
On taxing the robots
Bill Gates, a pioneer in computer innovation, recently opened the gates to a debate that will become mainstream in the coming decades. He suggested that taxing robots would be a perfectly acceptable policy to defray the losses stemming from automation.
His assumption is based on the correct fear that the advancements of high-tech industry threatens a large number of low to medium skilled jobs, of mostly equivalent income. This would lead to a high rate of unemployment in areas where automation may replace human labour. One wonders which sector is safe from the replacement of its labour force and the substitution of human capabilities. This is accompanied by reassurances of increased labour demand in the “caring sector” or other interpersonal service jobs. There is a growing need for caregivers for older people, people with special needs, helping children in education. It is logical to focus on areas that manifest increases in needs to reorient employment and then rely on innovation and automation to create wealth and increase productivity wherever possible, but there are some underlying assumptions which are left unchallenged, with regards to ease of reemployment, quality of remuneration and so on.
This also implies transfers in order to finance the caregivers and all of the other created jobs, many of which are governmental. Taxation is applied to human workers. Human workers represent labor and they are being replaced by robots. Why not tax robots, too? Human labour is substituted by machines, which do exactly the same job, maybe even better. There are two fine lines to this debate. The first is where the robot begins and ends for it to be taxed, especially since we are discussing automation, of which robotics is just a component. Many clerical jobs have been lost due to Microsoft’s Office suite, and we have not taxed the software in some special way to make up for the loss. The second is how to reconcile taxing some form of capital investment at a higher rate over others.
Using robots instead of humans means fewer wages for the employers to pay. That translates into more profit for the companies, or their owners. Automation increases productivity and enlarges the profits significantly. One possible solution would be to adopt a strategy based on redistribution of added value given by robots, rather than attempting to prevent automation. Workers will need retraining for new careers. One the other hand, imposing taxes on robots could also be designed to slow down the conversion from human workforce to robots, to prevent automation in areas where the marginal gain is still low enough to prevent automation and to establish transfers for the still resulting structural unemployment. However, those in favour of taxing robots are swimming against a strong tide. Taxing them will not eliminate unemployment, or the shift from human labour to (smart) machines. It will just stretch out the transition period and artificially prolong current economic structures, with a significant advantage to economic competitors (for instance, outside of the country) who go full speed ahead on automation. Inevitably, certain jobs are fated to be performed through automation, as the productivity could be increased significantly.
Taxing robots would certainly impede innovation. Economic growth in developed countries has slowed down and productivity has generally followed a similar dynamic. Investment and investors are slowly turning away from those markets. A predictable path to growth in those countries is the high-tech sector and the ongoing digital revolution. In those circumstances, the key is to use robots in a way that complements human workers, instead of replacing them wholesale. The Industrial Revolution was not just about destroying jobs, but about creating new ones and about workers discovering how to increase their personal productivity with the new machines. The same situation may be happening today with what we call “robots”.
Going back to the early Industrial Revolution, economist Gregory Clark from the University of California came to the conclusion that the first 40 years after an arbitrary starting date of 1760 were marked by a 10% loss in incomes. Only in 1830 did the figures turn positive, with a 20% increase. Wealth was rapidly created, but returns to labour did not follow the same rhythm. We are confronted here with the costs of transition, which are inevitable, but which we can confront with policies designed for their amelioration. We could also reject the remorseless logic of replacement and accept the same complementarity that exists in chess, where human-computer teams are superior to humans or computers alone. However, we are talking about individual business decisions from independent actors responding to market incentives, so our blandishments of human-robot coexistence may neither be relevant nor applicable in a practical sense.
In times of high-tech advancement and transition from one field of employment to another, governments could focus on wage subsidies to groups of people with unemployment problems or involved in job transition processes rather than subsidizing or protecting the businesses which are about to be destroyed by automated peers. Subsidies already take different forms that represent an inefficient way of allocating resources. However, this policy, and a political constituency arising from it, may easily result in the same decades long and “sacred” subsidies that the agricultural sector has enjoyed. Beyond the understandable concern for food security, agricultural subsidies have also focused on maintaining the viability of small farms, traditional farming methods and related activities of a cultural nature. We might find that this will become the case with labour-intensive manufacturing, generating its own political lobby and political idea that establishes permanent transfers. There could also be smaller contributions from market forces, as some categories of customers may create market demand for hand crafted goods that are superior in some way for their utility than the mass-produced variety. Whether this can sustain significant employment is doubtful, but, like the similar movement with regards to food, artisanal preparation and so on, it does present opportunities to moderate the alienating effects of technological transitions.
Taxing the technological advancement will not create more jobs or keep the existing ones already. At worst, it’s an imposition that will cost in terms of productivity and wealth. At best, it is a stopgap measure that buys society time to digest the changes without significant upheavals. Automation could also assist governments by enabling the reshoring of lost chains of production, which fled abroad searching for lower labour costs. They would not employ as many people, but they would provide some much-needed jobs, many probably better paid, and would contribute fiscally to the state, making the transition more sustainable. The measures that Bill Gates suggested boil down to falling into a protectionist trap that diminishes technological progress; and that is a paradox coming from someone who made his fortune from automation.