The Taylor rule and interest rate uncertainty in the U.S. 1955-2006

We use a Taylor rule with time-varying policy coefficients in combination with an unobserved components model for the output gap to estimate the uncertainty about future values of the Federal Funds Rate. The model makes it possible to separate ex-ante interest rate uncertainty into three components: 1) uncertainty about the Fed's future policy coefficients, 2) uncertainty about future economic fundamentals, and 3) residual uncertainty. The results show important changes in uncertainty about future short-term interest rates over time with peaks in the late 1960s/early 1970s, mid 1970s and late 1970s/early 1980s. While for one-quarter forecasts uncertainty about the Fed's policy reaction is more important than uncertainty about economic fundamentals this result is reversed for the two-quarter forecast horizon. Results from a modified model with regime shifts in the variance of the policy shocks confirm the previous findings but show changes in residual uncertainty to be important as well.