I wanted to say something about a very silly New Yorker piece by John Cassidy arguing that we should switch from our current method for measuring poverty (as a multiple of a household's minimum food budget, which for all its problems, keeps the idea tied to how well off people are in absolute material terms) to a relative scale that counts the "poor" as all those making less than half the median income. Fortunately, Will Wilkinson has a long post saying pretty much everything I'd wanted to and then some, so I can make this (ahem) relatively quick.
The obvious problem with the proposal is that even if you buy Cassidy's arguments for why relative position in the income distribution is important to people's well being in various ways, his proposal massively overcorrects by making it so that absolute income doesn't matter at all. On this methodology, if the income of every household in America quintupled overnight (in real terms, not just dollar terms), the poverty rate wouldn't budge at all. That's got to be wrong. Cassidy semi-anticipates this objection, and points out that the poverty rate can drop if the society's income distribution becomes more equal. But even that's not really true. Imagine the middle class gradually becomes miserably poor (in archaic absolute terms) while the top ten percent of the population becomes fantastically wealthy, leaving a sharp divide beween the Apollonian few and the Morlock many. That is, intuitively, a profoundly unequal society. But on Cassidy's methodology, the poverty rate might dwindle to almost nothing, provided the Morlocks were fairly uniformly immiserated.
Another odd thing about this is that Cassidy notes that it's problematic to calculate an "absolute" poverty line for a whole country: An income that leaves you at least somewhat comfortable in some rural town might not even cover rent in, say Manhattan—whether you're poor is going to be at least in part a function of what things cost locally relative to your income. But the same's going to be true when it comes to how people's comparison class affects their own subjective sense of poverty. Someone in that same rural town might feel poor if you dropped them on the Upper West Side, but not regard themselves that way at home. But at least with the absolute scale, you can measure local costs and try to adjust. How do you figure out what people's subjective comparison class for poverty or affluence is? And, incidentally, does this mean people in the developing world become poorer when they get access to global media and start comparing themselves to the rich West?