AI and the Economy: The Four Rates Which Define The Future (Short Version)

Abstract

The economic debate about artificial intelligence has produced a wide range of estimates and almost no coherent framework. This paper bridges that gap. Our modeling shows four implications:

• AI will displace 19.5–21.5 million workers and generate fiscal ca- pacity that scales with AI speedup. The $2.52 trillion gross produc- tivity uplift ($2.01–2.17 trillion net of demand drag) generates incremental federal tax capacity that exceeds the displaced workers’ permanent income loss at base case speedup (s = 1.38) and grows further at higher s. We show that a faster, more capable AI strengthens the fiscal position while leaving the number of displaced workers unchanged.

• AI will be more deflationary than the China trade shock, and more permanent. The combined supply and demand channels produce an average deflationary impulse of more than −1% per year; more than 50% larger than China’s peak contribution to US disinflation (hitting services rather than goods).

• AI’s footprint is already measurable in current data. We document $642 billion in excess productivity above pre-2022 trend, 2.9 million missing jobs in high-exposure sectors, and wages decoupling from productivity in direct proportion to each sector’s automation ceiling.

• A coherent framework for the AI debate. Via a single framework, we endogenously derive the impact of AI on GDP, jobs, prices, and vendor economics, producing internally consistent forecasts. 

Keywords: AI economics, labour displacement, productivity, Amdahl's Law, Workflow Completeness, general purpose technology, inflation, TAM