I'm not even entertaining the idea of using a bushing with no mandrel. Not saying it won't work well with the appropriate size bushing, but therein lies the problem.
Forget imperfections getting pushed to the outside. There is a much better reason to mandrel. Brass neck thickness variation between head stamps.
The bushing sets your tension/interference based on brass neck thickness. Any variation in neck thickness = different tension when using only the bushing.
I'm currently working on 6ARC loads. Hornady brass loaded neck diameter is .26805. Lapua loaded neck = .2705. If I were to use a bushing to set tension I'd need 2 different bushings or I'd get + .002 interference on the Lapua. I just use the same bushing for both followed by the same .2405 mandrel (gas gun). The mandrel eliminates that significant tension variation I would have had.
I've got like 7 different head stamps for 308, x 4 for 6.5C, X 2 for 6GT etc. The fuck if I'm going to try dialing in a bushing size for each one.
I just did a 6ARC load ladder that shows the value of using the same mandrel to set tension on Hornady & Lapua.
All brass trimmed same length. Loaded same BTO, loaded in once fired brass, annealed, LE Wilson FL bushing sized, .2405 mandrel.
29.3 Lapua = 2729 Hornady = 2733
29.6 Lapua = 2748 Hornady = 2760
29.9 Lapua = 2779 Hornady = 2784
30.2 Lapua = 2831 Hornady = 2835
Different bullet
30.3 Lapua = 2773 Hornady = 2770
Maybe I could get similar results with 2 different bushings?
Here is neck lubing tip that's been working great for me.
Just take the next size down VFG pellet on their mandrel soaked with the imperial graphite & brush it through the necks. The pellet holds the graphite in the fibers well & doesn't make a mess. I tap the pellet over the graphite container as you don't want a bunch of loose graphite caked on the pellet, just even coating in the fibers. This has by far given me the most consistent seating pressure.
Edit Pic added: