I study optimal taxation of robots and labor income in a model in which robots substitute for routine work and complement non-routine work. The model features intensive-margin labor supply, endogenous wages, and occupational choice. I show that it is optimal to distort firms’ use of robots, thereby violating production efficiency. However, it is not obvious whether robots should be taxed or subsidized. The optimal policy exploits general equilibrium effects to compress the wage distribution which relaxes incentive constraints and raises welfare. If robots polarize the wage distribution, a tax on robots compresses wages at the top but raises inequality at the bottom. The sign of the robot tax depends on which of the two effects dominate. In addition, occupational choice partly offsets wage compression, thereby limiting the effectiveness of a robot tax. I use the model for quantitative analysis based on US data. In the short-run, in which occupations are fixed, the optimal robot tax is positive and sizable, but its welfare impact is negligible. In the medium-run, with occupational choice, the optimal robot tax and its welfare impact are diminished further – and approach zero as the price of robots continues to fall.
How should redistributive governments change tax and education policy in response to skill-biased technical change? To answer this question, this paper merges the canonical model of skill-biased technical change due to Katz and Murphy (1992) with the continuous-type Mirrlees (1971) model. Workers of different ability face an extensive education choice to become high-skilled. Wages are endogenous. Although we allow tax schedules to be conditioned on education, optimal marginal tax rates are independent of whether individuals are low-skilled or high-skilled. Only the intercepts of the optimal tax functions differ between low-skilled and high-skilled workers. Optimal marginal income tax rates follow the same formula as in Mirrlees (1971). We show that education should optimally be taxed on a net basis. Moreover, optimal tax and education policies do not exploit general-equilibrium effects on the wage distribution to reduce pre-tax earnings differentials. Skill-biased technical change raises optimal marginal income tax rates, especially in the middle of the earnings distribution. SBTC has ambiguous effects on net taxes on education as both distributional benefits and distortions simultaneously increase. Numerical simulations demonstrate that, strikingly, optimal income taxes and education subsidies hardly change in response to SBTC. SBTC raises marginal taxes for middle-income groups while net taxes on education may either decrease or increase.