It is a question that is not going away: how much should we worry about intelligent machines stealing our jobs? Yet the debate feels stuck in a rut, fought over the narrow terrain of how many jobs might one day be automated. Is it almost half of jobs, as Oxford university academics Michael Osborne and Carl Frey predict? Or about a tenth, as the OECD believes? The discussion over this number seems to have become our yardstick for how much we should care. But we are so obsessed with this “how many” question that we have forgotten to ask one just as important: “where”?
If we have learnt anything from the economic transformations of the past few decades, it is that geography matters. Take MIT professor David Autor’s work on the impact of the “China shock” on US textile jobs. When China joined the World Trade Organization in 2001, there were fewer than 400,000 people working in American textile mills. As a proportion of the 150m-strong workforce, they were a tiny sliver. But in 57 counties in the south east of the country, they accounted for more than 15 per cent of all jobs. The impact of Chinese competition in these places, says Prof Autor, was “like a mini economic bomb going off over downtown”.
His research found the people who lost their jobs went on to suffer more job churn and lower lifetime incomes. Poverty rose. More men started to die from drugs and alcohol. The economic ripples from the loss of manufacturing jobs hit other local businesses.