How it actually works
Start with the parable from The spiral, and why UBI cannot stop it. One machine makes everything. It employs nobody. It belongs to one person. Nobody has a job, so nobody has income, so the government taxes the machine and sends everyone a cheque.
Follow one cycle. The state hands out $100. People spend it on goods from the machine. The state taxes the owner at a punitive 90 per cent and recovers $90. The owner keeps $10.
Next cycle the state has $90 to distribute. Then $81. Then $73. Nobody cheated, nobody evaded, nobody lobbied. The owner is simply entitled to a share, and a share taken out of a closed loop every time round drains it. Run the series to its end and the owner has quietly accumulated the entire original $100, one tenth at a time, while the payment to everyone else has fallen to nothing.
What people get wrong
The tax rate only changes the speed. At 50 per cent the disbursement halves every cycle — $100, then $50, then $25 — and the thing is over in a handful of rounds. To keep the payment steady at $100 forever, the state has to find the missing 10 per cent somewhere outside the loop every single cycle, by borrowing, printing, or taxing something that is also shrinking. The only rate at which the money comes back whole is 100 per cent, at which point the owner owns nothing in any sense that matters and you have nationalised the machine while keeping the share certificates as decoration.
The leak is not really caused by taxation, either. It is caused by the owner's surplus having nowhere to go. In an ordinary economy, retained profit comes back into the through two doors: the owner spends it, or the owner invests it. Both doors pay wages to somebody, and that is why the loop normally closes. A machine that needs no labour and no inputs bricks up both doors. Investment buys more machines that hire nobody. Consumption buys output from a producer that hires nobody. The surplus accumulates as a claim on a shrinking economy, and the state is left doing the entire job of putting money back into people's hands with a bucket that has a hole in it.
The numbers
Scale first. Displace a quarter of the US workforce and, at the average wage, roughly $3.2 trillion of annual income disappears. But the occupations most exposed to a machine that works through a screen are the well-paid ones — management work averaging $144,860 a year, legal occupations $129,880, computer and mathematical occupations $108,620, against an all-occupation mean of $69,770 — so the realistic figure is nearer $4 trillion. Total federal receipts were $5.24 trillion in FY2025. Replacing the lost income alone costs 60 to 80 per cent of everything the federal government collects, before anything else the government does.
Then the question of where the money comes from. Corporate income tax raised $497 billion in FY2025. Confiscate all of it, at a 100 per cent rate with no allowances and no avoidance, and you have covered roughly an eighth of the bill. Individual income tax — $2.60 trillion, or 49.6 per cent of all federal revenue — is paid overwhelmingly by exactly the people being displaced. Every worker replaced is simultaneously a new claimant and a lost taxpayer.
Why it matters to this crash
UBI is the standing answer to every question about mass displacement. It is what gets said in the last thirty seconds of the interview, and it is treated as an expensive but available option: a matter of political will and a tax bill. Our argument is that it is not available at all in the form usually imagined, and that the it is supposed to interrupt runs anyway, more slowly.
What would make this dangerous
The observable signature is a divergence between output and payroll tax receipts: GDP and corporate revenues holding up while individual income tax collections fall, because the displacement is concentrated in the high-earning screen occupations that carry half the federal revenue base. Employment counts will lag this. Withholding will not.
Watch, too, for corporate tax receipts failing to rise in step with the output that automation is supposed to deliver. Our reading of the boom is that AI succeeding dissolves barriers to entry rather than building them, which competes away the very profits a transfer scheme would be taxing. A world of cheap goods, high output and thin margins is a world with a large redistribution bill and no obvious base to levy it on.
The most dangerous version is the half measure: a UBI legislated at a fraction of lost wages, funded by deficit. It slows the spiral without stopping it, and it converts a demand problem into a compounding fiscal one, which is where it stops being an argument about labour economics and becomes an argument about .