liberal professional class, the book was a credential you could display. White Fragility, spine out on the coffee table, staged beside the succulents. A second copy rode in the tote bag, cover angled just far enough to be read across the train aisle. Nobody had to change a hiring policy. They only had to be caught holding the book.

If Portlandia had written the DiAngelo-era DEI script, it would have been hilarious: elite liberals demanding accountability from themselves, dragging the rest of the country along in a parade of mass guilt. Anyone want to guess the dress code?

DiAngelo’s defenders talk about identity politics as though she invented it. It has a serious origin: the Combahee River Collective, a group of Black lesbian socialist feminists who issued a statement in 1977 that remains one of the most rigorous political documents in American history. Their argument was not that race and gender should replace class analysis. It was that race, gender, sexuality, and class are interlocking systems of oppression.

What happened to that vocabulary between 1977 and 2018 is elite capture. The words — identity, intersectionality, power — were pulled out of the class politics that produced them. By the late 1990s the CDO had turned all of it into a job description. By 2018 it was the premise of DiAngelo's book. For the Collective, identity was where the analysis started. For the apparatus, it was where the analysis stopped.

I walked into Emerson Middle School in 1965, the year Malcolm was killed. Twelve years old, out of a middle-class neighborhood a mile away and into a building that was half white and half Black. Emerson stood on the same block as Flint Northern High School.

King was murdered on April 4, 1968. The block erupted. Fights first, then police, then tear gas. I went out a first-floor window and ran. I was fifteen, and the school could not protect me.

Classes resumed a few days later. The doors were locked for control. Robert Kennedy was killed in June. Then came the summer of 1968.

In September, I entered Flint Northern. English, history, and current events classes all collapsed into the same discussion. A room of kids from every income in the city, with the racial history of the country sitting there with them. The English teacher said: “This is going to be a tough year for all of us. I won’t lie to you.” Nothing was resolved. Nothing has been resolved since.

I went to the University of Michigan in the early ’70s as a political science major, when those same fights were the whole air of the place. I don’t recognize the movement of those years in DiAngelo’s theatrics. Too many of us who knew said nothing while the corporate version captured and replaced it.

Adolph Reed Jr., a political scientist and a fixture of the socialist left, was called a reactionary by liberal academics for making this argument, and disinvited from the platforms that had hosted him. He was right anyway.

He wasn’t alone, and none of them came at it from the right. Barbara and Karen Fields spent a book on how a language of race gets used to explain inequalities that class produced. Cedric Johnson traced how Black Power hardened into a Black managerial class, and Touré Reed showed how antiracist demands got severed from the economic reforms that would have delivered. Táíwò named it. The left critique of DEI is older and deeper than the one that got it killed.

Reed’s thesis, plainly: substituting racial identity for class analysis does not serve working-class Black people. It serves Black elites and their white liberal allies, who pose as representatives of Black America while backing charter schools, microfinance, and diversity in elite institutions. Those lift the Black professional class and leave the Black working class and poor exactly where they are. DEI is that swap, running on a budget. It produces Black faces in high places.

In 1996, in the Village Voice, Reed described a first-term Illinois state senator: “a smooth Harvard lawyer with impeccable do-good credentials and vacuous-to-repressive neoliberal politics… the point where identity politics converges with old-fashioned middle-class reform in favoring form over substance. I suspect that his ilk is the wave of the future in U.S. black politics.”

That was thirty years ago. The senator was Barack Obama, and Reed had described the whole career before it began: form over substance, the identity story carrying the weight the policy wouldn’t. Liberals got the president they called cool and competent — and the bank bailouts, the record deportations, the drone wars. In 2024 the party ran the play again, on the theory that the biography would do the work the policy wouldn’t.

In September 2016, at the dedication of the National Museum of African American History and Culture, Michelle Obama hugged George W. Bush. The photograph went viral. Commentators called it bipartisanship. The illegal war, the torture program, and Guantánamo went unmentioned. Nobody found it strange.

I called the liberal an heir to the Rockefeller Republican. Obama said it about himself first. Asked in December 2012 whether he favored a socialist model, he answered: “The truth of the matter is that my policies are so mainstream that if I had set the same policies that I had back in the 1980s, I would be considered a moderate Republican.”

The 1935 Social Security Act left out agricultural and domestic workers — most of the Black workforce at the time. The GI Bill was written universal and administered locally, so Black veterans collected a fraction of what white veterans did.

