Paul Krugman's piece for The Nation on the death of American economic mobility has been getting abundant link love over the last few days. I don't have much to say here, especially since I tend to agree that the Hubbard study he mentions has a number of problems that make it a poor ground for the strong claims often made by those who cite it. As I recall, for instance, it looks at people who filed tax returns in ten consecutive years. That may be necessary to get a good longitudinal picture, but it introduces some obvious selection bias.
Still, it's worth noting this:
But what these studies measure, as the economist Kevin Murphy put it, is mainly "the guy who works in the college bookstore and has a real job by his early 30s." Serious studies that exclude this sort of pseudo-mobility show that inequality in average incomes over long periods isn't much smaller than inequality in annual incomes.
I understand why that sort of mobility isn't the kind with which people are concerned, but it seems we've got to be consistent. I'm pretty sure that most claims about the degree of American income inequality don't exclude people who have low incomes because they're 20 and working in the college bookstore. If you count these folks when you're looking at inequality snapshots, but then exclude them when tracking mobility, that's going to skew your perspective as well.