To be explicit I'm not talking about neck sizing, I'm talking about the sizing of the neck portion which also occurs in full length sizing.
I'll give you an example via 308 Winchester since I have a bunch of the right numbers for it stuck in my head. You can turn it mutatis mutandis into whichever cartridge is actually in play.
I have some Lapua brass with neck walls .014" thick. Twice that plus .308" for a seated bullet means the loaded neck measures .336" outside diameter. I have a Forster full length die with its neck section honed to .333". So that die sizes the brass just right for a proper interference fit when seating the bullet, and the expander ball need not be used.
In contrast I have a Redding full length size die whose neck portion squeezes the brass down to a very tight .328" so the expander ball has a lot of work to do on its way out of the case. This causes increased effort withdrawing the case from the die and is a source of increases runout. Furthermore it wears out the brass faster. If the neck only needed to be sized to .333", then the trip down to .328" and back up to .333" over the expander ball added .010" of unnecessary cold working of the brass during this one prep cycle.
Now consider some thin WW brass I have with .012" thick necks. This means a loaded neck will measure .332" so if you run the numbers for this brass you'll see the Redding die is pretty close to ideal and the honed out Forster die is useless.
