Federal revenue forecasting

Federal revenue forecasts are crucial to debates about budget policy. But revenue forecasts are often very wrong despite being prepared by very talented analysts. This paper describes the forecasting process used by the Congressional Budget Office (CBO) and documents past errors. The errors tend to be serially correlated. That is to say, if CBO is too optimistic (pessimistic) one year, they are very likely to be too optimistic (pessimistic) the next year as well. The paper speculates about the reasons for this phenomenon. It concludes by arguing that the uncertainty of forecasts should play a much larger role in debates regarding budget policy.