Simulating a Post-Automation Economy

Abstract

We develop an agent-based, stock-ow-consistent model of an economy undergoing automation, built to ask which scal instrument reaches the durable surplus that articial intelligence creates. The model separates two channels: a competitive return on reproducible robotic capital, and a mobile, foreign-held intellectual-property rent earned by AI. Production is an endogenous nested-CES technology; wealth concentration is microfounded through heterogeneous, persistent returns to wealth; and taxation and capital mobility are modelled as behavioural responses. The central result is that the durable surplus is the foreign-held AI rent, a cross-border licence fee that corporate, robot, and compute or token taxes largely miss and that only a source-based levy (a digital-services-style tax or a withholding) reaches. The appropriate policy depends decisively on whether a country owns the automation or imports it: for a host that owns the rent the problem is domestic inequality, reached by progressive and wealth taxes; for a rent-importing host the problem is base erosion and a gradual transfer of capital ownership abroad, which a residence-based wealth tax cannot reach. We report conditional orderings, stress-tested with global (Sobol) sensitivity and a formal stability analysis, rather than point forecasts.