Depending on the org structure, pair programming can sometimes be a clearly superior version of a code review policy. Particularly when the review latency is high and, as a consequence, developers work on several branches in parallel.
It is, but with some differences; works better for remote teams (see github), is async (don't need to interrupt someone to get cracking), is more suitable to introverts (or anyone really, who can pair program for 8 hours / day or more?), and you get to sit down and review someone's code with a fresh pair of eyes instead of not catching mistakes because you wrote them or were there while they were being written.