Eric M. Leeper, Christopher A. Sims, and Tao Zha examine the effects of monetary policy on economic activity. They argue that monetary policy is not a single variable but rather a complex process involving interactions between policy, the banking system, and various economic variables. They use identified vector autoregressions (VARs) to analyze the data, which allow for the estimation of large time-series models and provide a more accurate measure of policy effects. The authors emphasize the importance of specifying and estimating behavioral relationships for policy, separating the regular response of policy to the economy from the response of the economy to policy. They challenge the assumption that models for policy analysis and forecasting are sharply distinct, arguing that economically interpretable models can have superior fit to the data. The authors analyze the data without imposing strong economic beliefs, using a single time frame and data set to check the robustness of results in the literature. They find that most specifications imply that only a modest portion of the variance in output or prices in the United States since 1960 can be attributed to shifts in monetary policy. They also point out substantive problems in models that imply large real effects on output or prices and argue that correcting these reduces the implied size of the real effects. A robust conclusion common across these models is that a large fraction of the variation in monetary policy instruments can be attributed to the systematic reaction of policy authorities to the state of the economy. This is what one would expect of good monetary policy, but it is also the reason why it is difficult to use the historical behavior of aggregate time series to uncover the effects of monetary policy. The authors use a class of models called identified VARs, which have recently become widely used. They describe the framework, summarize how it differs from other popular frameworks, and consider some common criticisms. They also discuss the ways in which they and others have put substantive meat on this abstract skeleton. The authors compare their approach with traditional simultaneous equations (SE) modeling and find that identified VARs differ in their requirements for identification. They also discuss the rational expectations critique of identified VAR modeling, which argues that disturbances are what is "omitted from the theory," and that therefore one cannot claim to know much about their properties. They argue that the identified VAR approach is not necessarily flawed, and that the assumptions on the model that justify doing so should be explicit from the beginning. The authors also compare their approach with the dynamic stochastic general equilibrium (DSGE) approach, which arose largely as a response to the rational expectations critique. They find that both approaches have their advantages and disadvantages, and that the inherent limitations of simulating non-stochastic shifts in policy rules are common to both DSGE and the newer SE-style models. The authors conclude that such strictly linear, weakly identified models do have limitations. They would not be comfortable extrapolating their estimates of policy effects to regimes of hyperinflation or to very different fiscal policy environments. However, they regard it as anEric M. Leeper, Christopher A. Sims, and Tao Zha examine the effects of monetary policy on economic activity. They argue that monetary policy is not a single variable but rather a complex process involving interactions between policy, the banking system, and various economic variables. They use identified vector autoregressions (VARs) to analyze the data, which allow for the estimation of large time-series models and provide a more accurate measure of policy effects. The authors emphasize the importance of specifying and estimating behavioral relationships for policy, separating the regular response of policy to the economy from the response of the economy to policy. They challenge the assumption that models for policy analysis and forecasting are sharply distinct, arguing that economically interpretable models can have superior fit to the data. The authors analyze the data without imposing strong economic beliefs, using a single time frame and data set to check the robustness of results in the literature. They find that most specifications imply that only a modest portion of the variance in output or prices in the United States since 1960 can be attributed to shifts in monetary policy. They also point out substantive problems in models that imply large real effects on output or prices and argue that correcting these reduces the implied size of the real effects. A robust conclusion common across these models is that a large fraction of the variation in monetary policy instruments can be attributed to the systematic reaction of policy authorities to the state of the economy. This is what one would expect of good monetary policy, but it is also the reason why it is difficult to use the historical behavior of aggregate time series to uncover the effects of monetary policy. The authors use a class of models called identified VARs, which have recently become widely used. They describe the framework, summarize how it differs from other popular frameworks, and consider some common criticisms. They also discuss the ways in which they and others have put substantive meat on this abstract skeleton. The authors compare their approach with traditional simultaneous equations (SE) modeling and find that identified VARs differ in their requirements for identification. They also discuss the rational expectations critique of identified VAR modeling, which argues that disturbances are what is "omitted from the theory," and that therefore one cannot claim to know much about their properties. They argue that the identified VAR approach is not necessarily flawed, and that the assumptions on the model that justify doing so should be explicit from the beginning. The authors also compare their approach with the dynamic stochastic general equilibrium (DSGE) approach, which arose largely as a response to the rational expectations critique. They find that both approaches have their advantages and disadvantages, and that the inherent limitations of simulating non-stochastic shifts in policy rules are common to both DSGE and the newer SE-style models. The authors conclude that such strictly linear, weakly identified models do have limitations. They would not be comfortable extrapolating their estimates of policy effects to regimes of hyperinflation or to very different fiscal policy environments. However, they regard it as an