On Taxing Robots

A few weeks ago, Bill Gates suggested robots that take human jobs should be taxed. His reasoning is the government taxes humans who do a particular job, so why shouldn’t you also tax the robot that replaces the human in that job? Enacting such a tax would not only ensure the government retains its tax base, but also slow the adoption of robotics, allowing people time to adapt to the changing world.

Of course, many people have disagreed with Gates’ idea, going so far as to call the founder of one of the most influential tech companies in the world a Luddite. But, I think it’s an interesting idea, worthy of discussion

The problem

Bill Gates is trying to solve one of the biggest problems of the twenty-first century, the displacement of human labor by automation. The problem is that automation results in gross inequality in society.  The manufacturers—the wealthy who own the businesses who replace human labor with robots—will benefit from the robotic revolution. They will be able to produce their goods at a much lower price, leading to increased profits.

The workers will benefit from lower prices on goods, but only if they actually have money to buy the goods. If they don’t have a job and are unable to find a job because their skills are obsolete in a robot economy, then this automation revolution is terrible for them. What’s more, there are likely to be a lot of displaced workers and relatively few robot-owners, so you end up with a society in which there are a bunch of poor people and a few really wealthy people.

This result isn’t good for anyone, since, once inequality becomes great enough and the poor realize the system is fixed against them, the poor are likely to rise up and destroy the system. I think we’re seeing some of that anger over inequality with the rise of Trump, and, if he doesn’t address the issues, then things could easily escalate in unpredictable ways.

Superficially, taxing robots would solve many of these problems. Replacing humans with robots would be less compelling, which would slow the drive to automation. People would still be displaced, but fewer people would be displaced each year, allowing people more time to adjust to these changes. What’s more, the government wouldn’t lose the tax revenue from the displaced workers, since they’d get it from robots instead.

The problems

Larry Summers has criticized the proposal, asking why Gates would tax in ways that reduce the size of the pie rather than distributing it differently. In a way, the “reduce the size of the pie” argument is silly, since all taxation does this. Too often, people who make this argument will only complain about the pie shrinking rather than actually proposing ways to redistribute it. From my perspective, if they aren’t willing to suggest reasonable ways to redistribute the pie, their “reduce the size of the pie” isn’t actually a serious argument, but rather an attempt to avoid arguments about pie redistribution entirely.

To me, the biggest problem with Gates’ proposal is that it’s impractical. Which labor-saving devices are robots? Is a stapler? What about a photocopier that staples? What about a photocopier that adjusts toner usage to increase contrast to improve the legibility of my document? What about a computer that allows me to copy a large chunk of text from one area to another? Or one that fixes typos when I do that cut and paste? What about one that paraphrases the text when I cut and paste? What about one that writes the text for me?

Thus, figuring out what qualifies as a robot is problematic. I imagine that if such a law were passed, an industry would arise that promised to de-roboticize your robots from a tax perspective (at least that would create jobs for humans). Depending on how tax law was worded, we could end up with people who look over the results of robots simply to rebrand robotic production into human production.

What’s more, if you add a robot tax, it could drive away business. As production has shifted internationally from countries with expensive workers to low cost workers, I imagine production by robots would shift from countries with high-cost, high-tax robots to low-cost robots. Thus, you’d lose the jobs of not just the factory-line workers, but also the people watch over and maintain the robots.

So, though Gates’ proposal has intellectual appeal, I don’t think it is feasible.

So what do you do?

Perhaps taxation of companies is one (relatively-unsatisfying) answer. When companies lower their costs—such as through automation—it increases their profits, and the tax system already has a way to ensure profits are taxed. Perhaps this tax could be increased on robot-producing and robot-maintaining companies. Of course, this would also result in such companies relocating to more tax-friendly locales, but that’s a general problem with taxation. To take care of such problems, the tax system is needs to be revamped.

Taxing the robot-providers and robot users still doesn’t solve the problem of displaced workers. This is where something like a guaranteed minimum income would potentially fit in. If a displaced worker doesn’t actually have to worry about starving to death or dying because they can’t afford hospital bills, then losing their job to automation isn’t nearly as big a problem as it would be otherwise. A guaranteed income would ensure workers have time and resources to retrain. There will still be some people left out in the cold by automation, but the consequences of being made obsolete would be far less severe.