The use of industrial robots has been increasing over the last three decades and whenever companies replace people with machines, the government loses the ability to tax workers, resulting in millions of euros worth of lost tax revenues each year. In this context, the role of legislators and politicians becomes central to preventing an unwanted situation, where technological revolution comes at the cost of mass unemployment and growing inequality, which is not compatible with Sustainable Development Goal 8, i.e. promote inclusive and sustainable economic growth, employment and decent work for all.
Several experts are concerned with the threat that advances in robotics and Artificial Intelligence (AI) could lead to substantial job losses or job polarization, ultimately leading to increased income and wealth inequality (Méda, 2016; Korinek and Stiglitz, 2017). Among these, Carl Frey and Michael Osborne (2017) predict that, in the next 20 years, technology may displace a large part of human workers. The World Economic Forum warned in a 2018 report [1] that more than 50% of current jobs will be automated by 2025. Although some studies raise doubts about the loss of technology jobs in advanced economies, there is consensus about its impact on developing economies, which are more dependent on factories and encounter robotic displacement (Carbonero et al., 2018; De Backer et al., 2018). Despite the benefits of technological development, unregulated technology can cause “serious social, economic and political harms” (Merola, 2022).
This article focuses on the role of taxation, as a possible tool, for sharing the gains from automation and robotics. In addition, since robotics seems to jeopardize routine and low-skilled workers, governments need growing public resources to be invested in education and training.
The expression “robot tax” refers to taxing the existence of robots or the operation of the robot’s labour in a company’s production and logistics (Bendel, 2019). The idea of a robot tax is yet to be fully defined, mainly because the robot tax suggests that companies should have to pay a tax in one of two ways:
- a company could pay income tax on each robot based on the displaced human employee’s salary; however, humans and robots often collaborate on tasks so it would be difficult to measure the work distribution between robots and workers for tax purposes (Merola, 2022); or
- a company could pay higher rates of corporation tax for using robots in their workforce out of their profits, which are likely to have increased due to the efficiency of the robotic workforce; though this could also have a negative financial impact, especially on small businesses (Merola, 2022).
South Korea is the only country to have a kind of robot tax, however, calls for a type of this tax are starting to emerge in the UK, USA, Japan and Canada. During 2017, South Korea introduced a robot tax. Korea has been very quick to adopt robots in the workplace, especially in manufacturing, which is dominated by semiconductors produced by robots. Another factor that is hurrying Korea along is that its unemployment rate has hit 1.7 million unemployed. However, the tax itself is not really a tax on robots – but rather discourages capital investment in technology.
Also, in 2017, the European Parliament called for automation legislation that would include a framework for their development and implementation, as well as assign responsibility for their acts. Most European leaders agreed that the growth of automation must be controlled, although they rejected the idea of introducing a robot tax to offset some of the potential negative effects on the job market. Six years later, the EU Artificial Intelligence Act will introduce regulation aimed at ensuring the human-centered and ethical development of AI in Europe – the measure still does not include any tax.
New York City Mayor – Bill de Blasio (2014-2021) – pushed for an automation policy aimed at protecting millions of jobs that could end by 2030. Revenue from the robot tax would be used to create new jobs in green energy, health care and education, helping to prevent the tax gaps.
Robots and automation put jobs at risk and therefore could cause a loss of income tax, but as the use of robots likely increases productivity and thus profits, one way to recoup the money loss is to tax these robots. Robot tax can also combat multinationals’ tax evasion, as they often divert tax profits to tax havens. Robot tax – as salaries – is calculated based on taxable income in the tax jurisdiction where the robot is located, thus avoiding shifting tax outside the jurisdiction.
Nevertheless, we must be careful with the definition of “robot” since its very broad and difficult to determine when a job is really replaced by a robot. As countries implement automation at different times, the level of legislation and definitions to be adopted should be considered. Except if all countries implement a robot tax with the same definition of a “robot”, companies can move their operations from countries that use a robot tax to countries that do not impose a robot tax, or to countries that have a definition that suits their purposes.
Alternatively, instead of taxing robots, governments can increase corporate revenues by raising corporate and capital gains taxes or requiring higher VAT rates on purchases of robot technology. Other possibilities – perhaps more controversial – are through governments to start and develop an investment fund or buy shares in technology companies to redistribute them to society after receiving dividends.
The robot tax can be seen as a tool to manage labour migration so that the workforce adapts at a rate that allows the development of alternative sources of work. A robot tax could reduce the unemployment rate caused by the destruction of jobs lost to automation. Because it takes time to acquire the necessary skills from one sector or job type to another – as do government schemes needed to provide adequate training – slowing the rollout of robots could give those workers time to train elsewhere. Alternatively, governments could consider reducing the tax burden on companies investing in people.
Robots can perform repetitive tasks faster and faster, resulting in less waste and cheaper products thus higher profits. Taxing robots can be seen as an opportunity to get something useful for society through taxation. Ryan Avant has argued that it does not make sense to tax certain types of capital that increase productivity and that a general wealth tax or property tax would be a better option.
Low wages could be prolonged by making it more costly to automate manual and repetitive tasks, which inhibits the ability to move from “bad jobs” to “good jobs”. A definition of “good work” is work whose quality supports people’s health and well-being, benefits them and serves the public interest.
If companies are less willing to invest in new technology because they fear an expensive robot tax, tech companies are unwilling or unable to develop it. Growth in rich countries has stalled for the last 100 years, indicating that it is increasingly difficult to find new ways of doing things. This, combined with reduced business investment, suggests that some robotics taxes should be set at levels that do not discourage business investment.
A robot tax can prevent the large-scale displacement of low- and medium-skilled workers and mitigate the negative effects of automation on the working class and the potential increase in wealth inequality, but the challenges of defining robots and encouraging innovation are difficult to overcome. Nonetheless, the resulting revenue can contribute to economic growth and the expansion of national social security budgets, making it worthwhile to delay the implementation of disruptive technologies which could be a way to combat inequality.
Soon, policymakers, companies and workers will need to prioritize this challenge.
[1] World Economic Forum, 2018, “The Future of Jobs Report 2018”.
Nuno Jacinto
November 2023