The Federal Housing Administration decided which mortgages the government would insure, and its underwriting manual told appraisers to mark down neighborhoods where racial groups were mixing. Whole sections of American cities were cut out of the mortgage market for a generation.

Historians still argue whether that was animus, political compromise, or administrative convenience. The argument barely touches the outcome. Nobody’s discomfort built it. Legislation did.

Barbara Fields reverses the usual order. Racism didn’t grow out of race. Race grew out of racism, invented to explain an arrangement that needed explaining.

Which is why the people who own that arrangement would rather you examine your soul. What legislation built, legislation can take apart.

Jesse Jackson built the multiracial, working-class coalition that could have done it. The Rainbow Coalition ran in 1984 and 1988 on shared material interest. Rent. Healthcare. Jobs. Workers’ rights. Jackson won roughly seven million primary votes in 1988 on a platform that made no distinction between the white factory worker in Detroit and the Black sharecropper’s son in Mississippi.

Martin Luther King’s last campaign, the Poor People’s Campaign of 1968, organized across racial lines explicitly, because King had concluded the civil rights laws had done what laws could do and what remained was class: redistribution, guaranteed income, structural economic change. He was murdered six weeks before the encampment opened on the National Mall.

They did to his memory what they did to his politics: they kept the dream and dropped the redistribution. It isn’t only King. The same swap runs through Harriet Tubman, Helen Keller, Rosa Parks, and a Walmart Juneteenth ice cream — I traced the pattern in Once You See It.

The Rainbow Coalition built a national movement on almost no money. Set that against the best-funded version of the other approach.

In 2024 the New York Times asked a simple question about the University of Michigan’s diversity program: what went wrong. Here is what nine years and $250 million bought. Black undergraduate enrollment rose from 1,255 in 2016 to 1,796 in 2024 — a gain of roughly one percentage point, in a state that is 14 percent Black. The Black Student Union had demanded 10 percent enrollment for decades. The program ended in 2025 without coming close.

Michigan’s own climate surveys showed perceptions of inclusion falling among the very students the program was built to serve. The Black Student Union called its goals left unfilled.

This was not a poorly implemented good program. It was the largest, best-resourced DEI initiative in American higher education — the strongest possible test of the theory. It got the fullest trial and moved the number one point. When the data came back flat, the apparatus did not revise the theory. It issued a Year 1 Progress Report calling the work “successful and promising,” then closed, without ceremony, when the federal government told it to.

A quarter of a billion dollars. Nine years. One percent.

None of this is a mystery once you look at who was doing the embracing. Elite universities are run by the professional-managerial class, people whose standing rests on credentials and on being seen to hold the right opinions. For that class, identity politics is cheap. It asks nothing of the endowment, nothing of tenure, nothing of legacy admissions. Class politics asks everything. Harvard sits on an endowment near $53 billion, has resisted sustained union pressure to pay wages matched to the cost of living in one of the country’s most expensive cities. It also assigns White Fragility.

They bought the workshop because it was cheaper and asked nothing of them. It is cheaper to hire a Chief Diversity Officer than to restructure who holds power. That was always the whole transaction.

Real progress has a shape. The racial wealth gap closes. Black homeownership matches white. Incarceration stops tracking race. The schools in Jackson, Mississippi and East Los Angeles are funded like the schools in Scarsdale. Those are material outcomes, and they take real policy: redistribution, labor rights, housing policy, criminal-justice reform, universal healthcare. That requires political power, which requires a coalition large enough to win.

Zohran Mamdani ran that plan in New York City’s 2025 Democratic primary — rent freezes, free buses, free childcare, municipal groceries in food-insecure neighborhoods, higher taxes on the wealthy — and built a coalition that crossed racial lines on shared material interest. He did not run on a DEI platform. He ran on what the Rainbow Coalition was practicing in 1988, calibrated to 2025. He took the primary, then the general, and has been mayor since January.

When Trump signed DEI’s death certificate, he only confirmed what was already true: the party that built it lost. The Democratic Socialists of America put replacements on the ballot. Adam Bojak in Buffalo. Mo Brown in Syracuse, who took out a fourteen-term incumbent by about a hundred votes. Claire Valdez and Darializa Avila Chevalier in New York, who beat sitting House Democrats, one of them the chair of the Congressional Hispanic Caucus. They beat the machine, in its own primaries, on its own ballot line.

Kshama Sawant spent a decade on the Seattle City Council and left undefeated. She recently ran in Washington’s 9th Congressional District against Adam Smith, the ranking Democrat on House Armed Services.

