This is crucial in the debate i think. If i understand the logic here:
If wealth "skew" and the fat tails were purely related to earnings one could make a plausible argument (though still dubious IMO) that this related to distributions of talent (and/or how talent interacts with production i.e. entrepreneurs skills are multiplied by all the people who work for them etc).
However, if wealth skew is > earnings skew that implies some other institutional process is at work that isn't really to do with individual talent or effort.
At a moral level this has big implications.
Put crudely, imagine the simple random returns to capital each period and there are no difference in talent, effort etc. This generates a simple lognormal distribution or (as they explain above) the fat ones when there is a reflecting barrier (i.e. birth / death).
In this model there is no "justification" for resulting differences in wealth -- they are purely "random".
This is basically the policy / moral background to this whole technical paper: are wealth distributions a result (largely) of random chance (and accumulation) or talent.
Because if the former then the resulting inequality has no moral legitimacy and no practical value. If the latter, there is, at the very least, an argument for practical value (in terms of rewarding talent / effort etc).