Nobody Reverses Course on a Full Stomach

A doomer's view of the coalition, the cushion, and what it would take to change direction

The Village That Signed

In December 1978, eighteen farmers in a village called Xiaogang, in Anhui province, gathered at night in a mud-walled house and pressed their thumbprints onto a sheet of paper. They were agreeing to carve the commune's collective fields into household plots, which at the time was the sort of thing that could get you shot. They also agreed that if any of them were arrested or executed for it, the rest would raise his children until they turned eighteen.

As the story is usually told, the village had been hungry for most of two decades. Anhui was one of the hardest-hit provinces in the famine of 1959 to 1961, and Xiaogang was known for sending its people out to beg. The next harvest, as I understand it, was bigger than the previous five put together.

Nobody in that room had read Hayek. They had read their own children's ribs.

And, looking to the future of the US (first) and the rest of us (consequentially), I am counting many ribs in our future.

Here is the principle: people do not abandon a bad idea because somebody proves it is bad. They abandon it when staying costs more than leaving, and not one day earlier. Proof is cheap. Pain is persuasive.

Below, I will argue (with facts) that certain things hold true for the composition of the US majority. Those facts have consequences that are driven by inarguable principles. And those consequences are dire.

People are just people. Put a few hundred million of them inside a set of incentives and they will do what the incentives pay for, sincerely, cheerfully and with a clear conscience. Nobody has to be a villain for a country to drift. They only have to be ordinary.

Who Holds the Keys

https://www.nbcnews.com/politics/2026-elections/exit-polls

Round the exit-poll numbers off and the picture barely moves from one election to the next. Something like nine in ten Black voters go Democratic. Around six in ten women. Roughly seven in ten of the under-thirties. About two in three Jewish voters, and well over eight in ten Muslim voters, although the Muslim sample is small enough that I would treat that figure as a direction rather than a measurement. Gay voters sit in the same solid majority. 

Young. Gay. Muslim. Jew. Women. Black. College Educated: Add them up and you have the spine of the coalition that has just taken both houses of Congress.

Generalisations hold only to a point, and we all know individuals contradict them every day. The point is that the institutions which set the tone of the country (universities, newsrooms, the bureaucracy, the HR department, the foundation with the nice lobby) are staffed by people who are drawn from, trained by and answerable to this coalition's moral vocabulary. Elections may change who sits in the Oval Office. They do not change who writes the hiring policy at a university.

The soft social sciences are the stronghold. The surveys I have seen put the left-to-right ratio among faculty in sociology and anthropology somewhere between eight-to-one and considerably worse. Law, medicine and teaching lean left too, but more mildly, because those professions have to deliver something a customer can check. A bridge either stands or it doesn't. A sociology seminar cannot fall down.

Left-leaning voters pride themselves on being ‘more college-educated’, and ‘Maga is dumb’. As we will see shortly, the Left only really dominates the soft social sciences, and that is a decision that has proven to be net-negative in ROI, so not really a ‘smart’ decision that you should be proud of.

The Threat Premium

What binds the coalition together, once you strip away the policy lists, is a shared sense of being marginalised, disadvantaged or hunted.

Some of that is accurate. Jews have been hunted for two thousand years, which is not a perception. Black Americans carry a documented history that does not need me to summarise it. Some of the feeling is stale, a grievance that was fact in 1965 and is habit in 2026. Some of it is amplified by media that profit from the amplification. The ballot box cannot tell which is which, and neither can the survey. Both record the feeling, and the feeling moves behaviour whether or not the fact underneath it still holds.

I call this the threat premium: the extra weight a person gives to protection over production when they believe they are being hunted. It is a perfectly rational thing to pay. If the outside world is hostile you want shields, quotas, regulators and referees. You do not want a growth rate. A growth rate is a rising tide, and a person who believes the boat is rigged does not care how high the tide goes.

The premium shapes what the coalition elevates: identity over merit, grievance over growth, redistribution over production. Nobody has to decide this. It falls out of the belief the way water falls out of a tipped bucket.

The Spreadsheet Nobody Reads

The belief reaches into the most expensive decision most young people ever make, which is what to study.

Take two eighteen-year-olds (a young woman and a young man, or the other way around if that makes you happy). One enrolls in psychology, sociology or social work. Four years later she holds a degree, a debt, and four years of earnings she never made. The other signs on as an apprentice electrician. He is paid from the first week, licensed by his early twenties and owes nobody a cent. Run the lifetime numbers after debt and forgone wages, and for a large share of the people on the first path the answer is negative, or at best behind the second. Skilled trades and plenty of self-employed paths beat it outright.

