Britain did not count the unemployed until it had to pay them. It has been choosing what to see ever since

It is worth knowing that for most of British history the state had no real idea how many of its people were out of work. Not a rough idea. None. Before 1911 there were, broadly, no government unemployment statistics at all.

What existed instead came from the trade unions. A skilled union paid its members a small sum when they were out of work, kept a ledger of who was claiming, and passed the totals to the Board of Trade. So the earliest picture of British unemployment, the figures historians still reach back for, running to the 1880s, is really a picture of a narrow, mostly skilled, mostly male sliver of the workforce that happened to belong to a union that happened to pay a benefit. Everyone else was invisible, because no one yet had a reason to look.

The reason arrived with a cheque. In 1911, Lloyd George's Liberal government passed the National Insurance Act, and Part II of it created the world's first compulsory unemployment insurance, the work, in large part, of a young Winston Churchill at the Board of Trade. It covered around two and a quarter million workers in three trades where the work came and went: building, shipbuilding, engineering. And the moment the state began paying men for being out of work, it began, for the first time, to count them. Not out of any curiosity about the labour market. As the bookkeeping of a benefit.

That is the pattern worth carrying forward, because it has never really changed. The state counts what it has to answer for. It measured the unemployed once it was writing them cheques, and not a moment before.

You can see how partial that instinct was in what followed. Britain paid unemployment benefit from 1911, but did not carry out a proper national census of the out-of-work until 1931, two decades later. For twenty years the country acted on a category it had not properly measured, because paying a defined set of claimants was one thing, and seeing the whole picture quite another, and only one of them was ever forced on it.

And once it was counting, it kept moving the line. The definition of who officially counts as unemployed has been redrawn again and again; by one retrospective tally, more than thirty times between 1979 and 1996 alone. Each change was defended on technical grounds. Each also happened to move the number. Counting, it turns out, is never neutral: who you decide to count is a decision about whom you are prepared to be responsible for.

Which brings us back to now. Britain does not count the disappearing first rung: the apprenticeship not offered, the graduate job never posted, the junior role quietly folded into a machine. We reported last week that when someone asked ONS for the figure, it wrote back to say it does not collect it. The history suggests this is not an oversight to be fixed with a better spreadsheet. The state has simply not yet decided that the vanishing of entry-level work is something it must answer for. The uncounted are never a random group. They are the people a country has not yet been made responsible for, and right now, the first job is one of them.