No, but found a Redpaper authored by Dawn May and others. They recommend '0'. Apparently system jobs, subsystem monitors, and secondary threads do not lock the library list anyway. It's only a regular job's primary thread that locks. It mentioned a performance benefit by reducing number of locks on the system.

Don't know what happens to an active job if you delete or rename a library that's in it's library list. Guessing it will probably would be OK until next time it tries to find an unqualified object, then it will end abnormally. (Not entirely unlike *PGM in a call stack, DLTPGM, then calls eventually return to the non-existent program.)

Before I clear, delete, or rename, I do WRKOBJLCK on *LIB; you'll lose that. To be safe you'd need a little program or command that lists all jobs and checks their library lists.

Save won't be affected, really. You'll still need save-while-active because of objects in the library.