I think you're assuming all of the recruited moderators can objectively/mechanically moderate, all with the same moderation philosophies, and the same values reflective of the relevant population.
In reality, that's not how it plays out.
In the real world, moderators will disagree, so you might need to set up appeals processes, formal guidelines, processes for approving and improving the guidelines, and on and on.
It's not merely a ratio of moderators to participants. There is overhead above that.