The party of Adam Smith is getting its own death certificate.

The line that never changes: you build the coalition on what people share across race and gender and origin — the rent, the wage, the hospital bill — and you win it at the ballot box. It is the expensive thing. It is the only thing that has ever worked.

Póvoa’s hospital has an office that files end-of-life directives into the SNS, the national health system that will, if I filled out the form correctly, know what to do with me when I can no longer say.

The office was off the main entrance, and then off that, a room maybe eight feet by twelve, two desks, two computers, two women. The one who took mine was kind, and her English was thin, which is a hard combination for the subject at hand. I have no Portuguese. We got by the way people get by when they try.

I had filled out the form at home. I downloaded the Portuguese document and ran it through Google Translate to select from a list of ten or twelve items, each with a radio button, each a small door: yes to this, no to that.

The woman identified a mistake. I had agreed to scientific research and then a button later not. She reached for the whiteout. There is something about watching a stranger correct your position on your own death plans with a small brush of white paint. The room was getting warm. The thought crossed my mind that I should apologize, exit and return when I could express my death wishes without whiteout.

I wasn’t dying today. But now sufficiently embarrassed, I was eyeing the door. We were done; the document was entered into her computer. Success! I thought.

Later that day I received an email from the SNS. My directive had been received and was recorded. I can’t get that woman out of my mind.

In the year he was killed, a Harris poll put his public disapproval near 75 percent. J. Edgar Hoover called him the most dangerous man in America. King’s own board at the Southern Christian Leadership Conference tried to censor him over his opposition to Vietnam. By 1967 he had stopped asking for integration alone. He said what he wanted from pulpits and conference stages, in his own words: “a radical redistribution of economic and political power” — a “revolution of values,” not a dream deferred into a monument.

Nineteen years later, in 1987, Gallup had him near 75 percent favorable. The same number, flipped. Nobody vindicated the redistribution. The speeches are still there, on tape, in the archives. They just stopped quoting them. What survived into the national holiday and the marble memorial is the speech about content of character, cropped from the man who stood in Riverside Church one year to the day before Memphis and called his own government “the greatest purveyor of violence in the world.” The occasion was Vietnam. The argument was where the money goes — the war budget eating the poverty budget. Not a line for a T-shirt.

There’s a name for what happened between those two numbers, and it isn’t new. Herbert Marcuse named it in 1965, in an essay called “Repressive Tolerance”: a society doesn’t need to suppress a threat forever. It only needs to survive long enough to tolerate — even celebrate — a domesticated version of it. Marcuse’s own term for the mechanism: tolerance that “neutralizes opposition” — a tolerance that inoculates people against imagining anything better — isn’t tolerance at all. It’s a more efficient form of repression, because it doesn’t look like one.

Nobody vetoed King’s redistribution. Nobody had to. They just built him a monument that left it out.

Harriet Tubman is the most celebrated abolitionist in American memory — a face nearly put on the twenty-dollar bill, the subject of a Hollywood biopic, the honorific “Moses” taught in every elementary school. What gets taught far less is what she did on the night of June 1, 1863.

She guided three Union gunboats up South Carolina’s Combahee River, after weeks spent running a network of formerly enslaved boat pilots to map the Confederate mines in the water. At dawn, the boats sounded their whistles and more than 700 enslaved people ran from the rice fields to the shore. Tubman became the first woman in American history to plan and lead a major military operation. Seven plantations burned behind her that morning — the economic infrastructure of slavery, destroyed by military force, on her direction.

The Underground Railroad made her a legend: quiet, patient, one family at a time, for a decade. The raid made her something the country has never quite known what to do with — a Black woman who planned an armed assault on American soil and won. The elementary school poster picked one.

Helen Keller is the only person on this list whose story ends in childhood. The well-house in Tuscumbia, Alabama — water over one hand, W-A-T-E-R spelled into the other — is where the national memory stops, and The Miracle Worker made the stop official in 1962. She lived another eight decades. She spent most of them as a socialist.

She joined the party in 1909, backed Eugene Debs, spoke against the war from the stage of Carnegie Hall, helped found the ACLU. Appointed to a Massachusetts commission on the blind, she found the blindness wasn’t evenly distributed — it pooled in the factories and the tenements, in industrial accidents and the untreated infections of the poor — and she said who profited from the conditions that caused it. The Brooklyn Eagle, which had spent years praising her intelligence, explained that her socialism sprang from the limitations of her development. Her answer: the editor had complimented her until she came out for socialism; only then did he remember she was blind and deaf and “especially liable to error.”