Why would anybody choose the worse path? Because the choice is not made on a spreadsheet. It is made on meaning, on peer approval and on advice. And the advice comes from the people who sell the degree and who, in those fields, vote about nine to one the same way. Groupthink is not a conspiracy. It is a monoculture that cannot see its own weeds.

Apply the law of large numbers. One graduate making a poor choice is a story. Four hundred thousand making the same choice is a price. The supply of one kind of credential swells, the market for it stays roughly where it was, and the wage falls to meet the gap. Meanwhile the country is short of engineers, electricians and the people who can fix a transformer. A good part of the demand for the soft-credential graduate is the state itself, in government roles, nonprofits, universities and compliance departments funded by somebody else's tax or tuition (I believe it is the larger part, though I cannot prove the share). The job market for the degree is, to a large extent, the tax base. When a sector grows faster than the base that feeds it, you have a bubble, and bubbles pop.

So, what have we got so far? Demographic sub-groups who feel victimised, feel entitled, believe they are smart, even whilst none of those things are true, yet have democratic weight to determine policy and outcomes.

What will they create?

What a Coalition Like That Decides

Policy follows worldview, always has. If you hold the beliefs above, certain decisions look obvious and kind. Expand access to higher education in whatever field, because access is justice. Prefer identity-based criteria in admissions and hiring, because the field is tilted. Regulate and redistribute more, on equity grounds. And treat dissent from the marginalisation narrative as something worse than error: as harm.

That last one is the flashpoint: it is the thing "woke" will come and rap you on the knuckles for saying. I would rather say it and take the rap. When disagreement counts as injury, the feedback loop is cut, and a system with no feedback loop is flying blind. I'd even argue that the modern HR department maps onto the commune's political officer, but that is another essay.

Growth-oriented reforms (merit-based selection, vocational priority, fiscal restraint) meet structural resistance, and not because anyone hates growth. They collide with the self-conception of the people being asked to adopt them. They collide with the benefits of being a victim in an environment where victimhood is a status symbol.

You cannot easily ask a coalition built on the premise (they strongly believe)  that the system is rigged to endorse a reform built on the (opposite) premise that the system broadly works.

The Slow Bleed

Supply and demand do not care about intentions. Oversupply the credentials, undersupply the skills, and productivity gets sluggish. The gap shows up in the budget.

Start with the federal books. The Congressional Budget Office has this year's deficit at about $1.9 trillion, close to 6 per cent of GDP, in a year with no recession, with debt held by the public at roughly 101 per cent of GDP. Net interest costs about a trillion dollars, around a fifth of all federal revenue, and the CBO has that climbing to roughly a quarter of revenue by 2036.

The deficit is not the work of one coalition. Both teams have fed it, and the biggest lines are pensions and healthcare for the old, which nobody under forty invented. The other side guards the revenue door as fiercely as this coalition guards the spending door. (Both teams hold a veto. They just guard different doors.) The coalition's contribution is not the original sin. It is the veto over fixing it.

Add the slow arithmetic of relative decline. 

America's share of world output has been drifting down for a generation, partly because other countries are growing from lower bases, which is catch-up and not a failure. But a shrinking share is a shrinking margin. A country that was once so far ahead it could afford any mistake is now only somewhat ahead, and "somewhat" is a much less forgiving word.

Then comes the zero-sum drift. If growth slows, every group starts counting its slice. Absolute gains shrink, comparisons sharpen, and friction rises, because friction is what a rising tide was hiding. The same threat premium that built the coalition now gets fed by the stagnation. A loop, in other words, and loops are hard to exit from the inside.

So, the obvious question is: what needs to happen before there can be a course correction?  The answer is that things have to get really bad first. And the answer is, looking at historical precedent, it has to get really bad before a majority becomes sufficiently incentivised top go against all the vested interest groups.

What Failure Actually Looks Like

The reversals I keep coming back to are China after 1978, Singapore after 1965 and Chile after 1973. In each case a country had swapped competence for ideology, and in each case it turned around, but only after the failure was undeniable. China's people were, by the World Bank's later reckoning, overwhelmingly in extreme poverty. Singapore had been expelled from its federation, had no resources to speak of and had just lived through racial riots. Chile had inflation running at several hundred per cent a year and empty shelves.

