What I mean is that making reviews happen is not a cost borne by publishers, thus is not a cost that should be recovered by publishers in the form of subscription fees.
Finding reviewers is hard: Good academics don't like to do it, because it doesn't contribute to promotion; bad academics are happy to do it because for them it does (at XYZ state), but they are at XYZ state for a reason, which means they aren't the best reviewers.
Finding reviewers is hard: Good academics don't like to do it, because it doesn't contribute to promotion; bad academics are happy to do it because for them it does (at XYZ state), but they are at XYZ state for a reason, which means they aren't the best reviewers.