It’s Not the Immigrant….. Stupid

The tsunami has already arrived. You just weren’t told.

📢 What you are about to read is the most complete and important piece I have written in my career.

That is not false modesty in reverse. I have written about Steve Keen and the private debt timebomb that orthodox economics refuses to see. About the machinery of extraction that funnels wealth upward while leaving households to survive on credit. About the .1% and how they have quietly purchased the institutions that were supposed to hold them accountable. About authoritarian politics and the collapse it tends to precede. And in Silent Echoes, about how conspiracy theorists are not madmen but messengers — ordinary people who sense, correctly, that something is being done to them, but who have been handed the wrong map and told the wrong destination.

This article is where all of that comes together.

What I am arguing here is not complicated, even though the system that obscures it has been made deliberately complex. We are walking into a wall. Not blindly, that would be forgivable. We are walking into it with our eyes open, nudged forward by people who will not be standing anywhere near it when we hit. The wall is the consequence of forty years of decisions that stripped the productive gains of an entire civilisation from the people who generated them and concentrated those gains in fewer hands than at any point in recorded economic history.

The blood nose is coming for all of us. The handful engineering it have already moved to high ground.

---

Dario Amodei, CEO of Anthropic, the company that makes the AI assistant you may be using to read summaries of articles like this one, issued a warning this week. A tsunami, he said, is coming. Artificial intelligence is about to reach human-level intelligence, and society has no idea what is about to hit it.

It’s as if this tsunami is coming at us, and it’s so close we can see it on the horizon. And yet people are coming up with these explanations, oh, it’s not actually a tsunami… that’s just a trick of the light.

He’s right that people are not paying attention. He’s wrong about the timeline.

The tsunami did not announce itself with a warning siren. It has been moving through the economy for twenty years, quietly, methodically, leaving behind it a trail of gutted industries, stagnant wages, hollowed-out towns, and a middle class that has been slowly drowning and blaming itself for not being able to swim.

The wave has already hit. What Amodei is watching build on the horizon is not the first one. It’s the second.

And here is the thing about tsunamis: by the time you can see them, the question is no longer whether to prepare. The question is who you blame for the damage already done. That question, as it turns out, has been very carefully answered for you. By politicians who knew. By corporations that profited. And by a press that, with some honourable exceptions, spent two decades reprinting press releases and political talking points instead of doing the one thing journalism exists to do: tell you what is actually happening.

What is actually happening is not a story about immigrants. It never was.

“It’s the Economy, Stupid”

In 1992, James Carville pinned a sign in Bill Clinton’s campaign headquarters that read: “It’s the economy, stupid.” It was a reminder to stay focused on the thing that actually determines how people feel about their lives. Not foreign policy. Not culture war. The economy. Whether you can pay rent. Whether your job will exist next year. Whether your children will live better than you.

Thirty years later, the same economic anxiety that Carville identified is driving political movements across the Western world. But this time, the campaign slogan has been rewritten. This time it reads: “It’s the immigrant, stupid.”

And a significant portion of the electorate has believed it.

Donald Trump built his 2016 and 2024 campaigns on it. Giorgia Meloni rode it to power in Italy. Marine Le Pen has made a career of it in France. Nigel Farage has never stopped saying it in Britain. Viktor Orban has governed Hungary on little else. The script is the same everywhere: your wages are low because immigrants are undercutting you. Your hospital is overrun because refugees are using it. Your child can’t get a job because foreigners took it.

It is one of the most successful political lies of the modern era. And it has been extraordinarily useful to a small number of people who understood, very clearly, what was actually happening to the economy and why.

What Was Actually Happening

Let’s start with a number: 74.

Between 1973 and 2013, US worker productivity rose 74 percent. The economy got dramatically more efficient. Companies produced far more output per worker than they ever had before. This is the foundational promise of capitalism: technology improves productivity, productivity generates wealth, wealth flows through wages, and everyone benefits.

The median worker’s compensation over that same period rose 9 percent.

Nine. Not 74. Nine.

