I think that this is a very high priority issue that is raised here.
The section editor role has been suggested as a work around to a situation where an editor will submit also her own articles for review. See for example @asmecher’s answer here: Hiding editor’s own article including reviewer identities?
If the situation really is as described above (the user has both section editor and author roles, but in that submission she is only assigned as an author), the whole role/permission level of a section editor becomes basically useless.