We have yet another study that debunks the widespread notion that robots — and other forms of automation, including “artificial intelligence” — will destroy our jobs and lead to a future of permanently high unemployment. According to the study, that would completely rewrite history, which has shown job creation to be an enduring strength of the U.S. economy.
The study (”False Alarmism: Technological Disruption and the U.S. Labor Market, 1850-2015”) comes from the Information Technology & Innovation Foundation, a Washington think tank focusing on technology. The study’s greatest virtue is to remind us that past changes have wiped out entire job categories without spawning a high-unemployment society.
This doesn’t mean that unemployment isn’t a problem. Obviously, it is, especially during recessions or for workers whose jobs have been lost to new technologies, products or competitors. But we need to remember that the U.S. economy has enormous recuperative powers. Indeed, the study argues that technology’s disruptive effects on today’s job market are much less than in the past.
I have written about this before, but the subject is worth revisiting because it’s encrusted in misleading myths. Using a database of occupations (examples: farmers, nurses, engineers) developed at the University of Minnesota, the study’s authors, Robert D. Atkinson and John Wu, examined how technology affected occupations in each decade since 1850. The more changes — either more or fewer jobs — the more disruptive technology’s impact.
There is often a cycle. A new technology creates or eliminates jobs. The effect continues or accelerates for some years or decades. Then the market matures, or other technological changes intrude. Jobs stabilize or decline.
Take elevator operators. In 1860, there were none, because builders hadn’t learned to construct tall structures with steel. Once this was mastered, the demand for elevator operators soared. By 1950, they totaled 114,473. But in the 1920s, Otis Elevators had developed self-service elevators, which spread after World War II. By 1990, the number of elevator operators was “essentially zero,” Atkinson and Wu report.
Or consider motion-picture projectionists. Their number rose in the early 20th century, reaching a peak of 31,000 in 1940. But in the 1950s more “Americans chose to stay home and watch TV,” while in the 1970s and 1980s, more movie theaters became “multiplexes [where] . . . one projectionist could now manage more than one projector.” Both developments reduced the need for projectionists. In 2015, there were slightly fewer than 5,000.
Technological change — both gaining and losing jobs — was disruptive but, in the long run, not destructive. The economy’s employment base adjusted. Technological changes created consumer demand for new goods and services, which sustained employment growth and offset job losses.
The great fear today, Atkinson said in an interview, is that this will change. Technology will eliminate jobs and not replace them. People worry that their “occupational capital [their job skills and knowledge] will be destroyed.” Machines will permanently substitute for people. “That’s what the debate about robots and artificial intelligence is all about,” he said.
But there is little precedent for this outcome. The reality, said Atkinson, is that new technologies have always jolted job markets. Individual industries and markets ebb and flow — often hurting workers and communities — but the larger process continues. Job losses in one sector are offset by job gains in another.
Consider the 1970s. During that decade, according to Atkinson, the number of farmers dropped 12 percent, and the number of telephone operators and typists both declined by 41 percent. But the overall job market expanded by roughly a quarter, or about 20 million jobs.
The lesson of history seems to be this: The robots won’t steal all our jobs, because their efficiencies will create more purchasing power for other spending or new products that require human involvement and oversight. For proof, consider smartphones. In 2012, they had created nearly 500,000 jobs related to “mobile apps,” up from zero in 2007.
Automation is not a new phenomenon. Understandably, it inspires fear. But so far in the United States, it doesn’t kill job growth.
Read more from Robert Samuelson’s archive.