Thursday, January 9, 2020
Daniel Hemel (Chicago), Does the Tax Code Favor Robots?:
In recent months, a number of scholars and commentators have articulated versions of the following argument:
- U.S. tax law favors capital over labor;
- Robots are capital;
- Therefore, U.S. tax law favors robots over labor.
Three implications tend to be drawn from this syllogism: (a) that U.S. tax law leads to inefficient investments in automation; (b) that automation — because it is capital-intensive and capital is tax-favored — will result in a reduction in tax revenues; and (c) that policymakers should respond to the automation trend either by imposing explicit taxes on robots or by raising taxes on all capital.
This short essay — an edited version of remarks at the Artificial Intelligence and the Future of Tax Law Symposium at the Ohio State University Moritz College of Law — seeks to illustrate why the line of argument above is misguided.
First, the claim that U.S. tax law is biased toward capital rests entirely on an unstated (and uncertain) normative premise: that the United States should tax income rather than consumption. If an income tax is the baseline, then U.S. tax law exhibits a pro-capital bias; if a consumption tax is the baseline, then U.S. tax law exhibits an anti-capital bias. Which baseline we choose depends on normative choices that claims of capital-favoritism tend to occlude. Second, robots do not only (or even primarily) represent “capital”; they also embed the labor of engineers and others. The labor of robot makers is often taxed at unfavorable rates relative to the labor of the workers whom automation threatens to displace. Third, the idea that U.S. tax law incentivizes firms to replace human workers with robots rests on doubtful logic, and the claim that automation will erode the tax base finds little support either.
This essay is not an argument against capital income taxation or a defense of the current Code, which does tax capital income but not all that much. The case for capital income taxation will be stronger, though, if it is based on firm foundations rather than on dubious claims of robot favoritism. The essay also is not a full treatment of the arguments for and against taxing capital. Its objective is to evaluate one such argument and to show why it is unpersuasive.
Part I of the essay examines the claim that the U.S. tax system favors capital over labor. Part II turns to the question of whether robots represent capital or embedded labor. Part III considers the case for explicit taxation of robots or broader taxation of capital once illusions about the tax code’s pro-robot bias are cleared away.