This article begins a three-part exploration into how robots and automation influence employment, drawing from recent findings by economist and Institute Professor Daron Acemoglu.
Across numerous American regions, automated systems have been substituting human workers throughout recent decades. However, the actual scope of this transformation remains unclear. While some technology experts predict a future where human labor becomes obsolete due to automation, other analysts remain doubtful about such extreme scenarios.
A groundbreaking study co-authored by an MIT professor now provides concrete data on this trend, revealing substantial impacts—though falling short of a complete robotic takeover. The research also highlights significant variations in how robots affect different industries and regions across the United States, potentially playing a meaningful role in widening income inequality.
"Our research demonstrates significant negative employment consequences," explains MIT economist Daron Acemoglu, who cautions against exaggerating the trend's overall impact.
Between 1990 and 2007, the research indicates that introducing one additional robot per 1,000 workers decreased the national employment-to-population ratio by approximately 0.2 percent, with certain American regions experiencing considerably more severe effects than others.
This translates to each additional robot implemented in manufacturing displacing roughly 3.3 workers at the national level, on average.
The increased workplace automation also depressed wages by about 0.4 percent during this identical timeframe.
"We identified negative wage effects, showing that workers in more heavily affected areas experience real wage reductions because robots compete effectively against them," Acemoglu notes.
The research paper, titled "Robots and Jobs: Evidence from U.S. Labor Markets," has been published in advance online form in the Journal of Political Economy. The authors are Acemoglu and Pascual Restrepo PhD '16, currently serving as an assistant professor of economics at Boston University.
Detroit's Automation Challenge
To execute this study, Acemoglu and Restrepo utilized data from 19 industries, compiled by the International Federation of Robotics (IFR), a Frankfurt-based industry organization that maintains comprehensive statistics on global robot deployments. The researchers integrated this information with American data on population, employment, business operations, and wages, sourced from the U.S. Census Bureau, the Bureau of Economic Analysis, the Bureau of Labor Statistics, and other authoritative sources.
The study also compared robot implementation in the United States with other nations, finding that America lags behind Europe in this regard. From 1993 to 2007, American companies introduced almost precisely one new robot per 1,000 workers; meanwhile, European firms implemented 1.6 new robots per 1,000 workers.
"Despite the U.S. being a technologically advanced economy, it trails many other developed nations in industrial robot production, usage, and innovation," Acemoglu observes.
In the United States, four manufacturing sectors account for 70 percent of robot implementation: automobile manufacturing (38 percent of deployed robots), electronics (15 percent), plastics and chemical industry (10 percent), and metal fabrication (7 percent).
Throughout the nation, the research examined robot impacts across 722 commuting zones within the continental U.S.—essentially metropolitan areas—and discovered substantial geographic differences in robot utilization intensity.
Given industry-specific robot deployment patterns, the country's most affected region is the heart of the automobile industry. Michigan displays the highest workplace robot concentration, with employment in Detroit, Lansing, and Saginaw experiencing greater impacts than anywhere else nationally.
"Different industries maintain varying regional footprints across the U.S.," Acemoglu points out. "The robot issue manifests most prominently in Detroit. Developments in automobile manufacturing disproportionately affect the Detroit area compared to other regions."
In commuting zones where robots joined the workforce, each machine replaced approximately 6.6 jobs locally, according to the researchers. However, in an interesting economic dynamic, adding robots to manufacturing benefits workers in other industries and regions—partly through reducing goods costs. These national economic advantages explain why researchers calculated that adding one robot replaces 3.3 jobs countrywide.
The Inequality Connection
While conducting their investigation, Acemoglu and Restrepo extensively examined whether employment trends in robot-intensive areas might stem from other factors, such as trade policies, but found no confounding empirical effects.
The research does indicate, however, that robots directly influence income inequality. The manufacturing positions they eliminate typically belong to workforce segments with limited alternative employment opportunities; consequently, automation in robot-utilizing industries directly connects to declining incomes among blue-collar workers.
"Significant distributional consequences emerge," Acemoglu states. When manufacturers implement robots, "The burden primarily affects low-skill and particularly middle-skill workers. This represents a crucial aspect of our broader robotics research—that automation constitutes a major technological factor contributing to increasing inequality over the past three decades."
Therefore, while predictions about machines completely eliminating human work may be exaggerated, the research by Acemoglu and Restrepo demonstrates that robot effects in manufacturing are substantial, carrying significant social implications.
"Our findings certainly won't support those who believe robots will eliminate all human employment," Acemoglu concludes. "But they do indicate that automation represents a genuine challenge requiring serious consideration."
