Several recent studies have reached quite different conclusions about which variable is the best indicator of the stance of monetary policy. These differences likely reflect varying assumptions about bank and Federal Reserve behavior. This paper takes a detailed and comprehensive look at the implementation of monetary policy and the identification of monetary policy shocks. The paper first outlines a general analytical model for studying and evaluating monetary policy procedures. The model is then used to estimate both the Fed's operational policy objectives and its intermediate objectives. The results can be summarized as follows: First, monetary policy shocks over the past several years have primarily affected the federal funds rate, even during periods when the Fed was reportedly targeting reserves. In addition, the paper finds a statistically-significant liquidity effect in all periods examined, although the effect is quite small. Finally, there is statistical evidence that suggests that the Fed's intermediate objectives have not been stable over time, and these differences appear to be economically important. Taken together, these results indicate that while monetary policy shocks can be uncovered by regressing the funds rate on appropriate variables in the Fed's information set, the reaction function should be estimated over subperiods rather than over the entire 1959-1993 period.