> >Average household incomes usually come in below per capita GDP, with
>>the discrepancy tending to widen with income. Milanovic uses
>>household survey numbers.
>>
>
>
>But it is difficult to estimate without launching a survey. Assuming that
>GDP is by definition the sum of all factors of production, we can
>approximate income inequalities if we know the GDP, the total population,
>and the percent share per quintile.
Uh, maybe I didn't make myself clear. Milanovic blended the results of about 80-100 national household surveys to synthesize a world household survey.
GDP includes corporate profits (on the income side) and investment (on the product side), whose share of GDP roughly rises with income. It'd be wrong to attribute that income/expenditure to households.
Doug