Over at Max's place Dean Baker has a post covering various issues about the consumer price index. I like his closing line. Short version: lots of people think the CPI still overstates the true rise in the cost of living, blah blah blah.
While I remember reading quite a bit about this stuff in my grad school days, one dimension of the issue I don't remember ever being addressed was how the CLI - the "ideal" cost of living index of which the CPI is an approximation - varies throughout the income distribution. That is, roughly speaking, the CPI tracks the price of a fixed basket of goods purchased by a "typical" consumer. I've never seen any research, though presumably it exists, about how that fixed basket, and resulting index values, would be different if we looked at the "typical" consumer in each, say, income decile.
