Actually, if you want to fix the problem, peer-review should be rewarded rather than publications, with accurate peer-review and unbiased evaluations. Publication count should not be nearly as valued, and peer-review should not be anonymous.
And how do you even guarantee this, 'accurate' and 'unbiased'? Any form of peer review will always have a slant, no matter how 'objective' it is intended to be.
Until that can be ascertained, I don't see how peer review is fundamentally superior. I would even argue that peer review can be detrimental to true science, especially those on the frontier.
Peer review of peer review. Instead of a couple of scientists looking at it, the peer-review process should be entirely open. So other people can scrutinize it.
How do you prevent movement to an equilibrium where reviewers give favorable reviews to the reviewers of their own papers?
Peer-review of peer review. If all peer review is public, then people can analyse the trends.