A Robot Tax Is a Bad Idea
The ideas of France's Benoit Hamon, the surprise front-runner in the presidential primaries of the Socialist Party, are far to the left of U.S. Senator Bernie Sanders and even the U.K. Labour Party's Jeremy Corbyn. This, of course, is the era of the expanding Overton Window, with radical ideas bursting into the mainstream. One Hamon proposal in particular, however, should never make it: A tax on robots, based on the premise that their proliferation is bad for human employment.
Hamon's most discussed proposal is a 750 euro ($805) monthly universal basic income for the French. It doesn't make much sense, primarily because 750 euros is well below the poverty line in France. But it's also expensive, so Hamon proposes new taxes to fund it. The robot tax is one of them.
It's not Hamon's original idea. In May, 2016, Mady Delvaux, a Luxembourg politician, suggested the link between universal income and a tax on automation in a draft report to a European Parliament committee. Delvaux's report recommended making companies report the effects of artificial intelligence and robotics on their economic results "for the purpose of taxation and social security contributions." It added that "in the light of the possible effects on the labor market of robotics and AI a general basic income should be seriously considered."
The logic is clear: Machines are getting smarter and better at tasks that previously could only be performed by humans, so why not tax the owners of smart machines and pass on the money directly to society?
Funding a universal basic income with more taxes on the wealthy is not necessarily a bad idea; a majority of Finns, for example, support it, and Finland is currently testing the idea in a limited experiment. But linking basic income to the destructive influence of automation is not supported by data.
Guido Matias Cortes of Manchester University in the U.K., Nir Jaimovich of the University of Southern California and Henry Siu of the University of British Columbia have long studied the decline of routine jobs, which involve a narrow set of specific activities performed by following well-defined instructions. They can be both manual -- operating a forklift or repairing home appliances -- and cognitive: secretarial work, bookkeeping, serving clients in a bank branch. In 1979, 40.5 percent of the working age population in the U.S. held such jobs; that share stayed constant for another decade, but then, by 2014, it dropped to 31.2 percent.
It's mostly people, both men and women, with a high-school education or less who accounted for the decline, Cortes, Jaimovich and Siu point out in a just-published paper; some of them ended up out of work, others took non-routine manual jobs -- became waiters and other service workers or, say security guards. That kind of employment has gone up.
Conventional wisdom says automation caused the less-educated Americans to lose their routine jobs and forced them either into unemployment or into the service sector. In their latest paper, Cortes, Jaimovich and Siu tested that hypothesis. They built a model with information and communication technology capital as a proxy for automation to gauge how much of the labor market change it could explain. The researchers concluded that the ICT capital would have needed to increase several times as fast as it actually has done if it were to account for both the occupational shift and the unemployment growth.
Automation did play a certain role in determining less-educated workers' life choices. But, Cortes, Jaimovich and Siu wrote, other factors were at least no less important. They specifically named "the share of high-skilled workers and their occupational choice, outsourcing and trade, and changes in policy affecting the incentive to participation in the labor market."
In other words, for workers in routine occupations to keep their jobs, it wouldn't be enough for a government to raise the costs of automation. It would also need to depress the population's general education level (for example, by making education less accessible), engage in protectionist policies and weaken the social safety net. The Trump administration in the U.S. appears to be engaged on all these three fronts.
To a leftist like Hamon, such policies would be anathema. Out of the entire available arsenal, he would only be able to pick increased taxes on automation -- but that wouldn't reverse the trend away from routine jobs. Those holding them now would still, at best, move to the service sector and at worst, end up trying to make ends meet on the insufficient basic income Hamon proposes. At the same time, French companies would have less incentive to innovate than their peers in other countries, making the French economy less competitive.
To make the trend more benign, growth in services and in non-routine manual crafts is the only useful option. People without a college degree will only keep working if they have enough such opportunities that pay better than the welfare state. The world doesn't have to be cruel to them despite the technological revolution -- it just needs to offer them feasible change rather than handouts.
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
To contact the author of this story:
Leonid Bershidsky at email@example.com
To contact the editor responsible for this story:
Therese Raphael at firstname.lastname@example.org