To its credit, the CBO actually acknowledges as much itself each year, publishing a document called “CBO’s Economic Forecasting Record,” which is updated each July. It is intended as a warning to legislators and others that economic forecasting is not an exact science, and that the constraints under which CBO is required by Congress to work (constraints regarding assumptions about what future legislators will do, and about the effects of various policies on growth and employment) make the agency’s forecasting even less of a science. This past July’s update, which you can read here, should rid anyone of the notion that CBO’s long-term forecasts really tell the future. Their scores of legislation are still very valuable, especially as short-term and mid-term measures, and as raw assessments of direct costs (rather than effects on the economy), but the assumptions about economic performance generally don’t look very good in retrospect.