The Economic Policy Institute, which has tracked this divergence for decades, no longer calls it wage stagnation. They call it wage suppression. The distinction matters. Stagnation implies drift, accident, market forces beyond anyone’s control. Suppression implies a mechanism. A choice. A beneficiary.

The beneficiary was not the immigrant. It was the shareholder.

The productivity gains went somewhere. They went up. Into profit margins. Into executive compensation. Into stock buybacks that inflated asset prices for people who own assets. The worker, who generated the productivity, got 9 cents on the 74-cent dollar.

This is the foundation of everything that follows. Keep it in your mind as we go.

The Manufacturing Story Nobody Told

Between 2000 and 2010, the United States lost approximately 5 million manufacturing jobs.

You know the story you were told: China. Globalisation. Trade deals. The jobs went overseas because foreigners would work for less.

There is some truth in that. But it is not the whole truth, and the part that is missing is the part that changes everything.

In 2020, economists Daron Acemoglu and Pascual Restrepo published research in the Journal of Political Economy that has since become the most rigorously cited empirical work on automation and employment. Their finding, covering the period from 1990 to 2007: each industrial robot introduced per 1,000 workers in a local labour market displaced approximately 5.6 workers in that area.

Not replaced temporarily. Not displaced into equivalent jobs elsewhere. Displaced. With wages in those communities suppressed simultaneously, compounding the damage to household finances already stretched thin.

Acemoglu and Restrepo’s later work, published in Econometrica in 2022, went further: between 50 and 70 percent of the increase in US wage inequality over the past four decades is attributable to the decline of workers in roles undergoing automation. Four decades. That is not a forecast. That is a forensic accounting of damage already done.

The story told to the people in those communities was not this story. The story told was: the Mexicans took your job. The Chinese took your job. The refugees took your job.

The robots took the job. The robots were bought by the corporation. The corporation kept the productivity gain. And then the corporation funded politicians who told you it was the immigrant.

The White-Collar Story That Is Happening Now

If you think automation is a blue-collar problem, you have not been paying attention to the last five years.

The entry-level hiring collapse is where the white-collar story becomes impossible to ignore. LinkedIn data shows that graduates from the 2021 to 2023 cohorts entered AI-exposed roles at lower rates and took longer to find first employment than any cohort before them. The gap began emerging before November 2022 — which means before ChatGPT went public, before the general public had a framework for understanding what was happening. Algorithmic displacement of entry-level white-collar work was already underway.

Big technology companies reduced new graduate hiring by 25 percent in 2024 compared to 2023. These are not positions that were paused and will reopen. They are positions that ceased to exist. Research by SignalFire found a 13 percent decline in employment for workers aged 22 to 25 in AI-exposed jobs since late 2022.

Bloomberg analysis of task-level exposure found that AI could replace 53 percent of market research analyst tasks and 67 percent of sales representative tasks. Managerial roles face 9 to 21 percent automation risk. The technology is eating the bottom of the career ladder first, which means it is removing the rungs that previous generations used to climb.

In the first six months of 2025, 77,999 US technology sector job losses were directly attributed to AI by the companies making those cuts. A 2024 survey found that 44 percent of companies using AI said employees would “definitely” or “probably” be laid off as a result. Not might be. Definitely or probably.

Paralegals face 80 percent automation risk. Bank tellers are projected to decline by 51,400 positions by 2033. Cashiers by 353,100. Customer service representatives are already being replaced at scale by systems that cost a fraction of a human salary, never ask for a raise, never get sick, and never unionise.

The IMF estimates that AI will affect nearly 40 percent of all jobs worldwide. This is not the tsunami on the horizon. This is the water already at your knees.

The Gender Dimension Nobody Mentions

Here is a number that almost never appears in mainstream coverage of automation, and its absence is itself a political statement.

In the United States, 79 percent of employed women work in jobs at high risk of automation. The figure for men is 58 percent. This reflects the concentration of women in administrative, clerical, and service roles that are precisely the categories algorithmic systems were designed to replace first.

