Richard Florida is a co-founder and editor at large of CityLab and a senior editor at The Atlantic. He is a University Professor and Director of Cities at the University of Toronto’s Martin Prosperity Institute, and a Distinguished Fellow at New York University’s Schack Institute of Real Estate.
In Dalton, Georgia, more than two-thirds of jobs are at risk for automation, according to a new analysis.
Dalton, Georgia, has long been known as the “carpet capital of the world.” In the early 2000s, Dalton produced almost half of the world’s carpet. Walking down the city’s carpet row—a stretch of carpet mills that spans a portion of Interstate 75—you could see miles of carpet mills, whirring activity as wool was spun and knots were tied. There were more than enough jobs for anyone who needed work.
Things are very different today. The economic crisis of 2008 hit Dalton’s carpet industry hard. Under the pressure of global competition, its carpet manufacturers laid people off and replaced human labor with robots. Although it still remains the nation’s “Carpet Capital,” that no longer means good jobs that require a low skill level to perform.
If the present is bleak, the future may well be bleaker. According to new analysis by my colleague Shade Shutters, of Arizona State University, more than two-thirds of Dalton’s jobs—37,574 out of 55,400—are at risk from automation, the highest proportion of any metro in the country. Shutters used occupational data from the Bureau of Labor Statistics to identify the types of jobs most at risk from automation across all US metros.
If the previous era of “robot shock” struck hardest at manufacturing workers and communities, the next wave promises to hit as hard or harder at even more vulnerable and lower-paid service workers and their communities across the nation. Indeed, the jobs most at risk from future automation are routine service positions like waiters and bartenders, taxi and limousine drivers, and retail jobs in clothing stores, department stores, sporting good and musical instrument shops, and jobs in the motion picture and video industries. Each of these positions are significantly more at risk than blue-collar jobs in the auto industry. Shutters’ findings extend this catalogue of at risk jobs to account for the people and places that are most vulnerable to automation.
While these findings conform to some broad and well-documented geographic patterns—the metros most at risk tend to be in the Rustbelt and the Sunbelt, while those least at risk are on the East and West Coasts—the divides themselves are far more granular. The Bay Area metros of San Francisco and San Jose are amongst the least vulnerable to automation, while Madera, California, just three hours away, and is among the most vulnerable metros. Furthermore, more than half (54 percent) of voters in Madera County voted for Trump, compared to just 9 percent who did so in San Francisco County.
The map above, made by Taylor Blake of the Martin Prosperity Institute, charts the results of Shutters’ analysis. It details the risk of automation across America’s metros: Dark purple indicates the metros that are most susceptible to automation, while lighter blue shows those that face the lowest risk. The map of places most at risk from automation spans the entire country from parts of New England and Pennsylvania down through central and southern Florida, across the Gulf of Mexico and into Texas, upward through parts of the Midwest and Great Plains and out West to Nevada, Washington and even parts of California.
The table below shows the 20 metros most at risk from automation. In these high-risk metros, roughly two-thirds of jobs are at risk from automation. This list of high-risk places draws heavily from small Rustbelt and Sunbelt metros with a deep history of manufacturing.
Metros Most at Risk from Automation
|Rank||Metro||Number of Jobs at Risk||Percentage of Jobs at Risk|
|9||Michigan City-La Porte, IN||21,132||64.9%|
|16||Florence-Muscle Shoals, AL||26,375||63.9%|
|17||The Villages, FL||11,634||63.9%|
|18||El Centro, CA||28,418||63.9%|
Dalton tops the list of the metros most at risk from automation, followed by Kokomo, Indiana. The list also includes hard-hit Rustbelt manufacturing centers like Elkhart, Indiana; Michigan City, Michigan; and East Stroudsburg Pennsylvania; and smaller Sunbelt metros like Daphne, Alabama; Sumter, South Carolina; Burlington and Jacksonville, North Carolina; Hammond, Louisiana; Florence-Muscle Shoals, Alabama; and The Villages, Florida, frequently mentioned as America’s fastest growing community.
The burden of automation falls the hardest on smaller metro areas (those with less 250,000 people). Here, 19 of the top 20 most vulnerable metros, and 87 of the top 100 metros most at risk from automation are those with less than 250,000 people.
In contrast to the burden faced by small metros, only one large metro (with more than one million people)—Las Vegas—faces a similar risk. In Las Vegas, nearly 63 percent of jobs are at risk from automation. Interestingly enough, the large metros that are at greatest risk from automation are all in the Sunbelt: Orlando, Jacksonville, and Miami Florida; Riverside, California; Louisville; Memphis; San Antonio; Birmingham; and New Orleans.
Conversely, the metros that face the smallest risk of automation are mainly larger, knowledge based metros and college towns, the table below shows. Here we see college towns like Durham-Chapel Hill, Ann Arbor, Boulder, Corvallis, and Charlottesville; and large knowledge-based metros like San Jose and San Francisco, Washington DC and nearby Baltimore, Boston, and New York.
Metros Least at Risk from Automation
|Rank||Metro||Number of Jobs at Risk||Percent of Jobs as Risk|
|370||Sierra Vista-Douglas, AZ||13,65||52.0%|
|377||San Francisco-Oakland-Hayward, CA||1,092,104||50.9%|
|378||Hartford-West Hartford-East Hartford, CT||284,449||50.7%|
|384||Ann Arbor, MI||87,863||47.9%|
|386||San Jose-Sunnyvale-Santa Clara, CA||458,363||46.7%|
|387||California-Lexington Park, MD||16,688||45.4%|
|388||Durham-Chapel Hill, NC||125,510||45.2%|
Even though they don’t make this top twenty list, larger Rustbelt metros face comparatively modest risk from automations, with 53 percent of jobs in Detroit being at risk from automation, 56 percent in Pittsburgh, 55 percent in Cleveland, and 57 percent in Buffalo.
Demographics and Automation Vulnerability
To get even deeper into what drives vulnerability to automation at the metro level, my colleague, Charlotta Mellander, ran a basic correlation analysis between the share of jobs at risk across metros, and key economic and demographic characteristics.
The share of workers employed in knowledge-based, artistic, and creative jobs appears to be the most significant factor in mitigating the risks of automation. It has the strongest negative correlation to the share of jobs at risk from automation (-0.84). On the flip side, metros with larger working classes are significantly more vulnerable to automaton. There is a significant positive correlation (0.53) between the share of workers employed in blue-collar working-class jobs and the share of jobs at risk from automation.
Politics factors in as well. Trump promised to “Make America Great” again by bringing manufacturing jobs back to the heartland. The share of residents who voted for Trump is positively associated with the risk of job loss from automation (0.50), while the share of people who voted for Clinton is negatively so (-0.47). This can be seen in the scatter-graphs above and below, which show the association between the share of Trump and Clinton voters on the one hand, and the share of jobs at risk from automation on the other.
A number of other factors also made metros less vulnerable to automation: Larger size, denser population, more education, more wealth, and concentration of the high-tech industry were all correlated with lower automation. One factor that made metros more vulnerable was the share of workers who drive alone—a key indicator of sprawl.
Like most everything else in America today, automation—and the risk of job loss that comes with it—reflects the deepening spatial inequality that vexes and divides our nation, hitting hardest at the most vulnerable groups in the least advantaged places.