Britain has promised to retrain its workers out of every crisis for fifty years. The record is more sobering than the promise.

The idea that you answer lost work with training is not new, and it is worth knowing how old it is before we reach for it again.

Britain built the machinery for it in 1974, when the Manpower Services Commission was set up to co-ordinate training as the post-war industrial settlement began to crack. Its most famous creation came in 1983: the Youth Training Scheme, launched as the pits and the factories closed, with around a billion pounds behind it and more than three hundred thousand young people passing through it a year. The promise was the one we still make. The old work is going; we will train you for the new.

What did it deliver? Honestly, a mixture. For some, the scheme was a real bridge, a first placement that became a job. For others it became a holding pen, a year stacking shelves for an allowance that critics called cheap labour, and evaluations from the time found the actual skill-building was often thin. In the coalfields, where whole communities were told retraining would carry them into whatever came next, the memory that survives is mostly of what did not arrive. The scheme's biggest single achievement may have been statistical: it kept hundreds of thousands of young people out of the unemployment count at a moment when that count was politically unbearable.

The instinct did not retire with Thatcher. New Labour rebuilt it as the New Deal, to move a workless class into work. The governments after rebuilt it again as apprenticeship levies and skills funds. And here we are once more, with reskilling for the AI age, the same reflex wearing this decade's clothes.

The lesson of fifty years is not that retraining fails. It is narrower and more useful than that. Retraining works when three things are true: it is patient, it is tightly aimed, and there is a real and growing kind of work at the far end of it. When those hold, it changes lives. When they do not, it becomes something else, a way for a government to be seen to act, to move a difficult number, to buy time. The uncomfortable question for our own moment is the third condition. Every previous scheme, for all its faults, could point to industries that were genuinely hiring: offices, services, retail, the whole white-collar economy that absorbed the people the factories let go. That economy was the far bank of the bridge. It is also, this time, precisely the ground the water is rising over.

History does not tell us retraining is hopeless. It tells us it has always depended on a destination, and that the one thing no scheme ever managed to manufacture was the jobs themselves.