For me a bushing die increases the chance that the neck will be out of alignment with the case. The bushings have enough clearance to move and that means they can twist or be off center and transfer that to the case. I don't have any idea how often that happens but it is one more variable to deal with.
Funny you mention that, I just worked through that recently with two different cartridges and die set ups.
1. Last night, loading 6.5x47 using a Redding FL bushing die. I have been using the practice of leaving a bit of a gap in the stem of the die so the bushing would float; read it somewhere.... For whatever reason, I decided to measure the concentricity. The fired brass was inside of .001. My just-sized case necks were .005!!! I screwed the stem all the way down leaving no play and re-sized them. The sized necks returned to inside .001. Simple to control, just screw the stem down.
2. A couple of months ago I switched to a neck only bushing die for .284W b/c it seemed to size more of the neck than the FL die. My accuracy deteriorated. I called Redding complaining about the fact that my FL die wasn't sizing enough of the neck. I thought I checked everything before I called. They were polite but were obviously dubious and offered to ship the die to them to have it checked out. I checked again after the phone call and realized I hadn't used the little spacer nut under the stem(i don't use the decapping pin or rod). Therefore the bushing was allowed to float up into the die. I added the spacer and the die sized %90 of the neck and my .284W returned to regular .3's. I called them back to admit my mistake and they mentioned that a lot of folks are seeing better case concentricity using FL dies and there has been a movement back to them from the neck/ body die combo. Since then, I have only purchased FL bushing dies.
If you use the dies correctly, they work well.
Two different methods:
1. Uniform the necks with an expander mandrel to even out dents from STM tumbling and possibly getting stepped on at the match, then reduce the neck to final desired size and "tension" using a bushing while the case is supported by the FL die walls.
2. Reduce the neck using a non-bushing, conventional die setup, possibly leaving some of the dings and dents, then drag an expander ball through the neck while it isn't supported by the FL walls to uniform and bring it back up to size that may or may not match your neck/ bullet thickness very well.
There is a lot less brass work in the first method b/c the necks are already expanded from firing and the mandrel doesn't even really stretch the brass unless it is already out of uniformity from handling. Often times, if I am shooting in a controlled environment and I am confident my necks aren't out of round, I'll simply omit this step all-together. Currently, until I learn otherwise, my thought is that concentricity is better and more consistently achieved by resizing/ forming the final case dimensions by the "shell" or walls of the die rather than by dragging a ball through the neck while the case is being pulled by the shellholder which may induce slop via the loose shellholder/ case head fit or the ram being even if slightly off axis. Perhaps those last two compenents take a lot of control for uniformity out of the die's hands vs final fitting being achieved inside the die ......and then perhaps the allowance for play in the shellholder allows the case to self-center on the expander ball. At this point, I am just opting for less work on the brass.