Andrew, with that test file, I do see noticeable banding with my current setup.
I think this test file is a rather extreme case, good for demonstrating the effect. But in real photos I don't encounter obvious problems very often with my current setup. The most typical exception is when skies are darkened during processing, where sometimes posterization sets in. So when that happens, how much of it is "real" and will print that way, and how much might be due to a video path that has insufficient bit depth?
So I guess the bottom line for me is, does a higher bit video path lead to a better match between screen and print a significant portion of the time?
As a test, I printed the 10 bit test ramp on my Epson 3880 via a color managed work flow in PS. I could see some slight banding in the print, but considerably less than what I could see in the soft proof on screen. Maybe a 10 bit video path would have allowed a better match?