The FBI kept a file on her anyway. The country kept the child at the pump.

Rosa Parks was the Montgomery NAACP’s secretary for twelve years before she refused to give up her seat. In 1944 the Montgomery NAACP sent her to Abbeville, Alabama, as lead investigator on the case of Recy Taylor, a sharecropper gang-raped by six white men. Parks turned the investigation into a national campaign, co-founding the Committee for Equal Justice for Mrs. Recy Taylor, with Du Bois and Langston Hughes signed on. She trained at Highlander Folk School — the Tennessee training ground for labor and civil-rights organizers — that August, four months before her arrest.

E.D. Nixon, the Pullman porter and union man who ran Montgomery’s NAACP, had already tried this once. Fifteen-year-old Claudette Colvin had refused to give up her seat nine months earlier, on the same buses, for the same reason. Nixon dropped her as the test case because she was pregnant and unmarried — not, in his own words, someone he “could win with.” He chose Parks on purpose, for exactly the qualities that would let the myth work later.

Historian Danielle McGuire’s phrase for what happened next: organizers built her into a “matronly ... symbol of virtuous black womanhood.” Twelve years of organizing, a rape investigation, a training course in nonviolent tactics — all of it dissolved into a story about a woman too tired to stand. The respectability was never an accident. It was the strategy. What got erased was that a strategy was ever needed.

The cycle used to take decades. It doesn’t anymore.

Juneteenth marks the day enslaved Texans finally heard they were free, two and a half years after the proclamation. It became a federal holiday in June 2021. Eleven months later, Walmart pulled a “Celebration Edition Juneteenth” ice cream — red velvet and cheesecake, a trademark symbol stamped next to the word, a white model in the matching T-shirt — after a backlash that named exactly what was missing. One viral reply said it in a sentence: no reparations, no voting protections, no end to the killings, just ice cream.

King took nineteen years. Tubman’s raid has taken a century and a half, and the country still hasn’t absorbed it. Keller watched it happen while she was alive — the film that froze her at the pump was in theaters six years before she died. Parks got no interval at all; her own movement built the myth while the boycott was still running, and the forgetting came later. Juneteenth took eleven months. The interval between the material claim and its gutted, celebrated substitute keeps shrinking.

Every name on this list made a material claim while it was still dangerous to make it. King wanted redistribution. Tubman wanted plantations burned, not commemorated on a bridge sign. Keller wanted socialism, not a miracle at the pump. Parks wanted six men prosecuted for what they did to Recy Taylor, not a statue in the Capitol for the day she sat down.

What survived, in every case, is the moral version — the dream, the miracle, the forgiveness, the legend — never the thing that made any of them dangerous in the first place.

Once you see the swap, you start looking for it everywhere. It doesn’t stay in the history section.

Cannabis gummies, however, are not available for purchase at the corner pharmacy. Or anywhere, really.

I live in Vila do Conde, in a country that looked at its catastrophic heroin epidemic of the 1980s and decided the problem was not criminal, it was human. It built treatment centers instead of prisons. The word “drogado” slowly, imperfectly, became “pessoa com distúrbio de adição.” For twenty-five years, Portugal has been held up as a model, with policy wonks descending on Lisbon to study it like a secular Lourdes. And yet: I cannot buy a gummy.

An hour north, across the Minho and into Galicia, there are cannabis clubs in Vigo. Not hundreds — that’s Barcelona. But they exist, they are open, and one of them is the Asociación Viguesa de Estudios sobre la María. The acronym spells AVE MARIA. María is slang for the product. So: a cannabis club in Catholic Galicia, named with a completely straight face for the devotional study of the Hail Mary. You ring a bell. Someone lets you in. The Atlantic is the same Atlantic you left behind. The policy is not.

This is Spain’s cannabis club model: private, members-only associations that collectively grow cannabis and distribute it among members. No storefront. No advertising. The legal theory is that collective personal cultivation is simply distributed personal cultivation. A legal fiction. It has held for a decade because no prosecutor has wanted the fight. The cannabis is, by all accounts, excellent.

Portugal looked at this and did not move.

I asked a local pharmacist about medical cannabis. She looked at me the way Portuguese pharmacists look at people who ask about medical cannabis — the way a librarian looks at someone who asks for a book that doesn’t exist, was never ordered, and whose author she suspects isn’t real.

