What’s at stake: “More human than human”, was the motto guiding the Tyrell Corporation’s engineering of biorobotic androids, in 1982’s Blade Runner. Fast forward to 2016, and Bill Gates argues that if robots perform human work, they should be taxed like humans. We review what economists think about this idea.
In a recent interview, Bill Gates discussed the option of a tax on robots. He argued that if today human workers’ income is taxed, and then a robot comes in to do the same thing, it seems logical to think that we would tax the robot at a similar level. While the form of such taxation is not entirely clear, Gates suggested that some of it could come from the profits that are generated by the labor-saving efficiency there, and some could come directly in some type of a robot tax.
The idea of regulating robotics also has appeal on the other side of the Atlantic, where the European Parliament (EP) has been discussing these issues over the past months. Earlier in February, the EP called for EU-wide legislation to regulate the rise of robots, including an ethical framework for their development and deployment, and the establishment of liability for the actions of robots including self-driving cars. But the EP rejected a proposal to impose a robot tax on owners, to fund support for retraining of workers put out of a job by robots. According to the Reuters report, the decision to reject the robot tax was hailed by the robotics industry, which says it would stunt innovation.
Noah Smith says the main argument against taxing the robots is that it might impede innovation. Stagnating productivity in rich countries, combined with falling business investment, suggests that adoption of new technology is currently too slow rather than too fast and taxing new technology could make that slowdown worse. Smiths argues that the problem with Gates’ basic proposal is that it is very hard to tell the difference between new technology that complements humans and new technology that replaces them, and proposes alternative options to deal with the disruption.
One idea is a wage subsidy for low-income workers, which would have the effect of making human workers cheaper. The easiest way to do that is to cut payroll taxes, which disproportionately fall on low earners. Another idea is to simply redistribute capital income more broadly. Income from capital gains, land rents and dividends now is highly concentrated among the wealthy, but policy could change that. Overall, instead of slowing innovation, the government should think about taxing humans less and redistributing the income of robots more.
Lawrence Summers thinks that robots are wealth creators, taxing them is illogical and Gates’ robot tax would amount to “protectionism against progress”. First, Summers disagrees with the singling out of robots as job destroyers, when there are many different kinds of innovation that allow the production of more or better output with less labour input.
Second, he argues that much innovative activity, even of a robot-like variety, involves producing better goods and services rather than simply extracting more output from the same input, and because of emulation and competition, innovators capture only a small part of the benefit. It follows that there is as much a case for subsidising as taxing types of capital that embody innovation. Third, a sufficiently high tax on robots would prevent them from being produced but Summers thinks it would be better for society to instead enjoy the extra output and establish suitable taxes and transfers to protect displaced workers.
Tyler Cowen wonders whether we should be taxing robots or rather subsidize wage labour. He argues that one reason not to tax the robots is that employers might substitute away from robots and toward natural resources rather than toward domestic human labor: for example, we could think of paying the energy costs to outsource to another nation and transport the outputs back home. A second issue is that of incidence: a general problem with a wage subsidy is that sometimes much of its value its captured by employers.
Coming to the incidence of a tax on robots, if the elasticity of the demand for robots is high, there will be a big shift away from robots and toward labor. It is at least possible that workers capture more of the gains this way than from the direct subsidy to their wages. On the downside, the employer fares less well under this scheme. So ultimately it depends on how labor and robot elasticities relate to each other: the robot tax would seem to do best when the elasticity of demand for robots is high, but the corresponding elasticity of demand for labor is low. As robots and labor become more substitutable, that difference in demand elasticities is likely to diminish.
Izabella Kaminska on FT-Alphaville argues that a call for robot income tax is really just a call for more corporation tax and/or a wealth tax, and it seems strange then for Gates to forget the argument against higher corporate taxes, just because the nature of the capital investment is now anthropomorphised. Kaminska highlights two paradoxes connected with Gates’ reasoning – a capitalist and a Marxist paradox – and concludes that “a simple moratorium on robot development would spare us all the angst”.
The Economist Free Exchange blog thinks that Bill Gates is “an unlikely Luddite, however much Microsoft may have provoked people to take a hammer to their computers”. His tax on robots is is an intriguing if impracticable idea, which however reveals a lot about the challenge of automation. Gates worries about a looming era of automation in which machines take over driving or managing warehouses.
Yet in an economy already awash with abundant, cheap labour, it may be that firms face too little pressure to invest in labour-saving technologies. When faster automation does arrive, robots might not be the right tax target: as machines displace humans in production, their incomes will face the same pressures that afflict humans. The share of total income paid in wages – the “labour share” – has been falling for decades: abundant machines will prove no more capable of grabbing a fair share of the gains from growth than abundant humans have.
Yanis Varoufakis also picked up the topic in Project Syndicate. He argues that the only way to simulate an income tax on robots is to use the corresponding workers’ last annual income as a reference salary, and extract the equivalent income tax and social security charges. This presents a number of problems, both in the need of a reference salary – which will be entail a degree of arbitrariness that will make it subject to disputes between producers and the tax authorities – and in the philosophical incongruence in taxing the robots but not the mechanical tools that the robot itself operates. The alternative is Gates’ idea to tax the installation of robots, but a lump-sum tax on robots would merely lead robot producers to bundle artificial intelligence within other machinery, so either the robot sales tax should be dropped or it should be generalized into a capital goods sales tax (which would generate an uproar).
Varoufakis argues instead for a universal basic dividend (UBD which he previously discussed), financed from the returns on all capital. A fixed portion of new equity issues (IPOs) would go into a public trust that, in turn, generates an income stream from which a UBD is paid. Effectively, society would become 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.
Republishing and referencing
Bruegel considers itself a public good and takes no institutional standpoint. Anyone is free to republish and/or quote this post without prior consent. Please provide a full reference, clearly stating Bruegel and the relevant author as the source, and include a prominent hyperlink to the original post.