Abstract
The literature on robot taxation has continued to expand since 2018 with numerous articles now referring to empirical evidence. The evidence presented in prior studies comprises abstract modeling and statistical pattern reviews with no statistically significant findings reported to date. Notably, one article is an advocacy piece by a tech lobbyist who at one point purchased priority Google results for the search "robot taxation." In some cases, technical errors are sufficient to reverse the stated results. Examples of error in empirical analyses include (i) motivated reasoning, such as the failure to model simpler or best explanations; (ii) lack of causal analysis; (iii) tax technical errors; (iv) omission of citations to conflicting theory or results; (v) errors in accounting methods; (vi) enhanced degrees of freedom in modeling parameters; and (vii) reliance on economic theories not reflecting robots as a fourth factor of production. The empirical evidence indicates that capital investment, such as in robots, occurs largely in higher tax nations, and that robot density is positively associated with high corporate tax rates, such as in Germany, Japan, South Korea, and the Nordic states, with little or no automation occurring in tax havens where the value of tax deductions for capital investment is zero.
This article summarizes the various reasons to be skeptical about the empirical evidence offered thus far in the context of robot taxation in Part II. A comprehensive review of the empirical literature on robot taxation is provided in Part III, where the existing empirical literature is categorized based on methodology applied, technical errors, degrees of freedom, and results.