China Merchants Securities' macro research team has published a report on how AI may disrupt taxation. It first compares past technology revolutions with tax-system shifts: Britain introduced personal income tax in 1842 after the first industrial revolution, corporate income tax spread worldwide after the second, and the internet revolution prompted a cross-border consumption tax framework, with the lag between technology and tax change shrinking from over eighty years to roughly a decade. The team argues AI mainly hits the labour tax base, since fewer jobs and wider income gaps make labour-related revenue less stable. New questions such as how to depreciate computing power, measure data and price models also weaken the taxability of capital, while pressure on household income indirectly affects the consumption base. Scoring AI exposure, labour share, reliance on labour taxes and fiscal fragility, the report finds Germany, the US and Japan most exposed and emerging economies relatively insulated. On responses, it judges that a new robot tax is unrealistic and that robots hardly qualify as taxpayers, so refining rules inside the existing system is the more feasible path.
News Detail
China Merchants Securities Macro: How Will AI Disrupt Taxation?
China Merchants Securities' macro team published a report on how AI could reshape taxation, tracing how each previous industrial revolution triggered tax-system change and noting that the response lag has shrunk from roughly 80 years to about a decade. It argues the core risk sits with the labour tax base: job contraction and wage polarisation weaken labour-linked revenue, while depreciation disputes over computing power, data measurement problems and AI pricing gaps erode the taxability of capital, and squeezed household income indirectly hits consumption. Scoring AI exposure, labour share, reliance on labour taxes and fiscal fragility, Germany, the US and Japan show the largest exposures while emerging economies are more insulated. The team views a robot tax as problematic and considers refining rules within the existing framework more realistic.
2026-09-28