Globally, according to IMF and ILO data, jobs most vulnerable to AI-driven task automation make up 9.6 percent of female employment in high-income countries. For male jobs, the figure is 3.2 percent.

Women are being displaced at roughly three times the rate of men. This is not a technology story. It is a story about which labour was valued least by the economy that built these systems, and which labour gets sacrificed first when efficiency becomes the only measure that matters.

You will not find this statistic in most political speeches about job loss. It does not fit the narrative that automation anxiety should be channelled toward immigrants. A 58-year-old man in Ohio whose manufacturing job was automated is a useful political figure. A 34-year-old woman whose administrative role was dissolved by a software package does not fit the casting.

What Steve Keen Would Tell You

The displacement story has a second layer, and it is the one that explains why entire economies now feel as if they are teetering.

Economist Steve Keen, building on Hyman Minsky’s framework of financial instability, has spent decades documenting something that orthodox economics systematically ignores: the role of private debt in determining economic stability. In Keen’s framework, it is not government deficits that threaten economic collapse but the accumulation of private debt in households and corporations. When private debt grows faster than income, the system becomes fragile. When income falls — through job displacement, wage suppression, or the gig economy absorbing displaced workers into lower-paid precarity — households sustain consumption through borrowing. The debt grows. The fragility deepens. And then, when the system snaps, it snaps catastrophically.

Apply this to the automation story and the picture becomes very clear.

The productivity gains from forty years of automation did not go to workers. They went to shareholders and executives. Workers, facing stagnant or declining real wages, sustained their standard of living through debt. Consumer credit. Mortgage refinancing. Student loans taken on the promise that education would restore the income trajectory that automation had disrupted.

The gig economy was not a new economy. It was an absorption mechanism. Uber, Deliveroo, TaskRabbit, and their equivalents absorbed workers displaced from stable employment and reclassified them as self-employed contractors, statistically invisible as unemployed, economically functioning as a precarious underclass servicing the consumption habits of people who had not yet been automated. Meanwhile, the household debt those workers carried continued to compound.

This is the kindling. The political manipulation is the match.

When people are economically desperate, economically confused, and economically angry, they are available for capture. The mechanism of that capture is simple: give the anger a face. Make it a foreign face. Make it a dark face. Give the displaced worker a villain who is not the CEO who signed the automation contract, not the shareholder who collected the dividend, not the politician who wrote the tax code that made automation more profitable than employment. Give them the immigrant.

An Economy That Eats Its Consumers

There is a logic failure at the heart of the automation story that is so obvious it should not need stating, and yet it apparently does.

An economy is not a machine for producing goods. It is a system for exchanging them. Production without consumption is not efficiency. It is inventory. And inventory that never moves is just waste with better accounting.

Henry Ford understood this in 1914 when he doubled his workers’ wages to five dollars a day. His reasoning was not philanthropic. It was structural. He wanted his workers to be able to afford the cars they were building. He understood that a factory that destroys the purchasing power of its own market is a factory running toward a wall.

The people replacing Ford’s logic with algorithmic efficiency metrics have apparently forgotten the wall.

When you displace a worker, you do not just remove a cost from your balance sheet. You remove a consumer from the economy. You remove someone who bought groceries, paid rent, took their children to the dentist, replaced their car every seven years, went to restaurants, bought shoes, paid a plumber. All of those transactions disappear with the job. The businesses that depended on those transactions contract. The workers in those businesses face pressure. The chain of consumption that an employed person represents is severed at the source.

Multiply this by millions. Multiply it by decades. And then ask yourself what you are left with.

You are left with an economy increasingly composed of people who own automated systems generating returns for a shrinking number of asset holders, and a growing population whose capacity to participate in the economy as consumers is being systematically destroyed. The AI and robots producing the goods will not buy them. They have no wants, no needs, no households, no children to clothe. They generate output but consume nothing.

This is not a sustainable economic model. It is a slow liquidation of the market itself.

