Stories

Robots may have given Trump an edge in 2016

Illustration: Sam Jayne/Axios

For two years, historians, economists and others have pondered whether western leaders, facing a growing populist challenge, must prepare for an even greater temblor resembling the French Revolution or 1930s fascism.

The big picture: In a new paper in the Oxford Review of Economist Policy, U.K. economist Carl Frey and two co-authors argue that the 2016 U.S. presidential election — and the effects of industrial automation during the decades before — may be a signal of worse to come.

Frey, co-author in 2013 of one of the most-cited papers about automation and jobs, writes that since the start of the Industrial Revolution in about 1780, new technology has transformed living standards but also "bred many political revolutionaries."

  • "The Industrial Revolution began with the arrival of the factory," they write, but "it ended not just with the construction of the railroads, but also with the publication of the Communist Manifesto."
  • A cause of political backlash: Three generations of ordinary Britons over six decades in the Industrial Revolution saw no benefit from mechanization. Instead, skilled craftsman were thrown out of work and replaced often by child laborers.
  • One shocking indicator of their suffering: Men in 1850 were shorter than they had been in 1760, they write.

Frey found similar statistics in the years leading up to the 2016 election. "The trajectories of the American economy over the four decades following the revolution in automation of the 1980s almost exactly mirror the first four decades of the Industrial Revolution in Britain," the authors write.

  • From 1979 to 2013, productivity rose by 65% but hourly wages for 80% of the workforce were up by just 8.2%. Since they were also working more hours as well, they may have lost ground.
  • And from 2000 to 2013, wages for 70% of workers were either flat or fell.

To assess any link between automation and the 2016 election, the authors used local mechanization data from the International Federation of Robotics and county-level voting data from Dave Leip's Atlas of U.S. Presidential Elections. They say they controlled for other factors in joblessness — like offshoring of manufacturing and trade deals. Among their results:

  • There is a correlation between automation in the years before the election, and how people voted.
  • At a 10% lower robot exposure, Michigan would have swung in favor of Clinton.
  • If robots barely increased in the years leading up to 2016, Clinton would also have won Pennsylvania and Wisconsin.

The electoral count, the authors say, would have been 278-260 with a Clinton win rather than the ultimate 306-232 in favor of Trump.

  • Frey and his co-authors write, "Although these findings naturally should be interpreted with care, it bolsters the view that that automation in recent years tilted the electorate into opting for radical political change."

What's next: Frey and his team recommend that governments act before political unrest worsens.

  • "Looking forward, automation is likely to become a growing political challenge."
  • "To avoid further populist rebellion and a looming backlash against technology itself, governments must find ways of making the benefits from automation more widely shared."

Go deeper: The anti-robot uprising is coming.