Home » Why Taxing Robots Is Against Human Labor

Why Taxing Robots Is Against Human Labor


The robots are coming. And that’s a good thing”. This is the title of a very recent MIT Technology Review article written by Daniela Rus, director of MIT (Massachusetts Institute of Technology)’s Computer Science and Artificial Intelligence Lab, and the novelist Gregory Mone. The key idea of this article, inspired by Daniela Rus’ book “The Heart and the Chip: Our Bright Future with Robots”, is that machines will augment human capabilities.

The presence of machines (e.g., robots, algorithms, intelligent technologies) in organizations has become massive. An extreme example is Motional, the Singapore-based driverless taxi service that started substituting traditional taxis since 2020. Another, even more radical, example is given by the international rum producer firm Dictador, which appointed on September 1st 2022 the first AI human-like woman robot CEO named Mika.

Although this technological revolution cannot be stopped, myopic populist governments are proposing policies to tax the use of machines, making them costlier for organizations, so as to force them to “use” relatively more human labor in their production processes. Some of these tax policies are truly imaginative, like those proposing to tax the virtual compensation that robots would have as a replacement for human beings. Besides fooling workers and voters, these policies ignore – naively or for political gains – the nature and the role of machines in organizations nowadays, and more in general what is the future of work.


To understand what machines do in organizations, let’s start with an ultra-simplified description of how organizations produce their goods without the help of machines. The first definition we need is the one of production function. In simple words, the quantity of goods produced by an organization is the result of the interaction of the traditional factors of production (e.g., human labor; physical capital as equipment). What happens when an organization adopts a machine? The machine can be considered as another factor of production that interacts with the traditional factors of production and positively affects their productivity.

After the adoption of a machine, an organization bears short-run additional costs not only due to the financial resources required to acquire the machine, but also in terms of adjustments to be made to the organization. For instance, integrating the machine into the production processes is costly, as well as training workers on how to use and supervise the machine. However, when the adjustment process is completed, the machine can either increase the marginal productivity of all the traditional factors of production or benefit one factor of production more than the others. Since full automation of jobs is extremely rare in organizations, and in the vast majority of cases machines and humans interact – with humans and machines complementing each other in executing the tasks of which the jobs consist – it is expected that the adoption of machines will be “human labor-augmenting”. Put simply, machines will make humans more productive, either allowing humans to specialize on the tasks where they are more productive and useful for the organization, or reducing their workload and creating opportunities for re-skilling and up-skilling.


What happens when a tax on the use of machines (typically labelled as robot tax in the news media) is imposed to organizations? At first blush, one may think that the employment of human labor becomes relatively cheaper than the use of machines, thus incentivizing organizations to hire more workers. For the government, more human employment translates into more tax revenues, also considering that in many Western economies, taxation on labor income accounts for a great share of total tax revenues (e.g., in the European Union such portion is well above 50%). However, despite potentially addressing a public finance issue in the short term, a robot tax creates a lot of problems to the whole economy in the long run.

First, the use of machines enables organizations to develop innovations in the form of new products and services. An example is given by humanoid robots assisting and emotionally supporting older people and remotely connecting them to doctors and nurses. On the one hand, innovation makes organizations competitive and less likely to exit markets and lay off workers. On the other hand, innovation needs complementary assets to be commercialized; thus, higher innovation rates create new demand for other organizations and new jobs.

Second, the adoption of machines is more diffused among larger organizations that typically operate in more than one country. Multinational organizations are mobile tax objects; thus, in the absence of an utopian – but strongly desired by politicians – coordination of robot tax policies across countries, multinationals can easily move their activities to another country, with negative backlashes on the domestic employment levels.

Third, the adoption of machines is not a decision that can be easily overturned. The adoption of machines entails fixed costs and other investments to pre-adapt the organization. Thus, in the case of a robot tax shock which negatively affects corporate profitability, the organization may be forced to lay off – at least temporarily – some workers, who would otherwise have joined re-skilling and up-skilling programs.

These are just three of the many reasons against the implementation of a robot tax. Moreover, it is worth stressing that in the last years, machines dramatically improved the work conditions of humans, for instance in terms of safety.


Hence, it is clear that the policy debate must not focus on how to discourage organizations to use machines, but on the implications of market adjustments to the wages of human workers with the advent of machines. According to Daron Acemoglu and Pascual Restrepo, among others, machine-led productivity growth won’t be passed onto the workers in the form of higher wages, but rather appropriated by the organization’s shareholders. Another scenario is that such productivity growth translates into fewer hours worked, with not necessarily negative effects on wages.

Taking the corporate perspective, many organizations are using machines to measure workers’ productivity, and subsequently determine their wage. On the one hand, this may increase incentives for workers to provide a higher effort in executing their tasks. On the other hand, excessive monitoring may lead to lower trust in the organization, with negative effects on workers’ wellbeing and excessive churn.

In sum, it is still not clear whether policies for the future should be leaning towards a “zero robot tax” – as suggested by the economists Joao Guerreiro and Sergio Rebelo, or towards an unemployment benefit generated by the machine-led productivity gains. What is clear is that governments must not face “the elephant in the room” – that is, the advent of machines – by implementing short-sighted policies to simply stabilize jobs to get political consensus, with the risk to distort market mechanisms and global value chains, thus forcing companies move to other locations. In Aldous Huxley’s dystopian “Brave New World” we saw what happens when stability is placed above all else, and change is deemed as a threat to stability. And we must avoid it.

Photo by Lenny Kuhne

You may also like

Leave a Comment