This paper shows that the Federal Reserve and private forecasts of the unemployment rate for 1983–2000 have much in common: they produce very similar mean errors and mean squared errors, and, while biased, they are efficient and accurately predict the direction of change in the actual series. However, the private forecasts encompass those of the Federal Reserve and are thus more informative. In light of these findings, we first address several questions related to the unemployment forecasts and then discuss such policy issues as the Federal Reserve's credibility, accountability, and transparency.