This paper studies uncertainty about out-of-sample interest rate forecasts implied by an estimated Taylor rule. It is shown that the Taylor rule leads to a decomposition of forecast uncertainty into an element that depends on uncertainty about the future state of the economy and another element that is related to uncertainty about the monetary policy reaction function of the Federal Reserve. Uncertainty about one-quarter ahead Federal Funds Rate forecasts from 1975 to 2007 is estimated and analyzed using a real-time data set for the U.S.