Should government tax robots to save people’s jobs? Bill Gates advocated a robot tax in a recent interview with Quartz. But it’s a terrible idea.
Begin with Gates’s arguments for the tax. He claims or suggests:
- It’s only fair for government to tax a robot at a “similar level” to the human worker it replaces. So, for example, if someone earning $50,000 were taxed $10,000, then the robot that does the same job should be taxed around $10,000.
- Taxes fund things such as education and care for the elderly, and a robot tax would help finance the human jobs in those fields.
- A robot tax would “slow down the speed” of the adoption of labor-saving technology and also help finance “transition programs” to retrain workers and the like.
Gates’s rationale might appear superficially plausible to some, but it falls apart at the slightest breeze of critical analysis.
Even leaving aside for a moment Gates’s presumptions about what government should be doing, his call for a robot tax makes little sense just on accounting grounds. Consider a simplified example in which a business owner hires one person at $50,000, then buys a robot for $100,000 to do the same job. The money for the robot goes to the robot manufacturer as revenues and so will be taxed (after expenses). If the business owner pays $10,000 per year to pay off the loan for the robot, he clears an extra $40,000 per year in profits. That money is taxed, and likely at a higher bracket. Alternately, the business owner might plow the funds back into his business, which would result in other people making more money. So Gates’s robot tax hardly substitutes one tax burden for another.
We should note that robots and other capital goods are already taxed by most states: It’s called the sales or use tax. And various localities tax businesses annually on the value of their property. Apparently Gates is talking about a new federal tax. But we should be talking about how to get rid of existing taxes on capital goods, as such taxes depress investment and expand red tape, not how to expand them.
One effect of Gates’s robot tax would be to create a new government bureaucracy to define what is and is not a “robot” and how much different robots will be taxed. Robots are just machines with computer components that require varying degrees of human maintenance and oversight. As Mark Herschberg suggests, drawing a firm line between machines such as power looms and tractors and “robots” is impossible. It is also impossible to accurately calculate how much human labor a robot offsets. As a consequence, the robot tax would be arbitrarily applied and manipulated by agenda-driven politicians, bureaucrats, and lobbyists.
Another problem, as Herschberg suggests, is that distinguishing “robots” from other sorts of technology is senseless. (Lawrence Summers also makes this point.) For example, are we to tax “robots” (however defined) but not the software that makes them operate better and improves business operations generally? Gates suggests that his tax would apply to automation generally, not just robots, but he’d run into comparable problems however he tried to draw the lines.
We don’t need a more complicated tax system; we need a simpler one. Even if we assume that the government spending Gates calls for is necessary and appropriate and that taxes would otherwise dip with increased automation, it would be less-bad to simply increase the income tax than to impose a bizarre new sort of tax.
Gates’s idea for a robot tax can be criticized on deeper economic grounds. Gates’s basic presumption is that automation expands unemployment. But that’s false, at least in the long run. New technology does not permanently throw people out of work—it frees up people to go into new fields and thereby meet ever more human wants and needs.
Consider agriculture. Over the Twentieth Century, something like two-thirds of agricultural jobs were lost even as output expanded, trends due largely to better technology. Does anyone think it would have been a good idea to discourage development of tractors, irrigation systems, artificial fertilizers, and so on, to “save jobs?” That would have been idiotic.
As some industries shed workers with automation, other industries expand. Gates himself contributed to the expansion of myriad industries with his software that helped drive the personal computer revolution. In 1880, nearly half of everyone working in the United States were in agriculture; in 2016 only about one percent were. And now roughly 80 percent of people work in services. The fact that technology enables us to do less back-breaking labor and expand the number of services that people offer each other is a good thing.
Gates is right to worry about short-term problems involving people who need to find new jobs. For example, nearly two million people drive heavy trucks for a living. As those jobs are taken over by computer-operated vehicles (are those “robots”?), as I figure they will be, truck drivers may need to look for work in other fields. But some of them may have a hard time training for a new position in a different industry. Part of this problem will be solved as fewer people train to drive trucks and some drivers retire, but many drivers may be thrown out of a job. What’s the right way to address such problems?
Gates presumes that it’s government’s proper role to retrain workers. It isn’t. Bureaucrat-run reeducation centers function poorly for the same reasons that bureaucrat-run industries generally do poorly (just ask the Venezuelans today). Individuals have a strong incentive to find new work, companies have a strong incentive to train those people, and private schools and training centers have a strong incentive to educate them. Government can push out some such efforts, but it can’t do a better job providing those values.
Indeed, bad government policies are a major drag on people’s ability to switch careers. High taxes, including payroll taxes, make it harder for people to save for their own rainy days. Badly run government schools often ill-equip people for a dynamic economy. Wage controls make it harder for people to earn their first jobs and enter certain unionized industries. Labor regulations make employers less likely to hire people. Government-controlled health insurance often binds people to their current jobs. Anti-competitive regulations and licensing also make it harder for people to enter certain fields. Subsidies often encourage people not to look for work. Land-use policies discourage people from moving from depressed areas to booming areas.
Bill Gates is wrong that today’s technologies pose a unique challenge to people finding work, and he is wrong that government-controlled programs are the answer to retraining. When it comes to people adapting to a changing economy, government is the problem, not the solution. Politicians and bureaucrats should get out of the way and let businesses and individuals, working by voluntary consent for mutual gain, solve the transitory “problems” of living in an ever-wealthier and more prosperous society.
In short, the answer is to free markets, not more tightly control them; to slash taxes, not expand them.
We’ve gotten so far from capitalism in principle and in practice that even today’s industrial giants such as Bill Gates often see a problem and reflexively reach for expanded government powers to solve it—temporarily forgetting that they often despise the people wielding that power. These days people often fear innovation and seek to throttle it rather than cheer it on and seek to unshackle it. Thankfully, capitalists continue to innovate, automate, and improve people’s ability to produce life-enhancing wealth, despite political hindrances.
Capitalism in this context means simply that government protects people’s rights to produce and to trade and leaves them otherwise free to go about their business. The fact that people’s business increasingly involves robots is not a reason to further undermine capitalism but to more fully embrace it.
Image: AZAdam