This paper has one tune and yesterday it played it again. Nobody is counting, nobody is counting, nobody is counting

Yesterday my colleagues found a gap and filled a paper with it. Let me do my job.

The first problem is that we always find this

Search this paper's archive for nobody is counting or nobody is measuring and you get fourteen pieces since July. Care, entry-level jobs, retraining, legal services, the gain from productivity, driverless taxis, and yesterday, redundancy forms. A paper that finds the same absence everywhere should ask whether it is discovering something about Britain or about its own method.

The second problem is that the speech was not the place

A leader at the United Nations is addressing other governments about frontier risk. Nobody stands at that podium to announce a change to a redundancy form. Judging a speech by what it left out is a cheap trick, and we would call it out if another outlet did it.

The third problem is the one that matters

Counting is cheap to demand. It sounds unanswerable, because who could be against knowing? But the thing being demanded here is hard. Attributing a particular redundancy to AI requires someone to decide what caused a decision that had many causes, and the people filling in the form are the people with the strongest interest in the answer. Put an occupational breakdown back on the HR1 and you would get a form full of restructuring and efficiencies, exactly as you do now.

Meanwhile the instrument this paper called the best in the system yesterday, payroll data, already tells you that 145,000 payrolled jobs went this year and that almost exactly that many were people aged 25 to 34. We reported it ten days ago. That is not nothing. That is a country that can see the shape of what is happening and is arguing about the cause.

Where I end up

I think the specific findings hold. The question was removed. The entry-level figure does not exist. The assessment has not been repeated. Those are facts and they are ours.

What I do not accept is the implied conclusion, that a government which measured better would act better. Britain measured the decline of manufacturing in enormous detail for thirty years. The numbers were excellent. The towns still went.