The biggest downside impacts of tech-fueled automation may come in the form of major labor market dislocation and social costs connected with that – rather than permanent job losses – said James Bessen, Director and Founder of Research on Innovation at Boston University School of Law, at a Brookings Institution event Dec. 12.
“The real policy challenge is not impending mass unemployment. There is a very disruptive change coming and there are many people who will lose jobs, but lots of jobs are going to be created,” Bessen said at the Brookings event focused on automation and the labor market. “Workers have to transition to new jobs. They have to get new skills, often move to a new occupation.”
Managing the social burden of that labor market transition, Bessen asserted, ought to be top of mind for policymakers as the automation trend heats up further.
Bessen explained why automation isn’t causing mass layoffs, citing simple economic principles: automation reduces the demand for manual labor; the reduced labor requirement cuts the prices of goods; and lower prices encourage more consumption.
For example, he said, the automated production of cloth didn’t end up destroying the clothing industry. Rather, instead of only owning a few outfits, consumers now fill entire closets with clothes.
“What gets automated are tasks, not jobs – very rarely is an occupation completely automated,” Bessen said. “Jobs require many diverse skills, and technology can automate some of these tasks but not all of them.” Certain occupations, like elevator operator, are some of the few exceptions to that rule, he said.
As evidence for his case, Bessen discussed a scholarly analysis of Dutch data on automation expenditures. It found that the first year after a company undertakes a major investment in automation, about two percent of employees will leave the firm. The study was unable to distinguish whether these workers leave because of layoffs or personal choice, but it did find that most often employees switch to a different industry.
Bessen called this effect “significant but not overwhelming.” He said that job loss due to automation varies by industry, but that income generally declines for all workers when automation is introduced to a firm.
This transition period is where policymakers can help the most, he said. Unemployed and underemployed workers affected by automation only make up 13 percent of the income they lost from government-assisted programs, according to the study Bessen shared.
What policymakers also should be aware of, Bessen said, is economic inequality that could spring from automation in the form of large firms increasing market domination. He said the transitioning labor market could benefit from job training, apprenticeships, certifications, and temporary social support as industries change. Encouraging the labor market through knowledge sharing, expanding trade secrecy laws, and promoting occupational licensing could stave off the fears associated with automation, he suggested.
