Yonatan Berman and LML Fellows Ole Peters and Alex Adamou have published a substantially extended version of their recent paper on American wealth distributions, available here
The dynamics of wealth distributions allow us to test empirically the ergodic hypothesis as it is made in this field of study. While relative wealth (personal wealth divided by population average wealth) is typically assumed to be ergodic, the data suggest otherwise. Ergodicity of this quantity would imply that each individual spends 1% of his or her time being in the wealthiest 1% of the population etc., and a requirement for this is that wealth tends to be re-allocated from rich to poor (stabilizing the distribution).
Wealth in the US has effectively been re-allocated from poorer to richer over recent decades. This necessitates a re-visiting of many fundamental beliefs in the field.