Work

Was Pareto right? Is the distribution of wealth thick-tailed?

Year: 2023
Type: preprint
Abstract: We fit log-normal, exponential, Pareto type I and Pareto type II distributions to US wealth data from 1989 to 2019 and examine the goodness of fit. Unlike earlier literature this paper uses high quality data, covering the entire US population, yielding powerful and unbiased tests. Beyond the 91st percentile the type II distribution consistently provides the best fit to the data and supports the hypothesis of a thick-tailed wealth (and by extension income) distribution. In addition, our results highlight the changing shape of the tail with decreasing concentration up to the 98th percentile and increasing concentration beyond. Our results suggest that practitioners modelling the distribution of wealth in situations where only limited data is available, a type I Pareto distribution might still serve as a valuable bias correction tool but should only be fitted to the top 1% of the population.
Authors: Wildauer, Rafael, Heck, Ines, Kapeller, Jakob
Language: English
Open Access status: green
OpenAlex ID: W7113794726
Cites: 63
Cited by: 4
Related to: 10