A monetary policy reaction function was developed from a theoretical model of Federal Reserve policy adoption. This reaction function was then estimated as a linear probability model and as a probit probability model over various subperiods from February 1953 to May 1985. The objectives of monetary policy were specified as forecasts of growth in output, price stability, unemployment, and international balance. These forecasts were generated from a vector autoregression model;The data were divided into subperiods according to three schemes in order to test three hypotheses. First, the data were divided into subperiods corresponding to the chairmenships of the Board of Governors to test the hypothesis that the chairman of the Fed significantly influences the objectives and priorities of monetary policy. Second, the data were divided into two subperiods: the pre- and postelection biennia, to test the hypothesis that the Fed's preelection reaction function differs from its postelection reaction function. Finally, the data were divided into two subperiods: 1974(1)-1979(9) and 1979(10)-1984(5) to test the hypothesis that the objectives and priorities of monetary policy changed when the Fed announced its new operating procedures;The results suggest that the chairman of the Fed may significantly influence the objectives and priorities of monetary policy. Furthermore, the Fed's preelection reaction function differs from its postelection reaction function which suggests that the "political business cycle" may, in part, be created by monetary policy. Finally, the results suggest that the objectives and priorities of monetary policy changed after the Fed announced a change in operating procedures.