A recent NYT blog article discusses recent work by Richard Burkhauser and coauthors on measuring trends in inequality in income net of taxes, transfers, health insurance, capital gains, and kitchen sinks. What a job that must be, I'm glad I haven't waded into that pool! The blog article contains some great quotes from leading economists in the field, and for me the biggest issue is raised by Gary Burtless: to figure out what happens to assets, Burkhauser et al. assume everyone gets the same market returns, and we know that's not true. It's hard to draw any definite conclusions before peer review hopefully exposes any kinks in methodology or lays concerns to rest.
We want to know what's happening to income inequality now, year-to-year, but my two cents is that "completed" or "cohort" inequality, what people have ended up with over their lifetimes after all the dust settles, is the outcome we're more concerned with, and not so much the year-to-year fluctuations unless those appreciably affect long-run outcomes. Do we know whether they do? Ah, the challenges of social science when, for very good reasons, it's very difficult to observe every vicissitude for an individual over many periods.