The numbers already show this. Consumer debt in the United States has exceeded four trillion dollars. Household savings rates have collapsed to historic lows. Retail bankruptcies have accelerated as the middle-income consumer — the engine of a consumption economy — has been hollowed out. The luxury segment thrives because the asset-holding class has never been richer. The mass market contracts because the wage-earning class has never been more precarious.

The economists who designed this outcome are not stupid. They understand the arithmetic. What they appear to have decided is that the transition period — the decades during which wealth concentrates at the top while the consumer base erodes — is long enough for them to extract what they need before the system reaches its structural limit. The short-sightedness is not ignorance. It is a calculated bet that the bill comes due after they are done collecting.

What is being accumulated in the trillions held by a handful of individuals and corporations is not wealth in any meaningful economic sense. It is a claim on a future economy that their own behaviour is making impossible. Elon Musk’s net worth, at its peak, exceeded the GDP of most countries on earth. That accumulation does not circulate. It does not build schools or buy groceries or start businesses in struggling towns. It sits in asset valuations, in offshore structures, in shareholdings that appreciate on paper while the physical economy that underpins them loses its capacity to function.

Greed at this scale is not just a moral failure. It is an engineering failure. You cannot extract the load-bearing walls of a structure and expect the building to stand. The middle class was not a political ideal. It was a functional necessity. It was the mechanism by which the productive output of an industrial economy was converted into the consumption that justified continued production. Destroy it, and you do not get a leaner economy. You get a smaller one, with a gilded top and a collapsed foundation.

The people who made this choice knew what they were doing. They made it anyway.

What Immigration Actually Does

Since we are dealing in facts, let us deal with the immigration facts too.

The claim that immigrants suppress wages, steal jobs, drain public services, and increase crime is not merely wrong. It is comprehensively, empirically, repeatedly wrong, tested across multiple countries over multiple decades and refuted every time with data that the politicians making the claim have consistently ignored.

Start with the three economies most shaped by immigration: the United States, Canada, and Australia. All three are, by any serious measure, among the most economically successful societies in human history. All three were built, at every stage of their development, by waves of immigrants who arrived with nothing except the willingness to work and the intention to stay. The Italians and Irish and Jews and Chinese and Vietnamese and Mexicans and Indians who built American industry, medicine, technology, and culture did not drain the economy. They were the economy.

Silicon Valley, the most economically productive geographic concentration of human activity in the modern world, was built substantially by immigrants and the children of immigrants. Google, founded by Sergey Brin, who arrived from the Soviet Union. Yahoo, co-founded by Jerry Yang, who arrived from Taiwan. eBay, founded by Pierre Omidyar, born in France to Iranian parents. Intel, co-founded by Andrew Grove, who fled Hungary. These are not exceptions. They are the pattern.

Canada has made immigration a deliberate economic strategy. Its points-based system is explicitly designed to import the skills the domestic economy needs, and the results are documented: immigrants to Canada pay significantly more in taxes over their lifetimes than they draw in services. They start businesses at higher rates than the native-born population. They fill labour market gaps in healthcare, construction, agriculture, and technology that would otherwise constrain economic growth. The Canadian economy does not tolerate immigrants. It depends on them.

Australia tells the same story. A country of 26 million people with one of the highest per-capita incomes in the world, built on a foundation of successive migration waves that were resisted and resented at each arrival — the Chinese in the goldfields, the Italians and Greeks in the postwar decades, the Vietnamese after 1975, the Lebanese, the Indians, the Africans — and which in every case ultimately enriched the economy and the culture that received them.

The wage suppression claim is perhaps the most persistently repeated lie in the immigration debate, and it has the most thoroughly demolished empirical record.

The Bank of England, which is not a left-wing institution and has no political interest in flattering migrants, conducted detailed analysis of the wage impact of EU immigration to the UK. Its finding was that the overall wage effect was negligible, and what small downward pressure existed was concentrated in the semi-skilled service sector, not in the broader labour market that political rhetoric claimed was being undercut.

