Bill Gates has called for a “robot tax” to help maintain government tax revenue when human laborers are let go because a “robot comes in to do the same thing.” He seems to be following the same line of argument that many make for a universal basic income: when a worker in a $50,000/year job is replaced by some automated process and then finds a $25,000/year job, the government will need the extra resources to sustain this worker’s quality of life, but will have lost that much in income taxes by the worker moving to a new lower-paying job.
Even setting aside the questionable benefit of government taxing and spending, there are some gaping holes in Gates’s argument.
We Must First Define What a Robot Is
The first and most obvious issue involves defining a “robot.” Must it be some human-shaped machine that beeps and boops and speaks in monotone? What about the giant mechanical arms that put doors on new cars in the factory? Does a “smart” tool that saves labor but doesn’t have any moving parts count? In that case all computers would be subject to a new tax. What about technologies that don’t necessarily involve computing or moving parts, but still reduce the amount of labor needed in some production process? In that case, every capital good going all the way back to the invention of the wheel and the control of fire would be taxed.
There is no non-arbitrary way to draw the line between “good” capital goods and “bad” capital goods. Capital is capital no matter how complex it is or whether it has buttons and blinking lights. All capital is employed by entrepreneurs for one reason: to produce more of what consumers desire, and to do so by using resources that have no other more valuable or more productive use.
Robots Increase Tax Revenue
Second, working through Gates’s thought experiment reveals that there is actually no lost taxable income on the whole. We can even make this scenario very complex, if we like. For instance, let’s assume a robot — let’s call it Bernard — just appeared like manna from heaven for an entrepreneur named Robert. The robot Bernard is exactly as capable as the human Theresa in Robert’s business and Bernard never breaks down. Suppose then that Theresa is totally displaced by Bernard and there is just no other job for her anywhere — she goes from earning $50,000/year (also her marginal revenue productivity for Robert) to $0.
Thus, in this scenario, $50,000/year of taxable income has been exactly transferred from Theresa to Robert because of Robert’s new robot, Bernard.There is no change in total tax revenue for the government.
Any relaxation of these assumptions to make the scenario more realistic would also only add to the amount of taxable income. Bernard is costly to make and repair, which means some robot producers and repairers have some income. Theresa could find another job elsewhere. If all of this happens in a progressive taxation system, Robert’s new income could even push him into a new tax bracket with a greater share of his total income taxed. If the Bernard robot is even more productive than Theresa, all the better for the government’s tax revenue, as all of the new income is simply added to Robert’s taxable income, at progressively higher rates.
Of course I’m not arguing that increased tax revenue is a good thing, just that Gates’s fears of lost tax revenue because of technologically displaced workers is unfounded.
Technology Makes Whole Populations Better Off
Finally, every historical episode of new significant technologies making production more efficient has resulted in whole populations becoming better off. While it is true that new technology can temporarily disrupt who is working where, the overall effect of increasing productivity is a positive one, even for those who lose their jobs. See the incredible “hockey stick of human prosperity,” in which the dramatic increases in human standards of living began with the Industrial Revolution and only accelerated with the introduction of new technologies like computing and the internet.
There is no war between labor and capital, or humans and robots, because all capital acts to increase the marginal productivity of labor. This has two effects: (1) we can produce more of the goods we want and need, and (2) our wages increase, both in real and nominal terms.
A farmer with a tractor is much more productive than a farmer without. A secretary with a computer is much more productive than a team of secretaries without computers. Can you imagine what your productivity would be without any capital? What sort of wage or salary could you earn without all of the fantastic capital goods at your disposal? We would return to a primitive existence, living hand-to-mouth without the ability to use capital goods. We would have all of the grueling and painful quality of life that comes with that.
Discouraging the use of capital is quite backward-thinking, therefore, despite the way Bill Gates positions his message as a forward-thinking one. It asks for more primitivism and less advancement. It stalls progress and prevents us from solving the problems that remain in the world, many of which Bill Gates has personally dedicated his life and wealth to solving.1
- 1And speaking of Gates’s personal accomplishments, it is especially frustrating to hear this kind of talk from a man who has saved more labor across the globe than any other person alive, possibly even among all of human history. But that would be impossible to tell, as all new productive technologies necessarily depend on the efforts of previous inventors and individuals who accumulated capital. We wouldn’t have the personal computer without having first harnessed electricity, plastics, metals, all kinds of engines, agricultural developments, and so on, as the quote from Mises makes clear:
If one were to measure the length of the period of production spent in the fabrication of the various goods available now, one would have to trace back their history to the point at which the first expenditure of original factors of production took place. One would have to establish when natural resources and labor were first employed for processes which — besides contributing to the production of other goods — also contributed ultimately to the production of the good in question. The solution of this problem would require the solubility of the problem of physical imputation. It would be necessary to establish in quantitative terms to what extent tools, raw materials, and labor which directly or indirectly were used in the production of the good concerned contributed to the result. One would have to go back in these inquiries to the very origins of capital accumulation by saving on the part of people who previously lived from hand to mouth. It is not only practical difficulties which prevent such historical studies. The very insolubility of the problem of physical imputation stops us at the first step of such ventures. (Human Action, pp. 490–1)
