Revisiting the Relative Forecast Performance of Fed Staff and Private Forecasters : A Dynamic Approach

This paper aims to extend the findings of Romer and Romer (2000) to a setup where the time variation of (relative) forecast performances is addressed in much greater detail. We show that the relative forecast performances of Fed staff and private forecasters are not stable in the presence of large macroeconomic shocks such as the Great Moderation and the oil price shocks of the 1970s. Furthermore, we show that the predictive ability of the staff outperforms that of private forecasters in the presence of specific factors, such as an increased uncertainty in the economy and the staff's better knowledge of the Fed's future interest rate