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.