>Actually, this does support my main point, that
>income distribution gets estimated much more than
>wealth distribution. The status of wealth distribution
>estimation in the US, a Fed survey every three years,
>is about where income distribution studies are in most
>LDCs, a half-baked survey every several years.
There's nothing half-baked about the Fed's Survey of Consumer Finances, though; not sure if you're implying there is. It's much more rigorous than the Census Current Population Survey, involving a 75-minute interview backed up by financial records - compared with the CPS's almost casual questions with answers based strictly on recall. Rich people are heavily oversampled in the SCF.