This paper explores the role of monetary policy in economic activity, emphasizing the need to specify and estimate behavioral relationships for policy. It argues that policy choices must evolve independently of economic conditions, and that the effects of monetary policy should be measured by the regular response of policy to the economy, rather than the reverse. The authors use identified vector autoregressions (VARs) to analyze the data, aiming to construct economically interpretable models with superior fit to the data. They find that most of the variation in output or prices since 1960 can be attributed to shifts in monetary policy, but that the size of these effects varies across different specifications. The paper also discusses the limitations of traditional econometric methods and the rational expectations critique, highlighting the importance of modeling policy as a random variable. The authors conclude that their approach, while recognizing the stochastic nature of policy variation, has limitations in extrapolating to extreme scenarios.This paper explores the role of monetary policy in economic activity, emphasizing the need to specify and estimate behavioral relationships for policy. It argues that policy choices must evolve independently of economic conditions, and that the effects of monetary policy should be measured by the regular response of policy to the economy, rather than the reverse. The authors use identified vector autoregressions (VARs) to analyze the data, aiming to construct economically interpretable models with superior fit to the data. They find that most of the variation in output or prices since 1960 can be attributed to shifts in monetary policy, but that the size of these effects varies across different specifications. The paper also discusses the limitations of traditional econometric methods and the rational expectations critique, highlighting the importance of modeling policy as a random variable. The authors conclude that their approach, while recognizing the stochastic nature of policy variation, has limitations in extrapolating to extreme scenarios.