The Congressional Budget Office is tasked with “scoring” legislation (i.e., figuring out the net costs of a proposed bill were it implemented). Unlike their counterpart the Office of Management and Budget, the CBO is traditionally considered a non-partisan agency. As such, people–especially members of Congress–tend to trust and rely heavily on the estimates CBO produces.
All that talk you’ve been hearing about health reform costing $1 Trillion over 10 years? That number–and variations around it–comes from the CBO. The thing is, the folks at CBO know that cost projections are an inexact science. That’s why they refuse to make long-term projections beyond the 10-year window. At that point, they concede, there are simply too many unknowns for reliable estimates to be made. Fair enough, but I think that tends to make us put just a bit more confidence in their 10-year numbers than perhaps we ought. After all, those projections–even in the first year–are based on a wide range of assumptions.
As Jon Gabel writes in yesterday’s New York Times, the CBO may be non-partisan, but it tends to err on the side of making conservative estimates. That doesn’t mean that estimates are biased to favor the Republican party, but rather that estimates “play it safe” — they overestimate new costs and underestimate savings. The result is that CBO scores tend to be higher than the reality that actually plays out. Gabel’s piece finds this by judging CBO estimates of health reform retrospectively. Looking back now, what did CBO project things would cost and how much did they actually wind up costing? He finds that, for the three major health reforms looked at, the CBO estimates were too high all three times. The reason? Well, I want you to read the article, so I’ll just say it has to do with making assumptions about unknowns when there is no historical precedent to rely on.
What does all of this mean? Just that you should take CBO estimates with a grain of salt, because after all, they predicted that you’d actually use two grains of salt.