Faber says "food" is 20% of income. You say "food" is 15% weight of CPI, which I presume is 15% of total expenses measured in CPI. Those are 2 different metrics and can't be compared at first glance.
You have a valid point. Faber does talk about income, and the CEX measures expenditures which are ultimately translated into weighting in the CPI. Which made me go back and look at the CEX figures and now I am confused. Looking at 2005 CEX I see that the figures people reported for all expenditures in the two lowest quintiles
exceeds their income before taxes:
Lowest 20 percent income before taxes: $9,676
Lowest 20 percent average annual expenditures: $19,120
Second 20 percent income before taxes: $25,546
Second 20 percent average annual expenditures: $28,921
If you look at Table 2 from the same 2005 report, you find it gets even more out of whack.
Consumer units with income before taxes of less than $5,000 show an income of $796, but total average annual expenditures of $19,684.
That same type of discrepancy continues for several categories up to the group labeled "$20,000 - $29,999".
The numbers do not compute.
But, back to your other points.
Faber is not claiming anything about pain, he is claiming that the CPI is bogus and does not show true inflation. If you compare what people spend to what they make then there is a measurement of pain they experience when something goes up in price. But measurements of inflation are based on what people
spend. Someone else will have to explain how a household with indicated income of $796 is spending 24 times that amount each year, but since inflation is a measurement of the increase in prices for the things they buy, all that matters is the accuracy of the expenses.
If I make $500 a month, and you make $5,000 a month, a 50 cent increase in the cost of a gallon of milk
hurts me a lot more than it would you (provided either of us even
buys milk). But the measurement of inflation is still based on that same 50 cent increase.