There are no cannabis dispensaries in Portugal. Not one. The country legalized medical cannabis in 2018 — a prescription, a pharmacy, a physician who must certify that you have exhausted every other treatment option first. The bar is not self-certification. The bar is demonstrating, to a skeptical GP, that nothing else worked. Most people don’t clear it. Most people don’t try.

Portugal grows medical cannabis. Processes it. Quality-controls it. Ships it north. It is Europe’s largest cannabis exporter — more than 32 tonnes in 2024, bound mostly for Germany, Spain, and Poland. Portuguese cannabis leaves in refrigerated containers for Berlin while Portuguese patients are still arguing with their doctors.

Vigo is an hour away. An hour. Across the Minho. Different country, different policy.

The rest of Europe is not much better.

Europe has solved this in the way Europe solves many things: partially, inconsistently, and with impressive paperwork. Germany has legalized cannabis but is still figuring out how to sell it. Amsterdam has tolerated it for decades while pretending not to know where it comes from. Switzerland runs pilot programs for registered residents. Malta has associations. No one has built a normal store.

In its 2025 report, the EU’s own drug agency put the continent’s illegal cannabis retail market at a minimum of €12.1 billion a year. That is Europe’s cannabis retail infrastructure: €12.1 billion, unregulated, untaxed, and functioning without permission.

Germany legalized cannabis in April 2024 and then discovered that legalizing a thing and selling it are two separate problems. It told its citizens it had legalized cannabis and asked for a minute.

That was two years ago. The 2025 election then handed Germany a government that doesn’t want to finish the job. No legal storefront is expected soon.

The gummy is available in Columbus, Ohio.

None of this is because the experiment failed. It worked — overdoses down, HIV down, prisons emptier, use no higher than before. It is the rare drug policy that is morally coherent. Which is what makes the rest of it so strange: the country that showed the world how to think about drugs grows cannabis professionally, ships it to Berlin, and tells its own patients to try ibuprofen first. The most progressive drug-policy nation on earth is a short drive from a cannabis club named for the Hail Mary, and has decided this is fine.

What Portugal decriminalized in 2001 was the user — possession, personal use, the human being with a problem. It never touched supply. Twenty-five years on, the state licenses companies to grow pharmaceutical-grade cannabis for export, sets a medical bar most patients can’t clear, and leaves everyone else where they always were: buying from the same illegal market the reform was meant to retire. The €12.1 billion isn’t the absence of a policy. It is the policy. The restraint was never philosophical. Supply was never on the table — only the user ever was.

I live in Vila do Conde, in the country that did it best.

I still can’t buy a gummy.

Mark Blondin writes from Vila do Conde, Portugal, where he has been slowly working through the prescription process. His GP remains unconvinced.

The United States spends 5.7 trillion dollars a year on health care — more than eighteen cents of every dollar the economy produces — and dies younger and sicker than any comparable nation on earth. The ordinary word for this is failure. The more useful word is design.

What the Evidence Actually Says

For a long list of ordinary conditions, the behavioral intervention sits in the same peer-reviewed literature as the drug, and condition by condition it holds up.

In healthy, low-risk people who have never had a heart attack, 200 to 470 have to take a daily statin for five years to prevent one heart attack. The rest take it every day and their risk barely moves. A daily walk, a changed plate, and a full night’s sleep move the same risk and cost the body nothing.

In a clinical trial, type 2 diabetes went into remission in nearly half the people who changed only how they ate. Remission — not management. Against a lifetime of metformin.

Cognitive behavioral therapy is the first-line treatment for insomnia in the guidelines themselves. It is almost never what gets handed across the desk.

For mild and moderate depression the antidepressant barely clears the sugar pill, and the exercise trials hold their own beside it. Severe depression is another country, and a medical emergency.

And physical activity is one of the strongest modifiable protections anyone has found against dementia. The 2024 Lancet Commission identified 14 such factors in all, and found that addressing the full set could theoretically prevent or delay nearly half of dementia cases — a figure drawn from following tens of thousands of people over years, not from running a drug trial. No single factor carries all of that, but the protection from staying active still outruns the demonstrated benefit of any dementia drug yet approved, because the drugs barely work.

The industry would rather you treat this as an open question. It should not want that fight, because the guidelines do not hide the behavioral case. They print it. Lifestyle first, then the drug. What the system withholds is not the recommendation. It is the delivery — the forty-five minutes, the billing code, the referral that actually leads somewhere. The evidence was never buried. It was left unfunded.

