Discourse 1 - Sidebar:
In looking at all this data,
the statistical variable fairy appeared before me and posited the question as
to whether the income data might, in fact fit some standard statistical model
and if so what type it was, what the fitting parameters were for that model,
and what the fit to the data looked like. The answers turned out to be
interesting, at least to the fairy and me. The data seems to fit an Exponential Distribution.
The fit is pretty accurate to a formula of Y = 17099e1.8126*X,
where y is income and x is log(1-Òpercentile groupÓ). What this means, if anything, is unclear to me or the fairy and hence has been
relegated to a sidebar. The one
thing it MIGHT be useful for is writing analytical expressions to calculate and
optimize tax rates, etc. My
intuition also says that the fact it fits such a nice formula is an indication
of a standard statistical process setting things rather than a ÒriggedÓ
distribution but this is based on nothing other than intuition.
