Income tax is a tax on going to work. Britain has never had to ask what it does if there is less of that

This desk got three things wrong about the OBR's fifty-year report yesterday, because it read a news agency's account instead of the report. The correction is published. I have now read the document.

The spending half is what everyone argued about. State pension spending goes from about 5 per cent of GDP to about 9 by 2075-76, or nearer 7 if you swap the triple lock for average earnings. The only fraction the OBR gives the lock is a fifth, and it is a fifth of the rise in all spending, not of the pension bill. Anybody calling it "about a third" of the pension rise, this newspaper included, is quoting something the OBR did not write.

The receipts half is the interesting one, and almost nobody read it.

Be precise about how the state is funded, because the imprecise version falls over on contact with any accountant.

National Insurance is a tax on employing somebody. This year an employee hands over 8 per cent of what they earn between £12,570 and £50,270, and 2 per cent above. On top, their employer hands over 15 per cent of almost everything above £5,000, with no ceiling at all. It is charged on pay. It is not charged on a dividend.

Now the objection, before somebody makes it for me. Profits are not lightly taxed. A company pays corporation tax, and when what is left reaches a shareholder, dividend tax is charged on top: 10.75, 35.75 or 39.35 per cent this year. Stack the two and the gap against a salary narrows.

What survives is narrower, and it is enough. National Insurance has no counterpart on dividends, rent or capital gains. The wage is taxed on the gross; the company on what is left after costs and allowances. And profit has places to go that a wage does not: retained in the company, held in a pension or an ISA, paid to a shareholder overseas whom HMRC cannot reach. A wage has nowhere to be. It is deducted at source, before it lands.

The state does not need capital to go untaxed to have a problem. **It only needs the money to stop arriving in the one form it cannot escape.**

Then, in the middle of the receipts chapter, the report does something forecasters do not usually do. On page 86 it opens a box called "The potential impact of AI on the labour share and tax receipts."

Somebody has put a number on it, and it is in the box.

The OBR runs an illustrative scenario. Wages and salaries fall from 40 per cent of GDP to 20 per cent by 2075-76, with total GDP unchanged. Tax receipts fall from about 38 per cent of GDP to under 35: nearly four percentage points. The mechanism is exactly the one above. The average effective tax rate on wages and salaries over that projection is 46 per cent. On profits it is 27 per cent. The extra tax collected on the profits does not come close to covering the tax lost on the pay.

Two honesties are owed. That scenario is not the OBR's baseline, which assumes the shares hold. And the OBR calls it highly stylised, and says the fall it models sits at the upper end of what the literature supports.

But note what it does not depend on. Not a recession. Not mass unemployment. Not a smaller Britain. Only a different answer to the question of who gets paid.

**Why it matters here.** Two things in this report already touch your money, and neither is speculative. The state pension age moves from 66 to 67 across 2026 to 2028, so the year you can claim depends on the month you were born. And income tax thresholds are frozen, so ordinary pay rises quietly move people into higher bands. This is information about how the system works. It is not advice about what to do with your money, and this desk does not give any.