Notice a detail that rarely gets mentioned. None of the three had to win a free argument against a majority that preferred the old way. A democracy has no such luxury. The reversal has to be voted in by the median voter, which means the pain has to reach the median voter, and stay long enough to change his mind. The recent democratic case is Argentina, which turned its politics inside out in 2023 after annual inflation passed 200 per cent.

So what would the American version look like? Take the thresholds from the historical cases, scale them to a rich country, and set them against where the United States stands this week.

Read the right-hand column and the problem jumps out. On the headline numbers America is not failing. It is doing something worse, for the purposes of reform. It is succeeding just enough.

Only two rows are even in view. Interest costs, which move slowly, and housing for the young, which hurts one cohort. And that cohort is the one to watch, because the under-thirties sit inside the coalition. If the dominant coalition is ever going to treat its own course as unsustainable, those are the people who would have to defect first, and they are the ones locked out of the house.

The Cushion

I call this the Cushion: the thickness of accumulated wealth, reserve-currency status and institutional momentum that lets a rich society go on being wrong long after a poor one would have been forced to stop.

Xiaogang had no cushion. There was nothing between those farmers and the grave, so the price of the old idea was paid in full, immediately, in children. An American family has a car, a mortgage, a retirement account that has ridden the market up, a pharmacy that opens on time and a government cheque if all else fails. The Cushion does not change the direction of a wrong idea. It decides the delay.

The Cushion has a long history of being mistaken for health. Rome clipped the silver out of its coinage for the better part of two centuries and went on paying its legions in ever-thinner coin until the day the coin stopped working. Argentina was among the richest countries on earth in the early twentieth century and spent the next hundred years proving that wealth is a lease, not a title. Britain in the 1950s and 1960s ran on the capital of an empire it could no longer afford, and by the 1970s was being called the sick man of Europe. The Soviet Union lived on oil and borrowed against oil, and looked immovable right up to the price collapse. In every one of those cases the wrong idea was visibly wrong to outsiders for decades. In every one of them it took the Cushion running out for the insiders to see it.

As an aside, backed by the gravitas of a thousand cable news panels and no research whatsoever: I suspect the best predictor of whether a society tolerates a bad idea is how recently it was hungry. Countries that have eaten well for three generations start to treat prosperity as a natural resource rather than an achievement. (And jokes about avocado toast would not be appropriate here 🙂)

This is why the better the numbers get, the longer the bad ideas survive. A record median income is not evidence the coalition's path works. It is evidence the Cushion is thick. Those are different claims, and the gap between them is where the doomer lives.

The Mandate That Wasn't

My reading of 2024 (and it is a reading, not a measurement) is that a large slice of American voters looked at the dashboard, saw the warning lights and decided the country was in an existential spot. The debt, the border, the schools, the cities, a bureaucracy that answered to nobody they could vote against. They picked the one man in the field who promised to put a boot through the dashboard. That was the opportunity: a rare moment when enough people had crossed the line to hand a government the chance to change the slow variables.

It was squandered. I am not going to argue why, because the reasons do not matter here and everyone reading has a favourite theory already. What matters is the result, which arrived this week. The Democrats took the House and the Senate. The count is still settling, but the House majority is far past the 218 needed, and the reporting on why points mostly at the cost of living, with the president's approval at record lows going in.

Backlash is not reversal. Read the last decade. Trump in 2016, backlash in 2018, Biden in 2020, backlash in 2022, Trump in 2024, backlash in 2026. That is not a country changing direction. It is a pendulum, and the pivot never moves. 

Each swing punishes whoever holds the keys for the pain the voter can feel, and each swing leaves the slow machinery (the universities, the HR departments, the credential pipeline, the agencies) exactly where it was.

Notice how little pain it took. Voters turned out a government with headline inflation at 3.7 per cent and core at 2.6, with much of the lift coming from petrol. That tells you what it costs to switch teams: a few points of inflation will do it. What it costs to switch systems is Argentina in 2023. Two thresholds, separated by a canyon. America crosses the first every second November and has never come near the second.

The Things They Feared

Take the conservative fear list, and I will pass no judgment on whether it was right or wrong: mass immigration, crime in the cities, cultural capture of the institutions, and the debt. Those fears drove the 2024 vote. What has happened to them?