Why Behavior Change Doesn’t Scale

None of which is easy, and an essay that pretends it is has already joined the sales force. Behavior change is hard. The diabetes that went into remission went there inside a structured program with people checking in, not on willpower alone in a kitchen at midnight. Sustained change holds only when something holds it up — the follow-up call, the group, the coach, the forty-five minutes that recur. A pill needs none of that. It scales on a swallow.

So the system met the hardest problem in medicine — getting a person to keep a change — and solved it the way a balance sheet does: with the product that asks for no support and bills every month.

I write this from Portugal, where universal coverage is not a policy debate. It’s a Tuesday. From outside the American system its assumptions become visible the way a room’s smell only reaches you when you come in from the street. Ozempic is the cleanest example. The same molecule sells here for a fraction of its American price. The patent costs money. The molecule doesn’t.

The Oldest Move the Industry Owns

“Personal responsibility” is co-signed by every player in the medical-industrial complex. It relocates the blame onto your bathroom mirror and calls the relocation a cure. The wellness aisle will happily sell you the wish to opt out as one more product. A treadmill threatens no one.

What threatens the model is the same act, named correctly and done together.

Power is supplied to institutions, not possessed by them. No board ever voted for your illness. It voted, quarter after quarter, for the returns that managing illness throws off — the same destination arrived at without anyone having to want it. A revenue model runs on participation, and participation can be withdrawn. Every unnecessary prescription not filled is a dollar pulled out of a revenue line. Not stopped cold — the preventive step first, the condition improved, then, with a doctor, the pill reconsidered. One person doing it is a private health decision. A neighborhood doing it is a market contracting.

Be honest about the size of the lever. The treatment of chronic disease in America runs to something like two trillion dollars a year, and not all of it is refusable. You cannot walk your way out of insulin, or a shattered hip, or chemotherapy, and no one should try. What becomes unviable is not medicine. It is the chronic-disease revenue model wearing medicine’s coat. The revenue model you can walk away from. Real medicine you will always need — and a sane country would fund it without flinching.

And you cannot walk your way out of a zip code. Movement is the cheapest medicine we have, and it is not equally available — not to the woman working two shifts, not on a street without sidewalks, not in a body that already hurts, not in a food desert engineered three federal subsidies upstream. The cheapness of the cure is not your liberation. It is the evidence of the crime. An intervention this effective and this systematically sidelined tells you precisely what the system is optimizing for, and it isn’t your health.

What Actually Moves Something This Large

Look at tobacco. American smoking rates have fallen by more than half since the 1960s, and the fall began from below. The surgeon general told the truth in 1964, the truth spread, and the norm shifted before the steep cigarette taxes, before the indoor bans, before the lawsuits. People changed their minds first. Then the structural levers — taxes, advertising bans, smoke-free laws — caught up, drove the decline the rest of the way, and made it permanent. Neither arm did it alone. The culture moved from underneath and the law locked it in from above.

The health version is the same shape. Collective refusal from below, and a public system built for prevention from above — the visit that takes a full history, the exercise prescription written on the same pad as the pharmaceutical one, the appointment paid for as the care it plainly is. Refusal alone is a protest. A system alone is a waiting room nobody enters. Together they move something this large.

Until then the eight minutes go on. The pad comes out. The commercial break fills with drug spots aimed at the disease the last break sold you. None of it is a malfunction.

The goal was never your health. It was your continued, managed, profitable illness. The walk they told you was a personal virtue was always a withdrawn dollar.

Take it back. All of you.

Subscribe now

On March 9, 2026, New Zealand released the Phase 2 findings of its official COVID-19 inquiry. The international press covered the release. The findings, they did not.

The contrast is documented. In 2020 and 2021, The Guardian ran a sustained volume of coverage anchored in a fixed vocabulary: “one of the best health outcomes in the world,” “successful,” “world-leading,” “model.” On March 9, 2026, when the Royal Commission inquiry findings were published, The Guardian’s headline read: “New Zealand Covid response among world’s best but ‘scars’ remain, inquiry finds.” The negative findings — on proportionality, on populations that bore disproportionate costs, on rights questions courts had raised — were absorbed into the word “scars” and filed as lessons rather than failures. The frame of success survived the verdict on the success.

The New York Times trajectory tells the same story in three headlines. May 23, 2020: “How Jacinda Ardern Sold a Drastic Lockdown” — the cornerstone of the leadership-and-empathy frame. January 15, 2021: “Why New Zealand’s Pandemic Response Worked (and What Comes Next)” — the success still praised, sustainability questions appearing in the subordinate clause. October 4, 2021: “Battling Delta, New Zealand Abandons Its Zero-Covid Ambitions” — the strategy finally conceded to have failed, but framed as a response to an external force (Delta), not as a policy reckoning. The arc moves from admiration to pragmatic adjustment. It does not move to accountability.

