I'm saying that it isn't. Modulo weird edge cases, the publisher delegates that decision to reviewers. If they choose to reject a paper, then it gets rejected.
Unless I'm really mistaken about how things work. My only first-hand experience is in PL, a field in which we publish in conferences rather than journals. I've sat in on a PC meeting where the decision to accept or reject papers was made. In cases where the reviews were clear, the paper was accepted or rejected without fanfare. For borderline papers, the decision was made through discussion among the reviewers who had reviewed that paper and the PC. I don't think the PC members were paid for it. The reviewers certainly weren't.
> Nobody cares what they think.
I care what they think. It would be really interesting and informative to have a system where paper reviews were public.
EDIT: What's your experience with publishers? Do you know more than I do?
Excuse me, but this decision is not being made by the publisher. It's being made by hard working volunteer reviewers, who are typically not paid by the publisher. These reviewers are the people actually deciding what papers are accepted or rejected, and thus deciding which you read, and they deserve credit for it.
I think you mean to point out that the publishers are the abstract entities to which prestige is attached (often for historical reasons), which they certainly are. Though that's slowly starting to change, as researchers move towards open access and away from for-profit publishers.