I'm re-reading your post (because it didn't quite make sense to me the first time around). Do you mean you are comparing two sections (to resist a given load) and one has a lower actual stress, but that same section also has a lower buckling stress?
If that's the case, it means that even though your total area increased (to get the lower actual stress), your least radius of gyration decreased (resulting in a lower buckling mode).
It also suggests that the section is in an area of elastic buckling and the stresses are being governed by that and not yielding of the section. If you have smaller kL/r values, the critical stress approaches the yield stress and small changes in L or r don't make as big of a difference in the critical stress.