I'm not totally an expert in that area, and somebody else wrote that chapter, but my understanding is that the assumption is that most of that difference is accounted for by birth rates but some of it may be accounted for by changes in identification.
On May 30th, 2006. See this statement in context.