Immigration moved first and furthest. The Census Bureau puts net international migration at 2.7 million in 2024, down to 1.3 million in 2025, with about 321,000 projected for 2026. The Brookings Institution estimates that net migration in 2025 was negative, somewhere between a loss of 295,000 and a loss of 10,000, the first time in at least half a century. Pew has the immigrant share of residents at 15.4 per cent, slipping from a recent peak of 15.8. Crime moved the same way. The FBI has the 2025 murder rate at 4.1 per 100,000, tying the lowest on record, with the first half of 2026 lower again.

So two of the four fears have shrunk, and the people who held them lost the midterms anyway. That tells you what wins elections. Voters pay for what they feel at the checkout, not for a fear that was resolved last year.

The other two fears, cultural capture and the debt, have not moved at all. That is the thesis in miniature. The fast variables can be changed by whoever holds the executive. The slow ones cannot.

And here is the catch for conservatives. Most of the immigration reversal was done by executive action, and what one executive does the next can undo. A Democratic Congress cannot reopen the border, but a Democratic president in 2029 can loosen the rules, and if flows return the fear returns with them. It cuts the other way as well, which the right does not enjoy hearing. Fewer working-age arrivals means fewer workers per retiree, and that makes the pension and interest arithmetic harder, not easier. You can want a closed border and a controlled debt, and still have to face that one wish eats into the other. (How hard it bites is argued over by people cleverer than me.)

A word, too, for anyone who thinks the coalition is a lock. Latino voters swung toward the Republicans in 2024 and, by most accounts, back again this week. That is how a group behaves when it votes on the checkout rather than the tribe.

The Rest of the World

A country the size of the United States is not just a country. It is the weather system the rest of us live in. Three things follow if the Cushion thins.

The first is the money. Treasuries are the world's safe asset. If interest costs reach a fifth to a quarter of revenue and stay there, creditors start asking what the safe asset is safe from, and the usual answers are inflation, financial repression (holding rates below inflation by decree) and, at the far edge, selective default. None of that is imminent. All of it sits in the playbook of every indebted state in history. A great many pension funds hold American debt, a good slice of Australian superannuation among them (I would bet), and every holder is a passenger.

The second is security. Allies have priced the American umbrella as a constant. We in Australia have staked a great deal on AUKUS and the alliance behind it. If the country holding the umbrella is turning inward, short of money and swinging its foreign policy every two years, the umbrella is a lease with a break clause, and sensible people start reading the fine print.

The third is contagion, and this is where the doomer view gets interesting. The same coalition arithmetic exists in Britain, Canada, Australia and most of Western Europe: a swelling credential sector, state-funded demand for it, identity-based institutional priorities, rising debt service. The difference is the Cushion. America's is the thickest. Smaller, poorer, more indebted countries hit the wall first. If I am right, the first democratic reversals will not come out of Washington. They will come from somewhere with less padding, the way Argentina's did. Watch the thin cushions. They are the canaries, and some of them are our friends.

Meanwhile the framework travels. Many of the world's future officials pass through the same universities that train the American professional class, and they carry the same vocabulary home.

And a note for home, because Australia is not outside this story. We run our own version of it: an education sector that exports credentials (I believe at a scale that would make an American provost blush), a public and quasi-public payroll that is one of the country's growth industries, and a Sydney housing market that sits, by the usual estimates, right where the table above puts the American breaking point. I am in no position to be smug about the Americans, and neither are you.

What I'd Watch

If the doomer view is right, there will be no headline announcing that the Cushion has run out. There will be tells, and they will be dull. I would watch six.

The first is the young. Under-thirties are the most loyal to the liberal, left-leaning coalition on paper and the most exposed to its costs in practice. When their Democratic margin starts to shrink at the same time as the housing numbers worsen, that is the defection beginning.

The second is where the clever eighteen-year-olds enrol. Apprenticeship and trade-school applications rising while the soft-credential enrolments fall is a quieter vote than any election, and it is the one that moves the supply side (I think it has already started in a small way, based on intuition rather than evidence).

The third is employers. When firms stop asking for the degree and start asking for the portfolio, the credential has lost its pricing power. A number of large companies and several state governments have already dropped degree requirements for plenty of roles.

The fourth is the long end of the bond market. When Treasury auctions begin to need a discount to clear, the Cushion is being eaten in public.

The fifth is people’s feet. The Census numbers already show California, Hawaii, New Mexico, Vermont and West Virginia shrinking while the South gains. That is the quiet feedback loop, and the nicest thing about it is that nobody has to win an argument. They just leave.

