Robot revolution fails to materialize

Last modified February 16, 2021. Published February 16, 2021.

Robots haven't boosted productivity as much as anticipated. (AP Photo/Carlos Osorio)

Robots have had a relatively minuscule effect on global productivity despite their widespread adoption in the manufacturing sector, according to a new paper by a group of French economists.

In the majority of the 30 countries that the researchers examined, including the United States, robots contributed to less than 0.2% of productivity growth per year from 1975 to 2019. The effect was larger in some countries for subsets of that time period, but never exceeded 1%. 

The paper comes amid conversations around the rise of automation and artificial intelligence, which some researchers say could lead to massive changes in labor productivity, especially when combined with robots. 

But despite such predictions, there is simply not evidence that robots have caused a broad upheaval in productivity so far, even if they have hurt workers in the manufacturing sector, the French researchers found. 

“The effect of robots both on growth and productivity was a bit disappointing,” said co-author Vincenzo Spiezia, an economist who studies the effects of innovation on economic growth. “The contribution of robots to productivity is limited to a few countries.”

Spiezia wrote the paper, which is forthcoming in the March 2021 issue of Economics Letters, alongside Gilbert Cette of the Banque de France and Aurélien Devillard of ASME France. 

While robots are used across nearly every manufacturing sector, they are most concentrated in automaking, a fact that was reflected in the researchers’ findings, Spiezia said.

In Japan, home to colossal car companies like Toyota and Mitsubishi, robots contributed 0.87 percentage points of productivity growth per year from 1976 to 1995 as the country underwent explosive economic growth. In Germany, home to Volkswagen and BMW, robots contributed 0.7 percentage points of productivity growth from 1996 to 2005.

These two subperiods were the high-water marks of robot-related growth, according to the researchers. After 1995, growth in robot-driven auto manufacturing in Japan cooled, and the fall of the Soviets and expansion of the European Union caused German automakers to gradually outsource manufacturing to poorer countries in Eastern Europe, spurring some robot-related growth in those locales at the expense of Germany. 

“Robots are really highly used in car manufacturing, and if you look at the periods of time, they match pretty well,” said Spiezia. 

U.S. automakers like General Motors and Ford have also embraced robots to some extent. However, the fact that many cars designed in the U.S. are built around the world, coupled with the immense size of the American economy, means that any robot-related growth did not significantly show up in the U.S.’s GDP. 

“There might be a lot of robots used for U.S.-branded cars, but it’s produced somewhere else in Asia for instance,” said Spiezia. “The weight is much bigger in Germany, for instance, than it is in the U.S.” 

Currently, about 45% of the world’s robots are used for manufacturing in the transport sector; 30% are used for electronic, electrical and optical manufacturing; 8% are used in the plastic and rubber industry; and 6% are used in the manufacturing of metal products, the researchers said.  

This concentration in a few manufacturing sectors means robots’ impact can vary widely between economies. To account for this in their research process, the economists compared labor productivity — the relationship between hours worked and economic output — to the number of robots per worker used in a given country each year.

Using the raw number of robots per worker is an imperfect measurement, Spiezia said, because it does not account for differences in types of robots’ value and productivity. A robot used in a warehouse may contribute differently to productivity than a robot used in an automobile factory, for example. He said an important step for further research on the topic would be to create a more dynamic measurement of robot productivity that incorporates such differences. 

Spiezia is also interested in examining the impact of robots on job losses. While the most visceral example of robots’ impact may be the blue collar assembly-line worker being replaced by a machine, robots can also impact the labor market in less visible ways. 

“A lot of blue collar jobs in manufacturing have already gone because of robots,” said Spiezia. 

However, he added, “As is usual with technologies, they have two effects: one is to displace jobs, and on the other side, they tend to complement existing jobs.” 

Some people who previously worked in dangerous or dull environments may find that robots can make their jobs safer and more interesting, Spiezia said. 

“There is little work on that and the issue deserves more,” said Spiezia of the relationship between robots and employment. 

The paper, titled “The contribution of robots to productivity growth in 30 OECD countries over 1975–2019,” will be published in the March 2020 issue of Economics Letters. The co-authors are Gilbert Cette of the Banque de France, Aurélien Devillard of ASME France and Vincenzo Spiezia. 

We use cookies to improve your experience on our site and to show you relevant advertising.