a bushing die will help with ES's
i prefer the bushing approach , with or without the expander , depending on how much sizing is being done , that has alot to do with the starting brass neck thickness , expander outside dimension , the inner dimension of the sizer die & , and how much its moving the brass on each operation
i think the bushing first to .001 under your intended diameter , which will leave stress in the brass/ spring back . if then opened up with the expander mandrel to your intended size , reduces the stress left from the bushing die and because the expander only sized a small amount , it would have not created very much spring back in the other direction , plus the remaining spring back will be in the same direction as when seating ( it will want to squeeze tighter ) -- this method also leaves you with the best run out possible IMO as long as your seating operation after doesnt create it
you could get away with just using a bushing without a die expander . but in that scenario id hope that all neck wall thickness's were very consistent ( turned ) and go with a slightly smaller bushing so when seating a bullet it basically will overcome that sized down stress/spring back , then basically use the bullet to expand it just past were where the neck wants to be relaxed ...
if you anneal after any way of sizing you go , itll create a very relaxed state of brass at your intended dimension , that is where that low ES happens IMO , especially if you add dry neck lube to that mix
the thicker the neck the more spring back memory you have