Robots, schmobots. The Rise of the Machines won't leave humanity on the dole

Worstall on Wednesday El Reg tells us that we journos are some of the least likely to have to worry about our jobs being eaten by the robots. Phew, gosh and that's lucky, eh?

Although I'm not really all that certain about this: a Worstall Article Generator, along the lines of the PoMo one, should be easy enough to generate.

Retell story, snark, claim Adam Smith was a bright guy, print. Seems easy enough. And I also work in financial journalism, exactly – as pointed out – where robots are, in fact, making inroads in “writing” what used to be the bread and butter of the trade: company announcement reports.

However, there's something else entirely from the world of economics which means that we don't have to worry about it all that much. The report suggests some 35 per cent of jobs might be destroyed by computerisation over the next two decades, which certainly sounds scary. There's 30 million or so at work at present, so by 2035 there could be 10.5 million roaming the streets, starving, as they've no useful labour they can perform.

However, there's two suggestions from economics that make this really not something to worry about.

The first is that we expect between 200 per cent and 400 per cent of all jobs to be destroyed in the next two decades anyway. Yes, really: there's something called “jobs churn”.

This is a combination of people leaving jobs for other ones and jobs being destroyed and people having to find other ones. Some of this is simply beause management is hopeless and a company goes bust. Some is because someone has a bright idea and a company is growing and sucking in people from elsewhere. Some of it is because technology is changing and thus what people can be gainfully employed to do is changing.

The flow of this through the economy is very much larger than most of us realise. It depends upon exactly how you want to measure it – the major difference is forced moves, as people are fired and companies disappear, and the wider measure of people jumping voluntarily. The lower measure equates to 10 per cent of all jobs in the economy each year and the higher is 20 per cent or so.

This is, as the numerate will note, very different indeed from the number of unemployed people at any one time. Indeed, a useful (although not entirely accurate) rule of thumb is that recessions and unemployment rises happen not when the job destruction rate increases but when the job creation rate falls. Unemployment as recorded being the residual of the flow across, from jobs disappearing to new ones being created.

And there's a further little bit here: the new jobs created are never quite exactly the same as the ones that have been destroyed. There's a tiny bit (and sometimes more than a tiny bit) of technological change in each job movement.