"Recurring change in a monetary policy function that maps endogenous variables into policy choices alters both the nature and the efficacy of the Taylor principle--the proposition that central banks can stabilize the macroeconomy by raising their interest rate instrument more than one-for-one in response to higher inflation. A monetary policy process is a set of policy rules and a probability distribution over the rules. We derive restrictions on that process that satisfy a long-run Taylor principle and deliver unique equilibria in two standard models. A process can satisfy the Taylor principle in the long run, but deviate from it in the short run. The paper examines three empirically plausible processes to show that predictions of conventional models are sensitive to even small deviations from the assumption of constant-parameter policy rules."