A presentation delivered for the Australia and New Zealand School of Government. A lot of the slides are based on Ryan Abbot and Bret Bogenschneider with a major influence in perspective from Moshe Vardi. Sources compiled through my own searching.
Amazon currently employs 125,000 people, it recently put its 100,000th robot to work.
Should robots be taxed? Joao Guerreiro, Sergio Rebelo and Pedro Teles (April 2018)
We use a model of automation to show that with the current U.S. tax system, a fall in automation costs could lead to a massive rise in income inequality. This inequality can be reduced by raising marginal income tax rates and taxing robots. But this solution involves a substantial efficiency loss for the reduced level of inequality. A Mirrleesian optimal income tax can reduce inequality at a smaller efficiency cost, but is difficult to implement. An alternative approach is to amend the current tax system to include a lump-sum rebate. In our model, with the rebate in place, it is optimal to tax robots only when there is partial automation.
Should Robots Pay Taxes? Tax Policy in the Age of Automation. Ryan Abbott and Bret Bogenschneider (Harvard Law and Policy Review, 2018)
Existing technologies can already automate most work functions, and the cost of these technologies is decreasing at a time when human labor costs are increasing. This, combined with ongoing advances in computing, artificial intelligence, and robotics, has led experts to predict that automation will lead to significant job losses and worsening income inequality. Policy makers are actively debating how to deal with these problems, with most proposals focusing on investing in education to train workers in new job types, or investing in social benefits to distribute the gains of automation. The importance of tax policy has been neglected in this debate, which is unfortunate because such policies are critically important. The tax system incentivizes automation even in cases where it is not otherwise efficient. This is because the vast majority of tax revenues are now derived from labor income, so firms avoid taxes by eliminating employees. Also, when a machine replaces a person, the government loses a substantial amount of tax revenue—potentially hundreds of billions of dollars a year in the aggregate. All of this is the unintended result of a system designed to tax labor rather than capital. Such a system no longer works once the labor is
capital. Robots are not good taxpayers. We argue that existing tax policies must be changed. The system should be at least “neutral” as between robot and human workers, and automation should not be allowed to reduce tax revenue. This could be achieved through some combination of disallowing corporate tax deductions for automated workers, creating an “automation tax” which mirrors existing unemployment schemes, granting offsetting tax preferences for human workers, levying a corporate self-employment tax, and increasing the corporate tax rate.
Robot Tax: https://www.techemergence.com/robot-tax-summary-arguments/. Is it trade or automation that is causing unemployment in advanced economies or developing ones? Multiple angles include:
- Whether a robot tax might even be needed. I.e. how quickly are robots going to replace workers
- How such a robot tax might work.
- What are the arguments for or against it.
When is a Robot a Robot for tax purposes?
Gates simply suggested that “some of it could come on the profits that are generated by the labor-saving efficiency there. Some of it could come directly from some type of robot tax.”
A segment of a proposed motion in the EU parliament which was rejected simply called for looking at, “levying tax on the work performed by a robot or a fee for using and maintaining a robot should be examined in the context of funding the support and retraining of unemployed workers whose jobs have been reduced or eliminated.”
Levying a fee on each robot sounds much easier in theory than it practice. How would we define “robots” for these tax purposes? What separates a tool from a robot or a complex computer program from AI? For example, a backhoe replaces human ditch diggers, but few people consider that a “robot.”
Is a vending machine a human-replacing robot? What about a vending machine that can automatically alert its owner when it is out of a product? What about a vending machine the size of a small store that can automatically stock itself and answer complex questions from customers?
If a country did try to place a fee on actual robots, it is likely it would be difficult for some companies to avoid the tax. Manufacturing robots, self-driving cars, delivery drones, and anything that looks like a classic “robot” or clearly replaces a specific category of job would likely be hit.
But there are a lot of marginally robot-like tools that manufacturers would aggressive lobby policy makers to exclude from the tax. Alternatively, manufacturers could modify the tools just enough to exploit some loophole in the definition of a robot.
Approaches to a Robot Tax
For example South Korea reducing the tax incentives for investing in automation is called a robot tax. It is possible to imagine the “robot tax” taking the form of higher taxes certain types of productively enhancing developments and/or the form of big tax breaks for companies that keep the size of their workforce stable each year.
A “robot tax” could also maybe take the form of a tax that target companies with high profits/revenue but small workforces. Something like a worker to profit ratio could serve as a proxy for deciding what company are making use of a lot of automation.
Why Should We Start to Tax Robots that are taking Human Jobs?
That’s because automation allows firms to avoid wage taxes, which fund social benefit programmes such as Medicare, Medicaid, and Social Security in the US, or National Insurance contributions in the UK. Firms are mostly responsible for just paying wage taxes for human workers to their governments.
In the US at least, there is a further incentive to automate because firms can claim accelerated tax deductions for automation equipment, but not human wages. Wage taxes are generally only deductible as paid. This structure allows firms to generate a significant financial benefit from claiming significant tax deductions sooner for robots.
Automation also results in indirect tax incentives. For instance, human workers are also consumers who are responsible for paying consumption taxes, such as retail sales tax (RST) in the US or value added tax (VAT) in the UK. Employers are generally thought to bear at least some of the cost of these taxes, as they may have to increase salaries in response to higher taxes on workers. Because robot workers are not consumers, they are not subject to these indirect taxes and so firms can avoid any associated burden.
