If meta-moderation powers were assigned randomly and uniformly, this could be gamed by just spam-creating tons of accounts. Any system that accepts user input needs to have a robust answer to the question "What if a significant percentage of my users are actually bots under control of a single person?"