There's been a great deal of buzz about the new "supplemental povery measure" released by the census, which seems to show a fair amount more poverty than we thought we had. Cue hand-wringing! But Mickey Kaus points out that neither the Census, nor the reporting, has talked much about the biggest difference between the supplemental measure and the old, boring, measure it's supposed to be supplementing: the old measure was an absolute figure, pegged to the cost of a basic market basket of food. The new measure is a relative measure, pegged to (approximately) how much the 33rd percentile is spending on various things.
The biggest difference between these two measures is not that the supplemental measure adjusts for government transfers, local cost of living and health care expenditures, though that's what you'd think if you read most of the reporting on it. The biggest difference is that while the old measure told us about peoples' ability to access the most basic neccessities, like food, the new measure tells us how well the bottom is doing relative to the second quintile.
In other words, under the old measure, it was possible, if unlikely, for the poverty rate to go to zero. Under the new system, it's not--at least, not unless we engage in really radical egalitarian levelling. The only way for the supplemental measure to fall is for the income of the bottom to rise relative to the income of that 33 percenter. It's a measure of inequality, not poverty.
You can certainly quarrel with the method for calculating the current poverty line, which is based on the average household budget of the early 1960s. But that's an argument for a supplemental poverty measure that calculates household needs in a more sensible fashion--not for transforming it into a measure of inequality rather than absoltute want. And if you ware going to peg the figure to incomes rather than the prices of basic goods and services, that should be emphasized in the reporting, so that people understand what they're dealing with.