Gender-specific top incomes: are they Pareto distributed?

Abstract

Estimating the Pareto alpha parameters from the top income distribution separately for women and men can provide information on gender inequality at the top (Atkinson et al. 2018). However, whether the top incomes for women and men are really Pareto distributed has not been tested. I fit a Pareto type I distribution to the Finnish total population administrative income data for the years 1995-2014 and find that for women, a Pareto distributional assumption is only plausible for certain years. Also, I find that Pareto models apply for very high incomes which are above the commonly assumed income thresholds for Pareto distribution.