From world output to human purchasing power
The original inquiry began from a deliberately crude benchmark: if global economic output is divided across the world's population, what does the resulting average tell us about the amount of purchasing power available to support high-value markets? The answer is not a welfare statistic and should not be confused with median income. It is a boundary-condition exercise.
Let nominal world output be G and population be N. A simple per-capita flow is:
g = G / N
If an illustrative world economy produced $100 trillion for 8 billion people, the arithmetic average would be $12,500 per person per year, or about $1,042 per month. If the economy were $160 trillion, the corresponding average would be $20,000 per year or $1,667 per month. These scenarios say nothing about distribution; they merely remind us that all nominal markets must fit inside a finite aggregate flow.
Purchasing-power-parity measures answer a different question: how much local real consumption a unit of income can command. PPP can be useful for comparing living standards and real output, but a company selling globally priced chips, cloud compute, aircraft or imported luxury goods cannot simply convert PPP output into equivalent dollar revenue. The distinction between real resource capacity and nominal monetizable flow must remain explicit.
Real capacity (PPP) ≠ cross-border nominal revenue capacity
The paper therefore uses two ledgers. The physical/real ledger asks what society can produce and consume. The monetary ledger asks how much money changes hands and who can capture it. AI can improve the first while compressing parts of the second.