This article: Region's Job Growth a Centrifugal Force starts badly "As a consensus builds that the Washington region needs to concentrate job growth, there are signs that the exact opposite is happening." and gets worse.
First, consensus implies agreement, clearly the market does not agree, hence the continuing suburbanization of jobs in the DC area. Since the people are alreadly living in the 'burbs disproportionately, employers seeking to be near their workers are following them out to the burbs. Only the "planners", (whom we might dub "centralizing planners") seem to have adopted this consensus, and even then, the planners for the outer counties surely do not agree.
Second, the evidence from Washington is that increasing suburban jobs decreases commute times by auto (though of course, not helping transit riders much). Commute times are not the only criteria, but they are what people personally care about. For historical reasons, Washington, like most cities, grew with a strong employment center, so that there are more jobs in the center than workers, and more workers living at the edges than jobs. This has been changing, but the dynamic is not fully played out.
So the question is really one of "first best" vs. "second best". The world is auto-oriented (in our book Access to Destinations, we describe climbing down "Mt. Transit" and up "Mt. Auto" in the past century). Trying to make transit-favorable policies (like centralizing jobs) is suboptimal in the suburbanized auto-reliant world we actually live in. Many, many other changes would need to take place for centralizing jobs to be truly optimal.