Automation Is a Job Engine, New Research Says

Photo
A new study on the effects of automation on the job market points to data that shows an increase in the number of bank tellers despite the pervasiveness of automated teller machines. Credit Marcelo Del Pozo/Reuters

The fear that technology is poised to kill jobs in unprecedented numbers is widely prevelent these days. Nothing is likely to ease that anxiety much, but a new research paper might prompt some second thoughts.

Using government data, James Bessen, a researcher and lecturer at the Boston University School of Law, examined the impact of computer automation on 317 occupations from 1980 through 2013. His conclusion, in a sentence, was: “Employment grows significantly faster in occupations that use computers more.”

Historically, it is well established that the advance of technology has generated more jobs than it has replaced, regardless of the angst of the moment. More than 80 years ago, the renowned English economist John Maynard Keynes warned of the “new disease” of “technological unemployment.”

Yet the paper by Mr. Bessen provides a detailed study of technology and jobs in recent years. “The idea that automation kills jobs isn’t true historically, and if you look at the last 30 years, it’s not true then either,” he said in an interview. “Right now, the best thing that can happen to you is to get some automation to do your job better.”

Mr. Bessen is scheduled to discuss his research on Monday at a conference in Greece, sponsored by the European Centre for the Development of Vocational Training, an agency of the European Union.

Mr. Bessen’s conclusions are based on occupation, salary and computer-use data from decennial, yearly and monthly surveys, mainly from the United States Census Bureau and the Bureau of Labor Statistics. In his paper, he also describes the effect of new technology on a handful of occupations.

Automated teller machines, Mr. Bessen writes, were widely adopted in banks starting in the 1990s and were pervasive by 2005. Yet since 2000, he notes, the number of bank tellers increased 2 percent a year, “substantially faster than the entire labor force.” What happened, he explains, is that A.T.M.s reduced the cost of operating bank branches. So banks greatly expanded their branch networks, with increased demand more than offsetting the jobs displaced by technology.

But something else occurred as well. Tellers were no longer just handling cash but also offering personal advice and selling new services to customers. “The nature of the occupation is changing,” Mr. Bessen said.

His comment echoes the main finding of a recent report from the McKinsey Global Institute, which concluded that many work tasks within jobs can be automated in the next three to five years. But the impact, according to the McKinsey report, will be to alter jobs rather than eliminate them.

Mr. Bessen’s research and the McKinsey report do not necessarily counter the more grim forecasts of the future impact of automation. Carl Benedikt Frey and Michael A. Osborne, researchers at Oxford University, projected in a paper published two years ago that 47 percent of American jobs were at risk from automation. But they were looking to a further horizon, 20 years from now.

The other big variable is the course of technological progress, especially in the field of artificial intelligence, and whether it will work more to improve human capabilities or replace them. There is plenty of debate on that front.

Three excellent books on robotics, published recently, each take up that subject and come to somewhat different conclusions: “The Rise of the Robots,” by Martin Ford; “Humans Need Not Apply,” by Jerry Kaplan; and “Machines of Loving Grace,” by John Markoff, who is a reporter at The New York Times.

In his research, Mr. Bessen found that while technology did not eliminate jobs over all, it was a source of turmoil in the labor market. It is a force for wage inequality as computer-intensive occupations displace other occupations. Desktop publishing, for example, greatly increased the demand for graphic designers, but they often replaced typographers in the 1980s.

The policy implication of his research, Mr. Bessen said, is to promote public policies that make it easier for workers to get vocational education and move to new jobs. By contrast, if technology becomes an inexorable job killer across the economy, then policies like a guaranteed minimum wage begin to look more appealing.