Oops, you’re right. I must have read the 2001 number as 1991. I stand corrected.
No problem. BTW, here’s another strange “artifact” of statistical analysis. When there is a recession, sometimes the numbers get driven UP, while the tendency in the “boom” part of a recovery is to drive the numbers down.
This is because we usually speak of “wage-earners” (Note that in THIS study, it was only wage-earners with FAMILIES and CHILDREN). Recessions and their job losses tend to hurt those at the lower end more than the upper end, as those workers are more easily done without, and more easily brought back. So we lose a lot of low-wage-earners, which RAISES the wage averages in all the quintiles. Sure, some higher earners are effected, but so long as a larger number of lower-income workers leave, it raises the averages.
When the hiring first starts up again, it’s never across the entire spectrum of wage-earners, and tends to be concentrated at the lower end of the spectrum. So the unemployment rate drops, things are “better”, but the wages for the lower quintiles drops.
Except not the lowest quintile as much, beause it is effected by the fixed-benefit model driven by tax dollars. The bottom quintile was pretty much at the lowest end of the plausible pay range, so whoever gets added back can’t really draw the lowest quintile down that much.