As a simplified thought experiment, consider an a priori perfectly equal, static (equal numbers of people of all ages) society with the following characteristics:

* People are in school until age 20.

* Assets at age 20 are 0.

* People work until age 65.

* People die at age 85.

* Assets at death are 0.

Assume that inflation and investment returns are 0. After-tax income while working is uniform, say at $100k/year. Expenses are uniform. We'll pretend expenses before age 20 are 0 (covered by taxes). A really super-simplified model.

Total lifetime after-tax income is $4.5 million. This is 65 years worth of expenses, so expenses are $69231/year.

Peak wealth, at 65, is $1,384,605. Wealth around that peak is distributed like so as a function of age:

  59 -- $1,199,991
  60 -- $1,230,760
  61 -- $1,261,529
  62 -- $1,292,298
  63 -- $1,323,067
  64 -- $1,353,836
  65 -- $1,384,605
  66 -- $1,315,284
  67 -- $1,245,963

The top 10% of the population by wealth are all in that age range (because 9/85 > 1/10). Note that by construction this is a perfectly equal society.

The point is, looking at wealth numbers without considering lifecycle effects is usually pointless. Yes, having a paid-off house and a bunch of saved-up retirement money is not something most people have. It's something that generally only people approaching retirement would have. Younger folks have neither yet; someone who has been retired for a while may have the former but no longer the latter.

Now looking at wealth distributions while holding age constants is a more interesting exercise. You can still run into issues because of differences in lifeycles, of course. As a simple example, what can one say about the lifetime earning prospects of a 30-year old making $60k a year? Well, it sure depends on whether they're an elementary school teacher or a neurosurgery resident.

Unfortunately, good analysis of this stuff is hard. It's even hard to figure out what sort of questions make sense to ask....