Re: The CBO’s Crystal Ball
Earlier today Jonah asked if there were studies of the accuracy of CBO’s economic forecasting. There have been a few over the years, but the most useful are actually the ones that CBO itself does every year — they’re very frank assessments of their own accuracy, comparing their predictions with what really happened. Here is the latest version, from July of 2009.
The bottom line is what we might expect: their short-term forecasts are pretty good, their long-term forecasts are pretty bad. Assumptions about the ten-year effects of a bill that will reshape a sixth of the economy and is full of unrealistic assumptions about what future congresses will do are very likely to be way off. But they’re still useful benchmarks.
I think Paul Ryan had it just right on this subject at yesterday’s summit: CBO does an excellent job given the parameters under which they are forced to work, but those parameters are often not very realistic. You can see in just about every one of the CBO reports about the Democratic health-care bills in the past year that just below the calm surface of the text are analysts waving red flags and screaming “we’re talking about politicians here, this will be much much more expensive!” They do their best to communicate that clearly, but not too clearly.