Why Computers Don’t Replace Workers

“Technology replaces workers, and we should be very afraid.”

Writers continue to spin this time-worn fallacy with “click-bait” articles that add fuel to an ongoing hysteria of Luddism. Thankfully, economic studies help to demystify these bogus claims.

Technological innovation spurs productivity gains, thereby encouraging economic growth. Of that much, we are certain. By studying the linkages between computer use and occupational categories, James Bessen, a professor at Boston University, finds that employment grows faster in occupations that use computers more, as compared to occupations with lower computer use. Occupations that use computers can attribute up to 0.9 percentage points of their employment growth to the use of computers. As a comparison, national employment grew on an average of 1.8 percent during the 1990s.

Why is this important? Naysayers make the case that jobs get eliminated when organizations adopt information technology, like computers. By this logic, occupations that use more computers should see employment grow slower, or even experience negative growth. But Bessen finds that computer use triggers an evolution of skill requirements, thereby increasing the need for some positions while reducing the need for other position within the same occupational category.

He cites the textbook case of automatic teller machines’ effect on bank tellers during the late 1990s. Prior to ATMs, bank tellers basically performed the routine task of counting money. But the spread of ATMs also coincided with bank teller employment increasing 2 percent yearly, higher than the annual job growth rate. Because costs of operations went down, banks were able to open more branches, and concurrently, the occupational skill needs of a bank teller shifted towards more abstract tasks.

Bessen finds that occupations that rely more on abstract tasks adopt computers faster than occupations that have more routine tasks nowadays. Routine tasks are typically codified as short, repetitive tasks without any face-to-face interactions, while abstract tasks require problem solving tasks, and managing or supervising other workers. He suggests that such could be the case because initial waves of computer adoption targeted and were successful at automating routine tasks, but newer innovations in computer technology targets enhancing the productivity in more abstract task-intensive occupations. He points to the late 1990s where most of the gains to automating routine tasks were fulfilled and computer use became a more important feature for occupations heavy in abstract tasks.

Overall, Bessen’s findings point to the reality that computer use doesn’t affect the total number of jobs in the economy. At the economy level, occupations that use computers substitute for occupations that do not. As with the case of bank tellers, technological change changes the nature of the jobs, and workers need to adapt and pick up new skills in a changing economy.