I've spent more time trying to understand the Armour, Burkhauser, and Larrimore paper discussed in my last post. The idea of that paper was that you should count gains or losses in the value of assets during a year as income for that year. That seems reasonable, at least for assets that could be sold easily. Their paper was based on a comparison of 1989 to 2007, on the grounds that those two years were at the same point in the business cycle. Investment returns were very different in those years: by their calculations, stocks went up by 27% in 1989 and 6.4% in 2007. So suppose that in 1989 someone had a salary of $200,000 and $300,000 invested in stocks. Their income was thus about $280,000 (salary plus 80,000 increase in the value of the investment). Suppose that in 2007 that person had a salary of $250,000 and still had $300,000 in stocks. That comes to an income of about $270,000--that is, a decline from 1989.
But which is better, having a salary of $200,000 and $300,000 in investments or having a salary of $250,000 and $300,000 in investments? Of course, if there was actually a long-term decline in the rate of return on investments, then that would be bad for affluent people, but that doesn't seem to be the case. Investment returns (especially stocks) just go up and down from year to year--there's some relation to the business cycle, but it's not very strong. So in effect, they're adding an error term (investment returns). Doing that naturally makes it harder to see the trend (the rich have gotten richer).