Continued automation and declines in low-skill shares of GDP have been widespread
globally and linked to inequality. We examine the long-term, global consequences of
policies that foster automation or address the distributional consequences of it, using a
six-region global macro model. Results depend on whether welfare criteria are Rawlsian,
emphasizing the performance of low-skill households, Benthamite, which aggregate
pecuniary measures, capital-owner friendly, or simply based on real GDP. Even where
automation delivers only bias against the low skilled, we find that the fostering it is a
dominant strategy under all but the Rawlsian criterion. We then consider a post
automation scenario in which worker displacement is significant, examining inequalityconstraining
but balance-preserving fiscal interventions, such as tax-financed “earned
income tax credits”. These generate only small international spillover effects and are for
the most part not preferred under all criteria except the Rawlsian one.