I'd take CBO estimates over any of the others that the government produces without any question. The other sources are just about always more interested in arriving at a particular result rather than trying to produce a quality estimate.
In this case it may be explicit rather than implicit. I'm going from vague memory, but I think I heard someone who had looked into the process say that rather than starting with rosy assumptions (and then tweaking them if they didn't quite get the result they wanted) these guys literally put in the conclusion they wanted and told the modelers to solve for X as a growth rate to get there.
Could be I'm remembering wrong or it could be that the person was wrong, but it wouldn't surprise me given how cynical these guys have been.