A major investigative reframing of the pandemic record — the kind the Times applied to Iraq WMD or the Afghanistan withdrawal — has not appeared. Time ran the elimination strategy as proof that competent governance works. Its archives show no major headline centered on the Phase 2 inquiry findings. The BBC’s pattern was set early. On April 19, 2020 — six weeks into the pandemic — its headline read: “Coronavirus: How New Zealand relied on science and empathy.” The piece credited the response as “early, strict and compassionate,” noted that Ardern’s leadership had “won global attention,” and offered it as a “model response” defined by “clarity and trust in science.” The BBC did not return to that framing with equivalent prominence when the inquiry found the model had structural failures. Calm technocratic approval through the consensus years; minimal global amplification of the retrospective.

Celebration was front-page. Reassessment is archival. That asymmetry is a choice, and it is not a journalism question only about New Zealand. It is a question about how the press manages its own prior commitments.

What the press built

Between 2020 and 2022, the international media constructed an Ardern narrative that was less reporting than secular hagiography.

The machine had many components.

In June 2020, the Economist Intelligence Unit ranked New Zealand’s COVID response the best among 21 OECD countries. Bloomberg named New Zealand the best country in the world to live in during the pandemic in November 2020. Academic journals published peer-reviewed papers calling her communications strategy a “masterclass.” The Harvard Kennedy School produced a formal case study — “Jacinda Ardern: Leading New Zealand Through the COVID-19 Pandemic” — deployed in undergraduate, graduate, and executive education programs across the country. The press did not report on this institutional apparatus critically. It became part of it.

What distinguished Ardern’s coverage was its fusion of policy validation with identity politics. Female leaders, multiple studies argued, had simply handled the pandemic better. Her Facebook Live broadcasts from her baby’s nursery were not merely covered. They were celebrated as a new model of leadership. The coverage did not distinguish between the leader and the strategy. Criticizing the elimination policy meant criticizing her. And criticizing her, in the media environment of 2020-2021, carried obvious costs.

What the coverage did not do, during this entire period, was identify, examine, and critique countries like Denmark. New Zealand is an island nation with a relatively young population, lower comorbidity rates, and the most significant geographic advantage any country could have in a respiratory pandemic: two oceans and no land borders. Denmark is densely populated, land-linked to continental Europe, and carries a slightly older median age. It played a harder epidemiological game without sealing its borders, without running a government lottery to determine which of its own citizens could come home, and without mandates that persisted in schools and retail settings into late 2022.

By the time the pandemic stabilized, New Zealand’s COVID death toll was approximately 1,100 to 1,200 per million. Denmark’s was approximately 1,400 per million. The gap is real. But Denmark achieved it without binary border closure, without the Alert Level framework that triggered automatic nationwide lockdowns at single community cases, and despite being geographically embedded in the continent where the virus circulated most freely. Adjust for the age differential between the two populations and the gap narrows further. The press that ranked New Zealand’s response best-in-class was not systematically asking what the geography was doing versus what the policy was doing. That question was not asked because asking it was not consistent with the story being told.

What the press suppressed

While building the model, the press was simultaneously participating in the suppression of the scientists who disagreed with it.

On October 4, 2020, three epidemiologists — Martin Kulldorff of Harvard, Sunetra Gupta of Oxford, and Jay Bhattacharya of Stanford — published the Great Barrington Declaration, an open letter arguing that pandemic policy should focus protection on the elderly and clinically vulnerable while allowing lower-risk populations to resume normal life. It was signed by thousands of scientists and co-signed by Michael Levitt, a Nobel laureate at Stanford.

Four days later, Francis Collins, then Director of the National Institutes of Health, sent an email to Anthony Fauci that has since become the machine’s paper trail. Obtained through a Freedom of Information Act request and confirmed in congressional testimony, Collins wrote: “This proposal from three fringe epidemiologists who met with the Secretary seems to be getting a lot of attention — and even a co-signature from Nobel Prize winner Mike Levitt at Stanford. There needs to be a quick and devastating published take-down of its premises. I don’t see anything like that online yet — is it underway?”

Note what Collins did not write. He did not write: “We should convene a public scientific debate.” He did not write: “Let’s examine their evidence.” The Director of the NIH wrote that there needed to be a “devastating takedown.” Within ten minutes, Fauci responded by forwarding a piece from Wired magazine.

