Taxing robots? Maybe in alternative ways

By Kim Hyeong-woo

In a recent interview with Quartz, Bill Gates, Chairman of Microsoft, proposed that governments should tax robots. By doing so, we will be able to generate social funds to help workers who are displaced by robots. Such a "robot tax" would slow down the speed of automation, thus give more time for society to handle its potential challenges.

He may sound like a modern-day version of a Luddite, a group of English textile workers who destroyed weaving machines in the 19th century, but I believe he is starting an important and timely debate. Advances in robotics and artificial intelligence are likely to result in job losses not only in manufacturing but also in transportation (for example, driverless trucks) and service industries (the fintech revolution). When new technology replaces human workers, it may create the following problems.

First, job losses may result in a substantial decrease in government tax revenues. Based on a Bureau of Economic Analysis (BEA) report, personal income taxes contributed 51.4 percent ($1,946.9 billion) of the total government tax receipts in the third quarter of 2016. This number corresponds to 31.4 percent of the government current expenditures and 70.1 percent of government social benefits that include social security payments, Medicare, and unemployment insurance payments. That is, decreases in income taxes may put the existing social safety net at risk, pushing displaced workers into even harsher lives.

Second, automation may lead to further income inequality across households. When human jobs are replaced by robots, all the labor income (minus operational costs of the robot) will be transformed into profits, which will be allocated to the company's shareholders via dividends, or greater compensation will be given to high-skilled workers who survived automation. Consequently, this is likely to widen the income gap across the human workers in the absence of a tax code overhaul.

Notwithstanding many other socio-economic issues that may stem from automation, these two potential challenges alone may be enough to trigger discussions on a robot tax. If taxing robots or the owners of robots is inevitable, how should we do it? Is it practically feasible and desirable? If not, are there any alternative methods?

It seems that there are many practical issues in implementing robot taxes. As Yanis Varoufakis points out, it is not easy to assess proper tax rates. Bill Gates suggested that if a robot replaces a human worker, we should tax the robot at a similar level as a human would pay. Even in such cases, it is not clear what reference salary should be used over time for the tax base, because machines will not ask for a salary-negotiation every year. Further, there may not be any prior reference salary if the government tries to tax a new company that uses no human workers.

Alternatively, the government may impose a lump-sum tax on the installation of a robot, which will deter automation by creating tax disincentives. However, it would be very difficult to distinguish such taxes from conventional taxes on capital goods, which complement human workers instead of replacing them. Furthermore, heavily taxing corporate investment might impede innovation and R&D activities. That is, distortive effects of such tax schemes may outweigh the potential benefits of slowing down the social costs of automation.

So, it seems that there are many practical obstacles in implementing a robot tax. Alternatively, governments may raise tax rates on corporate profits or they may employ more progressive tax schemes that target the top income earners to compensate for losses in tax revenue due to automation. If properly implemented, such a tax overhaul would provide a reasonable solution to the two problems we mentioned earlier. That is, the government will be able to compensate for losses in income taxes, and the new tax policy would alleviate income inequality. However, I recognize how difficult it has been to put forward a tax overhaul toward a more progressive system even during times with no contemplation of economic consequences of automation.

Fortunately, we are not out of luck. We may consider a practically feasible and less distortive solution such as the establishment of a sovereign-wealth fund like the one proposed by Miles Kimball. Under these schemes, we may create a public fund that has the same level of independence as the Federal Reserve System or the European Central Bank. The fund buys equities of existing companies as well as initial public offerings (IPOs) for the fund using available tax revenues. The fund generates a cash flow, from which dividends can be distributed evenly to every citizen. This is consistent with the idea of a universal basic dividend (UBD) suggested by Yanis Varoufakis as an alternative to a universal basic income (UBI).

I do agree with Gates that we need to start discussions about the consequences of automation and desirable solutions to deal with potential challenges from it. It seems that there are many practical obstacles in implementing a robot tax, even though we must find the new source of incomes that complement tax losses due to automation. However, creating a public fund that renders partial ownership to society might serve an attractive and practically feasible solution to robot taxes.

Dr. Kim Hyeong-woo is a professor of economics at Auburn University. He received his Ph.D. from the Ohio State University, and his B.A. and M.A. from Seoul National University. He published over 25 SSCI journal articles since 2009 in the areas of macroeconomics, financial economics, and economic forecasting. He can be reached at gmmkim@gmail.com.