George Borjas at Harvard, who is the most cited academic supporter of the wage-suppression claim, has been repeatedly challenged by economists including David Card, who won the Nobel Prize in Economics in 2021 substantially for his work on immigration and wages. Card’s findings, replicating across multiple natural experiments, showed that large influxes of immigrants had minimal to no negative wage effects on native workers. The Mariel boatlift — the arrival of 125,000 Cuban immigrants in Miami in 1980 — produced no measurable wage decline for Miami workers. The economy absorbed the labour and expanded.

On crime: the data is consistent and it runs entirely against the political narrative. Immigrants, including undocumented immigrants, commit crimes at lower rates than native-born citizens in every jurisdiction for which reliable data exists. This is not a marginal finding. It is a robust, replicated result across the United States, Europe, and Australia. The reasons are not complicated. People who have undertaken enormous personal risk and sacrifice to establish themselves in a new country have strong incentives to stay out of trouble. Deportation is a consequence that the native-born do not face. The stakes of transgression are categorically different.

A Stanford study found that undocumented immigrants in the United States are 45 percent less likely to be incarcerated than native-born citizens. The Cato Institute — a libertarian think tank with no particular affection for immigration for its own sake — has repeatedly found that both legal and illegal immigrants are less likely to commit crimes than native-born Americans.

None of this is secret. None of it is contested at the level of serious empirical research. It is simply ignored by politicians for whom the immigrant serves an indispensable function: they are the available face for economic anger that, if correctly directed, would implicate the politicians themselves.

The immigrant pays taxes. Buys goods. Starts businesses. Fills labour shortages. Commits crimes at lower rates. And has been the foundation of every successful immigrant-receiving economy in modern history.

The only thing the immigrant cannot do is serve the narrative of people who need a scapegoat for the damage they caused.

Who Benefits From the Lie

Let us be specific, because vagueness is where political manipulation hides.

Donald Trump built his political career on the immigration scapegoat while simultaneously overseeing, in his first term, the most aggressive corporate automation investment in American history, incentivised by the 2017 Tax Cuts and Jobs Act, which reduced the cost of capital investment and made replacing workers with machines more financially attractive. The workers who voted for him in 2016 believing he would bring back their jobs watched the opposite happen, at scale, with legislative support. His donors, who included technology investors and corporate executives whose firms were automating aggressively, benefited directly from policies his administration enacted.

In the UK, Boris Johnson and Nigel Farage built Brexit substantially on the claim that EU migration was responsible for wage suppression and public service strain. The actual evidence, including from the Bank of England, showed minimal wage impact from EU migration. The wage suppression was structural, driven by the same automation dynamics affecting the entire Western world, compounded by a decade of austerity that destroyed the public services the same politicians then blamed migrants for degrading.

In Italy, Giorgia Meloni’s Fratelli d’Italia has governed on a platform of cultural and demographic threat from migration while presiding over an economy in which youth unemployment remains structurally high, automation exposure in northern manufacturing is among the highest in Europe, and household debt continues to compound on stagnant wages.

The pattern is not coincidence. It is strategy.

The corporations automating labour have no interest in voters understanding the cause of their economic distress. The politicians funded by those corporations have every interest in providing a different explanation. The media outlets dependent on advertising revenue from those corporations have limited appetite for investigative journalism that names the mechanism. And so the lie circulates, elections are won, automation accelerates, and the displaced worker votes again for the politician who promises to deport the person who did not take the job.

The Press Failure

James Carville’s original insight was that political campaigns win by keeping voters focused on the thing that actually affects their lives. The corollary, which has been industrially exploited for two decades, is that you can win by keeping voters focused on the wrong thing.

The press was supposed to be the correction mechanism. Investigative journalism exists specifically to identify the gap between the story being told and the story that is true. It has, with honourable exceptions, failed at this task so comprehensively that the failure itself requires explanation.

Some of it is structural. Advertising-dependent media has limited appetite for attacking the corporations that fund it. Some of it is the collapse of local journalism, which was the layer of the press most likely to document the impact of automation at the community level: the factory that closed, the call centre that replaced fifty people with a chatbot, the paralegal firm that reduced headcount by 40 percent after implementing document review software.

