This paper investigates the nature of nonlinearities in the monetary policy rule of the US Federal Reserve (Fed) using the flexible approach to nonlinear inference. We find that while there is significant evidence of nonlinearity for the period to 1979, there is little such evidence for the subsequent period. Possible asymmetries in the Fed's reactions to inflation deviations from target and the output gap in the 1960s and 1970s may tell part of the story, but do not capture the entire nature of the nonlinearity. The inclusion of the interaction between inflation deviations and the output gap, as recently proposed, appears to characterize the nonlinear policy rule more adequately. Copyright © 2005 John Wiley & Sons, Ltd.