In the olden days, when machines were brought in to replace human workers, the workers revolted—the Luddites famously rose up to smash the automated looms that had been deployed by factory owners and were erasing their livelihoods. In modern times, when machines are brought in to replace human workers, the workers vote Trump.
Okay, that’s a bit of an oversimplification, but new research adds to a growing body of work that suggests there’s a distinct link between regions hit by automation and voting Republican, and voting Trump especially. In 2016, areas where industrial robots have eliminated jobs—mostly in the Rust Belt and the South—saw a sizable upswing in voters turning to Trump. (Remember, automation likely played a much larger role in accelerating job loss over the last decades than did other factors like offshoring.)
Now, research at the Brookings Institute again affirms that Republicans are more likely to live in areas ravaged by automation than Democrats. “Our data confirm both a stark history of automation in Trump country and substantial future exposure,” the authors write, “exposure that points to more work flux, more job uncertainty, and potentially more political disruption.”
The reasoning behind the link is fairly straightforward—when people and communities experience widespread economic loss and anxiety, they’re more eager to seek out radical political changes (a la our Luddites). In 2016, that meant voting for the blustery reality TV star who promised to go kick the D.C. establishment in the teeth. Economic anxiety also tends to bring traits like anger at elites and racial animosity to the fore, which synergized conveniently with Donald Trump’s nakedly anti-immigrant platform.
With a Democratic candidate offering little in the way of amelioration to those whose jobs had been automated away, it makes a certain amount of malign sense that voters would turn to Trump to vent their rage. These arguments have been made before, though automation has not been placed as centrally to the phenomenon as it should be. Especially because automation is continuing apace—thanks, somewhat ironically, to Trump’s own signature corporate tax cuts, which enabled businesses to purchase more efficient, labor-and-cost-saving machines—and because it’s continuing to hit Republican districts and Trump country the hardest.
“At the state level, all but one of the ten states most heavily exposed to future job market changes cast its electoral votes for President Trump in 2016,” the report notes (emphasis mine). And all but one of the top 20 congressional districts most susceptible to automation are Republican ones. All but four of the top 50 districts most exposed to automation elected Republicans in 2018. Educated workers are most insulated from automation—as we all know, Trump “love[s] the poorly educated”—and tend to vote Democrat. The 50 districts least-exposed to automation all voted Democrat in the last election cycle.
Which is certainly in line with previous studies on the subject. A paper published in March 2018 by Oxford researchers Carl Benedikt Frey, Thor Berger, and Chinchih Chen, found that “support for Donald Trump was significantly higher in local labor markets more exposed to the adoption of robots.”
They went so far as to assert that “Michigan, Pennsylvania, and Wisconsin would have swung in favor of Hillary Clinton if the exposure to robots had not increased in the immediate years leading up to the election, leaving the Democrats with a majority in the Electoral College.”
Mere months before that paper debuted, the New York Times’ Thomas Edsall proclaimed that “Robots Can’t Vote, But They Helped Elect Trump.” The column referenced a National Bureau of Economic Research study that showed that “commuting zones” home to more industrial robots were the same ones most likely to vote GOP.
“The swing to Republicans between 2008 and 2016 is quite a bit stronger in commuting zones most affected by industrial robots,” MIT’s Daron Acemoglu, one of the NBER researchers, told Edsall. “You don’t see much of the impact of robots in prior presidential elections. So it’s really a post-2008 phenomenon.”
And it’s one that we should be paying attention to. The economy may have stabilized since the fallout of the 2008 financial crash and recession, but it has stabilized in a more precarious form, with gig and part-time work running rampant, and secure, higher-paying union manufacturing jobs giving way to contract jobs and non-union warehouse and service work. In 2014, one in three laid-off workers said automation was to blame, and that kind of distress is viral. To many, the specter of automation looms larger than ever.
“No single factor such as tech-driven worker anxiety determines local political behavior,” the report’s authors write. “But there’s no mistaking that districts that voted Republican in the 2018 election are subject to higher levels of automation exposure.”
Already, we’re seeing Democratic presidential candidates devoting more of their attention to economic anxiety and even automation specifically—one, Andrew Yang, has essentially built his entire campaign around addressing the issue. Both Senators Bernie Sanders and Elizabeth Warren have made securing protections for middle-class workers central planks of their campaigns, in a bid to give the automation exposed an alternative to angry teeth-kicking.
Trump, for his part, barely seems aware that automation even exists.