On the general point that individuals aren't the same as aggregates, I don't have a cite handy and I won't have time to look it up 'til tonite, but data does show that income mobility in the US has been declining in precisely the same time frame (post-1975ish) as the shift in income gains to the top cohorts. That is, studies that do track on individual or families rather than aggregatesALSO show that the poor are becoming ever more likely to stay poor. Or another way of looking at it, an individual's income is now more highly correlated with their parent's income than was the case from the 1930s to 1970s.

On these mobility measures, the US actually now scores _worse_ than many of the West European countries, if I recall correctly, though not quite as static as the UK.

As to Robert's specific point, from what I recall the income concentration data still holds when looked at the level of individual wage earners rather than by households. That totally removes the divorce argument -- the wage earner data simply compares paychecks with no reference to household composition.