Should robots pay income tax?

Economists critique a novel proposal from the founder of Microsoft

In the futuristic television series “Humans”, Joe, a middle-manager in a manufacturing company, is sacked from his job and then replaced with a ‘synth’ – an artificial human who can do his job better at a fraction of the cost. If the owners of the synth had to pay a tax on the income the synth generated, would that have helped keep Joe’s job? Or at least could the money from such a tax have been used to help Joe find a better job or pay for other public services? In essence, that is the question raised by Bill Gates’s suggestion, made in an interview earlier this month with Quartz, an online publication, that there should be a tax on the income generated by robots.

This scenario is no longer considered science fiction. Earlier this year Ryan Avent, of The Economist, published a book called The Wealth of Humans predicting enormous social upheaval as developments in robotics and artificial intelligence render millions of workers obsolete. Persuaded by such arguments, Gates, like many CEOs and politicians, backs the creation of a universal basic income (UBI), to provide a safety net for citizens whether rendered jobless or not by the hypothesised technological revolution. He has mooted a tax on robots as one way of financing the UBI.

Gates’ central idea is that human labour is taxed (through income taxes, social security taxes and so on), but the income that robots or machines generate for their owners is taxed only indirectly in the form of profits taxes. Taxing it more directly could therefore be fair and efficient:

“Right now, the human worker who does, say, $50,000 worth of work in a factory, that income is taxed and you get income tax, social security tax, all those things. If a robot comes in to do the same thing, you’d think that we’d tax the robot at a similar level.”

Furthermore, he argues that a tax on robots would at least slow (if not halt) the investment in such labour-replacing devices, thereby giving time for alternative employment to emerge as the economy adjusts. The tax revenue generated could then be used to finance training schemes to help workers regain employment or provide a basic income to prevent poverty.

The idea is gaining momentum. The French socialist presidential candidate Benoit Hamon supports it and earlier this year, the European Parliament discussed the proposal, but ultimately rejected it. However they accepted the need for a new ethical and legal framework for the use of robots. But Gates’s proposal has, so far, been given short shrift by most economists and commentators.

Izabella Kaminska, a journalist on FT Alphaville, is not impressed. She thinks Gates’s robot tax is just another call for more taxes on corporates or wealth. And she argues that taxes on investment are a bad idea because they will deter productive investments which can deliver higher output and living standards. But in any case, she’s sceptical that we will face the scenario envisaged by Avent, Gates and others:

“To be blunt, robots and AI systems will only displace human workers in any great number in a finite resource world if the overall unit costs of building, educating, maintaining, powering and resting them is lower than the average cost of birthing, nurturing, educating, feeding and caring for human workers especially relative to their overall productivity potential.”

She also thinks that human beings have some very large advantages over any potential robot including our capacity for creative thought and ability to self-power and maintain ourselves. Her conclusion is that, considered in the round, the business case for investment in robots is weaker than it first seems. Even if it is, a tax on robots is no more persuasive than other taxes on capital or profits.

Over at The Economist Ryan Avent and his colleagues aren’t impressed with Gates’s idea either. They agree with Kaminska that automation is a form of capital investment that should increase productivity, raising incomes and lowering prices, which makes deterring it through a tax is probably harmful to the supposed beneficiaries of the policy – the working poor.

“Investments in robots can make human workers more productive rather than expendable; taxing them could leave the employees affected worse off. Particular workers may suffer by being displaced by robots, but workers as a whole might be better off because prices fall. Slowing the deployment of robots in health care and herding humans into such jobs might look like a useful way to maintain social stability. But if it means that health-care costs grow rapidly, gobbling up the gains in workers’ incomes, then the victory is Pyrrhic.”

They add that the slowdown in productivity growth experienced by most countries over the past few years is evidence, if anything that investment in new technology is too low, rather than too high. Returns to capital are falling, but profits are rising, probably because of the dominant market position of some leading firms in what has become a “winner takes all” economy. A tax on those firms or measures to disrupt their market power might be better policies than a tax on robots, they argue.

Yanis Varoufakis, a Greek economist and former Finance Minister, who agrees with Gates about the potential plight facing legions of displaced workers, points out some thorny practical problems with the idea. First, how are we to measure the income of a robot? If we use the income generated by the human whose job has been replaced as a reference point, then how do we adjust for the expected future changes in income? And what if the robot does a job which no human has ever done? What is the income then? But perhaps most tricky of all: how are we to distinguish between robots and other types of machinery. After all, why tax a robot but not a combine harvester? Can ‘autonomy’ really be a sound basis for that sort of distinction?
Instead he argues there is a simpler policy that could have similar benefits:

“There is an alternative to a robot tax that is easy to implement and simple to justify: a universal basic dividend (UBD), financed from the returns on all capital. Imagine that a fixed portion of new equity issues (IPOs) goes into a public trust that, in turn, generates an income stream from which a UBD is paid. Effectively, society becomes a shareholder in every corporation, and the dividends are distributed evenly to all citizens. To the extent that automation improves productivity and corporate profitability, the whole of society would begin to share the benefits.”

This would mean that future companies would be part-owned by the state, and a share of profits would be disbursed to everyone in equal amounts, but companies would retain the freedom to automate or use robots as they see fit. That, of course, would mean the part-nationalisation of all publicly-listed firms, a move many would oppose and which goes far broader than a measure to target job-destroying automation.

There is plenty of reason for scepticism both about whether intelligent robots will really cause mass unemployment (or the also widespread lesser prediction, increased inequality of incomes) and the practicality of targeting a tax on specific forms of automation. Robots have been displacing humans in car manufacturing for decades already. Nevertheless, amid such a technological transformation, creative thinking about how and on what to levy taxes is much needed. So Bill Gates has done a service by kicking off the debate.

Edited by Bill Emmott