I think the problem is that salary/tenure/promotion is tied to hazy, individual, non-standard assessments of impact, of which the calculated "Impact Factor" of journals that people publish in is one factor. My boss (and Public Library of Science founder) has a blog post related to this, http://www.michaeleisen.org/blog/?p=911 . Essentially, hiring/tenure committees don't explicitly sit down and plug in the impact factors of journals that a candidate has published in, but most people have a hierarchy of journals in their mind, with Science, Nature, and Cell (for biologists) at the top.