I'm pretty sure that's a real phenomenon. I have actually looked at the BLS and census data from 2000 vs 1990 and there is a widening when you compare deciles. But what I cannot fathom is the degree to which the conveyor plays a role. Here's what I mean by that.

Imagine that upward (and downward) mobility are conveyor systems on which the taxpayer sits. For simplicity, let's just consider the upward trend. There is a continuum from one end of the conveyor to the other. Taxpayers get off and on at various places along the way and they move at different speeds because of circumstances beyond their control and because of circumstances under their control. From the bottom end there are lots of people continually arriving on the conveyor. As time passes and circumstances evolve, fewer people stay on the conveyor. At the far opposite end a few people continue to lengthen the end of the conveyor.

Bot