Some of it is laziness, which is the most uncomfortable explanation and therefore the one most rarely offered. It is faster to reprint a politician’s statement about immigration than to build the empirical case linking automation to wage suppression to household debt to electoral desperation. The former fills column inches. The latter requires the kind of sustained, data-driven, source-building journalism that takes months and does not trend.

And so Dario Amodei can warn this week about a tsunami on the horizon, and the press can report his warning, and the cycle of future-focused alarm can continue, and the twenty years of damage already done can remain unnamed and unexamined, and the voters can continue to be told that it was the person who crossed the border, not the machine that crossed their pay stub.

What the Numbers Actually Say

This is not speculation. The evidence is documented, peer-reviewed, and sitting in academic journals and institutional databases that are publicly accessible and almost never cited in political coverage.

Acemoglu and Restrepo established that robots displaced workers at scale between 1990 and 2007, before the AI revolution, before the automation conversation even reached mainstream politics. The damage was already done before most voters had a framework to understand it.

The Economic Policy Institute’s productivity-pay gap data shows four decades of workers generating wealth they did not receive. That wealth did not disappear. It went somewhere. The data shows where.

The entry-level employment collapse in AI-exposed sectors is documented in LinkedIn and SignalFire research that is not obscure or contested. It is simply not reported as the structural story it represents.

The IMF’s estimate that 40 percent of global jobs face AI exposure is not a fringe projection. It is a mainstream institutional finding that has produced almost no policy response commensurate with its scale.

And Keen’s framework for private debt dynamics sits beneath all of it, explaining why the economic fragility that makes populations available for manipulation is not the result of poor individual choices but of a system that extracted the productivity dividend upward and left the costs of that extraction as debt on the balance sheets of ordinary families.

This is the story. It has always been the story. The numbers have been there. The mechanism has been documented. The beneficiaries are identifiable.

The immigrant did not take your job. The algorithm did. The politician who told you otherwise was funded by the people who bought the algorithm. And the press that should have told you watched the tsunami build for twenty years and called it a trick of the light.

Sources

1. Acemoglu, D., & Restrepo, P. (2020). Robots and Jobs: Evidence from US Labor Markets. Journal of Political Economy, 128(6). **https://www.nber.org/papers/w23285**

2. Acemoglu, D., & Restrepo, P. (2022). Tasks, Automation, and the Rise in US Wage Inequality. Econometrica, 90(5). **https://onlinelibrary.wiley.com/doi/full/10.3982/ECTA19815**

3. Acemoglu, D., & Restrepo, P. (2019). Automation and New Tasks: How Technology Displaces and Reinstates Labour. Journal of Economic Perspectives, 33(2). **https://www.aeaweb.org/articles?id=10.1257%2Fjep.33.2.3**

4. Economic Policy Institute. Productivity-Pay Gap tracker. **https://www.epi.org/productivity-pay-gap/**

5. Economic Policy Institute. Wage Suppression and Inequality. **https://www.epi.org/unequalpower/publications/wage-suppression-inequality/**

6. International Monetary Fund. World Economic Outlook: AI and Labour Markets, 2024.

7. International Labour Organization. Gender and Automation Exposure data, 2024.

8. MIT Work of the Future Task Force.

https://workofthefuture.mit.edu/

9. SignalFire Research. Graduate hiring and AI-exposed job trends, 2024.

10. Bloomberg Intelligence. AI Task Exposure by Occupation, 2024.

11. Amodei, D. Interview with Nikhil Kamath. WTF Is podcast, February 25, 2026.

12. Landymore, F. “Anthropic CEO Warns of ‘Tsunami’ on Horizon.” Futurism, February 27, 2026.

13. Keen, S. Debunking Economics (revised edition). Zed Books, 2011.

14. Keen, S. The New Economics: A Manifesto. Polity Press, 2021.

15. Minsky, H. Stabilizing an Unstable Economy. McGraw-Hill, 2008.

Aldo Grech ”The future is embedded in the choice”. Books and private advisory.