The sixth is the machines, and it is the one that could make the other five beside the point. The old picture of where GDP comes from has three legs: primary, resources pulled out of the ground, secondary (turning those resources into something, and tertiary (services). Labour runs through all three, and it is what wages, taxes and the entire credential economy are built on. Add a fourth leg, AI and automation that do the work instead of helping someone do it, and labour stops being the primary source of value. It becomes one input among several, and no longer the scarce one. The tell is dull too: output rising while headcount stays flat, entry-level and graduate hiring thinning out first, and labour’s share of national income sliding (I think I can already see the first two, based on anecdote rather than data). Consider what that does to this coalition. Its institutions assume a degree is a ticket to a wage, and a wage is the tax base. Pull value away from wages and you pull it away from the tax base, the credential and the redistribution funded by both. The Cushion gets eaten from the bottom while the interest bill eats it from the top. Or the machines grow the pie so large that the arithmetic stops mattering. I do not know which, and I distrust anyone who says they do.

The Case Against Me

The first problem is the monolith. These groups are not one thing. (Jewish Americans are among the most educated groups in the country, and Muslim Americans are, by most accounts, well represented in medicine and engineering. Women are now the majority of college graduates and the backbone of nursing and teaching, two fields with real shortages and real wages.) But I have addressed this by acknowledging there IS generalization involved, but the raw numbers don’t lie. 

So the low-return degree story is not merely about the groups. It is about a layer: a professional and managerial class that writes the rules, selected and trained and paid by a single set of beliefs. The mass of the coalition simply votes on threat. If you read this essay as saying these people are stupid, read it again. I am saying a thin layer sets the incentives and everyone else responds to them sensibly.

The second is the mirror. The other coalition has its own threat premium, built around the forgotten man, the coastal elite and the sense that the country was stolen. It has its own low-return behaviours, its own veto and its own talent for blaming someone else. The incentives are the same, and so is the arithmetic. I would be a poor contrarian if I only put one team under the microscope. Many pundits on Reddit and Tik Tok bemoan the ‘uni-party’ state, and there is merit to that.

The third is the data, which is not on my side. Record median income. (But understand inflation.) A murder rate last seen in the 1950s. Poverty falling. Forecasts of American decline have a poor record: the 1970s had one, the Japan scare of the 1980s had another, and 2008 had a third. I may be adding to the pile. I like a collapse story. It is the interesting option, and the interesting option has led me astray before. I know it does.

The fourth is the strongest, and it is Britain. In 1975 British inflation touched 24 per cent. The IMF was called in the next year, and by the Winter of Discontent the rubbish was piling up in Leicester Square. The country elected Margaret Thatcher in 1979 and began, painfully, to turn. That is a democracy reversing at a threshold well below Chile's. Which is the real argument against my numbers. The trigger in a democracy may be closer than I say, closer to the 8 to 15 per cent band than to hyperinflation. Even so, it took a crisis, and of course one swallow does not make a Summer.

And the Cushion cuts both ways. The same wealth that delays the lesson can also fund a way out. Productivity from new technology, cheap energy, the dollar's privileges and the country's unmatched ability to attract talent could each thicken it for another generation. I cannot rule that out. I can only say that a thicker Cushion is a longer delay, not a different destination.

Where That Leaves Us

Nobody in Xiaogang was brave in December 1978. They were out of options, which is a different thing and a far more reliable engine.

America still has options, and anything can happen in the next two years. But a two-term Democrat President succeeding Trump in 2028 may well be a bridge too far.. I would not call ‘having options’ good news. A cushion is not a cure. It is a longer fall.

Yes, it is pessimistic. But, I believe, with good reason.

Sources: U.S. Census Bureau, Income in the United States 2025 (census.gov/library/stories/2026/09/median-household-income.html) and poverty figures as reported by Fox and Vision Monday; CBO, The Budget and Economic Outlook 2026 to 2036, with interest-to-revenue shares from the American Action Forum's analysis of it; BLS September 2026 CPI as reported by Trading Economics and Investing Live; FBI 2025 crime data as reported by Stateline; Census Vintage 2025 population estimates and Brookings net migration estimates as reported by RealPage and Business Standard; Pew immigrant-share analysis as reported by Fox; NAR first-time buyer data via National Mortgage News; 2026 midterm coverage from Al Jazeera, MS NOW and AP-based reporting. Exit-poll shares are rounded from the outline and should be checked against NBC's exit poll page before publication.