A mathematical analysis argues that national averages and individuals’ long-term income experiences should be reported separately.
Economic growth is usually summarized by the change in gross domestic product per person, an average across incomes. The study examines how that average can differ from the growth experienced by an individual over time when incomes follow a random pattern in which changes build on previous income.
The authors develop mathematical properties for both rates and argue that relying on the average alone can give an incomplete picture. They illustrate the framework with income data from the United States and France covering the past 75 years.
Two ways income grows
Under the study’s model of random multiplicative income growth, the average growth rate of income exceeds the long-run growth rate experienced by individual incomes. The authors attribute this difference to the process being non-ergodic, meaning that an average across people does not necessarily match the long-run experience of one person.
They establish mathematical properties of both growth rates and identify a natural measure of inequality that grows at the difference between them. The framework is illustrated using income data from the United States and France over 75 years.
// Source
Philosophical Transactions of the Royal Society A Mathematical Physical and Engineering Sciences · 2026 · DOI: 10.1098/rsta.2025.0163
Authors: Alexander Adamou, Yonatan Berman, Ole Peters
Institutions: King's College London, Santa Fe Institute, London Mathematical Laboratory