Yes, per subject per journal but then how many non-profits do you need? How do you organize them? How do you get a coordinated best-effort, etc...
I mean: corporations do not exist in a vacuum, they (usually) DO provide benefits to the society also.
I insist: I am not trying to defend abuses, I am trying to clarify that a for-profit corporation dealing with those many editorial issues is not bad per se.
Per subject, per journal, the same person who edits a journal today at Elsevier could do the same thing for the same salary at a university consortium-backed non-profit.