Overnight interbank loans are a closed system, by definition.
The part that isn't a closed system is when people use Libor as a metric for other purposes. Which isn't a reasonable complaint, because that's explicitly not what the Libor was created for, and it was never intended to mean anything else.
That is: Things grow beyond "what they were created for". To say that any such use is fair game for abuse is both unreasonable and unwise.
No, it's not at all.
It's more like when people complain that memcache doesn't make a good tool for session handling, or that it's not fault-tolerant and has zero replication. Of course not - it was never meant to be used that way, and the documentation explicitly says it shouldn't be used that way[0]. If you still want to, fine, but don't complain that memcache is a bad system when you're simply using it against its intended design.
[0] https://github.com/memcached/memcached/wiki/ProgrammingFAQ#h...
So, to see which analogy fits better: Is there a document (issued by someone authoritative) that says that Libor should not be used the way it was (i.e., to set interest rates for non-interbank loans)? If not, then what we have, essentially, is your claim that it shouldn't be used that way, against everyone using it that way.
Except that it's not a closed system. That would be the point.