The press covered this story. Not in real time — after the emails were FOIAed, years later. It did not cover it as it was happening. The coordinated dismissal of three credentialed scientists from three of the world’s most prestigious universities was not reported as an institutional scandal in real time. It was, for the most part, amplified. The media that received the “devastating takedown” talking points ran them. The scientists were fringe. The science was settled. The model held.

The same pattern played out around Sweden. Anders Tegnell, Sweden’s state epidemiologist, implemented a voluntary, proportional response rather than compulsory lockdown, and was subjected to an international press pile-on documented in a peer-reviewed study published in Globalization and Health in July 2020. Rachel Elisabeth Irwin of Lund University identified six recurring narratives in international media coverage of Sweden, including that Sweden had a “herd immunity strategy” and was “not following expert advice.” While those narratives were partially grounded in reality, the language and framing used systematically distorted the accuracy of the reporting. Tegnell repeatedly argued that critics were emphasizing differences rather than similarities between countries. The press was not interested in the correction.

Sweden was necessary, as a story, to make the New Zealand model look correct by contrast. Jonas Ludvigsson, a clinical epidemiology professor at the Karolinska Institute, published peer-reviewed findings suggesting the risk of severe illness from COVID-19 in children was low. The findings directly challenged the case for keeping schools closed. He faced such a sustained barrage of personal attacks and threats that in February 2021 he announced he was leaving COVID-19 research entirely. The President of the Karolinska Institute publicly condemned the harassment. That story received minimal coverage in the outlets that had amplified the attacks’ underlying premise.

The silence as evidence

Journalism criticism often focuses on what is published: the false story, the agenda-driven frame, the institutional bias in the coverage itself. Rarer is the examination of what is not published, because absence requires documentation. The inquiry’s findings provide that documentation. The press covered the report briefly. The findings, with one or two exceptions, it did not engage.

New Zealand’s Royal Commission of Inquiry into COVID-19 Lessons Learned found that the elimination strategy imposed disproportionate costs on specific populations: young people, families separated by closed borders, workers who could not perform their jobs remotely. It found that the Alert Level framework, which triggered automatic national restrictions at single community cases, lacked the proportionality mechanisms required to weigh those costs against benefits. It found that New Zealand’s borders were effectively closed to the government’s own citizens for periods that courts questioned as lawful. These are not peripheral findings or administrative footnotes. They are the formal conclusion of New Zealand’s own government that the model had structural failures the model’s architects did not adequately account for.

The press that built the model has chosen, in large part, not to cover the verdict on the model with equivalent energy. The Guardian’s March 2026 headline — “world’s best but ‘scars’ remain” — illustrates the pattern precisely. The inquiry’s critical findings are present; they are present as subordinate clause. The success frame, established in 2020, governs the 2026 retrospective. Covering the inquiry honestly means covering the gap between what was written then and what is documented now. It means reporters examining their own archives. It means the institutions that amplified the “devastating takedown” of the Great Barrington scientists asking whether the science they said was settled was, in fact, settled.

That examination has not happened. Its absence is itself evidence of the institutional dynamic this essay is describing.

The kicker

In November 2024, President Donald Trump nominated Jay Bhattacharya — Stanford epidemiologist, co-author of the Great Barrington Declaration, one of the three scientists Francis Collins wanted to “devastatingly” destroy — to lead the National Institutes of Health. The Senate confirmed him on March 25, 2025.

This fact will be used by some readers to dismiss this essay. Bhattacharya was appointed by Trump, confirmed on a party-line vote, and operates within an administration that has made NIH restructuring politically contentious in ways that have nothing to do with COVID epidemiology. All of that is true.

It does not change what Collins wrote in 2020. It does not change what Fauci forwarded. It does not change what the New Zealand inquiry found. And it does not change the institutional record: a Director of the NIH solicited a “devastating takedown” of three credentialed scientists from three of the world’s most prestigious universities; major outlets ran the takedown; the same outlets later covered the elevation of one of those scientists to lead the agency that had tried to silence him, and they covered it as a political story about Trump.

That choice — to make the story about Trump rather than about the institutional record — is itself a continuation of the dynamic this essay describes. The press that labeled Bhattacharya fringe in 2020 did not cover his confirmation as an accountability story about what happened to a scientist the institutional machinery tried to silence, and whether that machinery, which included major journalism, owes its audience a reckoning.

The press that built the model did not lose the ability to revisit its own archives. It declined to.