A new study by economists Grace Lordan and David Neumark finds that minimum-wage increases make it likelier that low-skilled workers whose jobs can be automated will become unemployed. The study is especially relevant given two recent trends: first, the incorporation of dramatic minimum-wage hikes into the Democratic doctrine, and second, the continued progress of automation. The coincidence of those trends, if Lordan and Neumark are to be believed, does not bode well for low-skilled workers.
The study examines the effect of minimum-wage increases on workers whose jobs consist of “routine” tasks, which “involve a repeated sequence of actions, are easily codifiable, and [are] therefore substitutable with technology.” Such jobs exist across a variety of industries, and are especially abundant in finance, retail, manufacturing, and public administration. Automating data entry, upholstering, or assembly-line work is easier than automating jobs requiring adaptation to unpredictable conditions or abstract problem-solving. And as the unstoppable march of technical progress continues apace, the technology required to automate those jobs becomes more sophisticated — and less expensive.
Raising the minimum wage means raising the cost of labor. But when many workers are performing tasks that can be done more efficiently by computers or machines, management will hesitate to pay them $15 an hour. Sure enough, the authors find, “minimum wage increases cause a statistically significant reallocation of labor away from automatable tasks.”
The average numbers are bleak — “a $1 increase in the minimum wage leads to a 0.43 percentage point decrease in the share of automatable jobs done by low-skilled workers” — but the numbers under the hood are even worse. Older workers in manufacturing are hit especially hard, as are women in public administration and blacks in the transportation and manufacturing sectors. Increasing the minimum wage also causes workers to switch jobs, adding to the precariat. So, too, does it leave them vulnerable to reductions in hours.
Lordan and Neumark see their study as filling a gap in the minimum-wage literature, which they say “usually focuses on very low-skilled workers.” Rather than examining the effect of a wage increase on teenagers, for instance, their analysis takes a look at all the jobs which can feasibly be automated. That’s timely information, and not just for bank tellers: The driverless car, the bricklaying robot, and, further out, Ross the lawyer may render more occupations exposed to the pressures of automation than before.
It’s true that the share of non-routine jobs in our labor force has risen over time as routine jobs have either moved overseas or been automated away. Automation hasn’t cannibalized jobs; it has led to the creation of more opportunities for high-skilled workers. But there’s a swathe of people to whom the modifier “high-skilled” doesn’t currently apply. Balancing the creative destruction brought about by new technology with the need for an employed populace is a difficult challenge that simply raising the minimum wage won’t meet.
This study is evidence that raising the minimum wage in a world of automation means fewer low-wage jobs. If we had a national productivity strategy or a coherent approach to immigration, that could be a good thing. Such a counterfactual requires a possible world far away from the one we live in, however. If implemented here, the Democratic “fight for $15” could cause catastrophe for millions.