Perhaps most concerning, these policies result in dramatically reduced tax revenue for the government. That’s because most government revenue comes from wage and consumption taxes. Corporate taxation now represents less than 9% of the overall tax base in the US and is likely to significantly decrease following the recently enacted Tax Cuts and Jobs Act of 2017.
When firms replace people with machines (or elect to automate initially), the government loses the ability to tax workers. This is not compensated for in the form of higher taxes on corporate earnings. In the aggregate, this could amount to hundreds of billions of dollars a year in lost tax revenues if robots replace workers to the extent predicted by many experts.
Why Taxing Robots is Not a Good Idea
He argues that today’s robots should be taxed—either their installation, or the profits firms enjoy by saving on the costs of the human labour displaced. The money generated could be used to retrain workers, and perhaps to finance an expansion of health care and education, which provide lots of hard-to-automate jobs in teaching or caring for the old and sick.
A robot is a capital investment, like a blast furnace or a computer. Economists typically advise against taxing such things, which allow an economy to produce more. Taxation that deters investment is thought to make people poorer without raising much money. But Mr Gates seems to suggest that investment in robots is a little like investing in a coal-fired generator: it boosts economic output but also imposes a social cost, what economists call a negative externality. Perhaps rapid automation threatens to dislodge workers from old jobs faster than new sectors can absorb them. That could lead to socially costly long-term unemployment, and potentially to support for destructive government policy. A tax on robots that reduced those costs might well be worth implementing, just as a tax on harmful blast-furnace emissions can discourage pollution and leave society better off.
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. Why refit a warehouse when people queue up to do the work at the minimum wage? Mr Gates’s proposal, by increasing the expense of robots relative to human labour, might further delay an already overdue productivity boom.
But 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. Labour abundance is partly to blame; the owners of factors of production in shorter supply—such as land in Silicon Valley or protected intellectual property—are in a better position to bargain. But machines are no less abundant than people. Factories can churn out even complex contraptions; the cost of producing the second or millionth copy of a piece of software is roughly zero. Every lorry driver needs individual instruction; a capable autonomous-driving system can be duplicated endlessly. Abundant machines will prove no more capable of grabbing a fair share of the gains from growth than abundant humans have.
Waves of automation might necessitate sharing the wealth of superstar firms: through distributed share-ownership when they are public, or by taxing their profits when they are not. Robots are a convenient villain, but Mr Gates might reconsider his target; when firms enjoy unassailable market positions, workers and machines alike lose out.
Robots creating a wages and employment 'death spiral' warns IMF
The worst outcome under the IMF modelling is where robots only replace low-skill workers.
"While skilled labour enjoys continuous large gains, the wage for low-skill labour decreases in the short/medium run under conditions much weaker than in the benchmark model [where robots can do any job]," the IMF found.
"Nor is there any assurance that growth eventually raises the low-skill wage. Quite the contrary: there is a strong presumption the real wage decreases more in the long run than in the short run.
"The magnitude of the worsening in inequality is horrific."
Under the research modelling in this case, the skilled wage increases from between 56 to 157 per cent in the long run, while wages paid to low-skill labour drop between 26 to 56 per cent.
The low-skilled group's share in national income also decreases from roughly a third to as low as 8 per cent.
Even in the scenario where robots only compete for some jobs, and the impacts on wages and growth are reduced, the IMF paper said inequality gets worse.
"Allowing for tasks that complement robots does not help as much as one might think, partly because more and more workers compete for those jobs, driving down the overall wage.
"In addition to the fall in the average wage and the rise in the capital share, unskilled workers suffer large decreases in absolute and relative wages."
Even in areas where robots can't compete, the news isn't great from the IMF team.
"This also does not really help, again because there are only so many of those jobs to go around, and labour chased out of the automatable sector tends to drive down wages."
As for solutions, the IMF broadly targets two possible ways to limit mounting inequality: through education, and tax.
Sadly neither option looks overly promising.
While education can be seen as an investment to convert workers from unskilled to skilled labour, it has its limitations.
"Can it offset the huge real wage cuts unskilled labour suffers and the decrease in labour's overall income share at an acceptable cost? And if the answer is yes, how long will it take for wages to increase for those who remain unskilled?" the economists ask.
As for tax, as governments around the world are already aware, it is not easy to track down and get a fair share of the profits and capital accumulation of big corporations.
- Improved education and access to skills, which may require major changes in the system of education finance and admission
- Reforms of labor market institutions to boost workers’ bargaining power and including a higher minimum wage
- Corporate governance reforms and worker co-determination of the distribution of profits
- Steeply progressive taxation that affects the determination of pay and salaries and the pre-tax distribution of income, particularly at the top end
Katina Michael, When Uber Cars Become Driverless: “They Won’t Need No Driver", IEEE Technology and Society Magazine, https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=7484838
Australian "Robo-Debt" Scandal
Data shows 7,456 debts were reduced to zero and another 12,524 partially reduced between July last year and March. At least 20,000 Centrelink debts were either wiped or reduced in a nine-month period, newly released figures show.
The event is taking place at the Shangri-La Hotel, Sydney. Katina is coming in via Zoom.
Date: Wednesday 5 September
Time: 9:00-10:00 AEST
Site Visit: Google (The Millennial and Technology Story)
Participants: ATO Senior Leadership Program