This paper presents an optimization-based econometric framework for evaluating monetary policy. The authors develop a structural model to assess the effects of proposed monetary policy rules. They argue that traditional econometric methods fail to account for the Lucas (1976) critique, which suggests that policy evaluations based on past data may not hold under different policy regimes. By deriving their model from intertemporal optimization, they ensure that the model reflects rational expectations and individual behavior, allowing for more accurate policy evaluation. The paper outlines five steps: first, they estimate a vector autoregression (VAR) model of interest rates, inflation, and output to identify the actual monetary policy rule used by the Federal Reserve. Second, they propose a theoretical model to explain the responses of output and inflation to monetary policy shocks. Third, they combine the structural model with the VAR to identify shocks to the structural equations. Fourth, they simulate the consequences of hypothetical monetary policy rules. Finally, they use the model to compute the welfare consequences of different monetary rules. The authors find that the Federal Reserve's monetary policy rule can be described by a feedback rule that sets interest rates as a function of current and past values of output and inflation. They estimate this rule and find that an increase in GDP relative to trend raises the funds rate, and an increase in inflation relative to its target level raises the funds rate by a greater amount. They also find that the short-run multiplier for output is larger than the long-run multiplier. The paper also presents a simple model of output and inflation determination, which incorporates decision lags and allows for a better fit with the predictions of the VAR. The model is based on an infinite-lived representative household that maximizes utility subject to a budget constraint. The model includes a Calvo-type price-setting mechanism, where prices are adjusted at exogenous random intervals. The authors estimate the parameters of the model so that the model's predictions regarding the effects of a monetary policy shock fit those estimated by the unrestricted VAR as closely as possible. They find that the structural parameters can be identified from the impulse responses, although some parameters cannot be uniquely identified. They calibrate two of the unidentified parameters based on independent evidence, allowing them to estimate the remaining parameters from the impulse responses. The calibrated parameters include the frequency of price changes and the elasticity of the marginal disutility of producing output with respect to an increase in output.This paper presents an optimization-based econometric framework for evaluating monetary policy. The authors develop a structural model to assess the effects of proposed monetary policy rules. They argue that traditional econometric methods fail to account for the Lucas (1976) critique, which suggests that policy evaluations based on past data may not hold under different policy regimes. By deriving their model from intertemporal optimization, they ensure that the model reflects rational expectations and individual behavior, allowing for more accurate policy evaluation. The paper outlines five steps: first, they estimate a vector autoregression (VAR) model of interest rates, inflation, and output to identify the actual monetary policy rule used by the Federal Reserve. Second, they propose a theoretical model to explain the responses of output and inflation to monetary policy shocks. Third, they combine the structural model with the VAR to identify shocks to the structural equations. Fourth, they simulate the consequences of hypothetical monetary policy rules. Finally, they use the model to compute the welfare consequences of different monetary rules. The authors find that the Federal Reserve's monetary policy rule can be described by a feedback rule that sets interest rates as a function of current and past values of output and inflation. They estimate this rule and find that an increase in GDP relative to trend raises the funds rate, and an increase in inflation relative to its target level raises the funds rate by a greater amount. They also find that the short-run multiplier for output is larger than the long-run multiplier. The paper also presents a simple model of output and inflation determination, which incorporates decision lags and allows for a better fit with the predictions of the VAR. The model is based on an infinite-lived representative household that maximizes utility subject to a budget constraint. The model includes a Calvo-type price-setting mechanism, where prices are adjusted at exogenous random intervals. The authors estimate the parameters of the model so that the model's predictions regarding the effects of a monetary policy shock fit those estimated by the unrestricted VAR as closely as possible. They find that the structural parameters can be identified from the impulse responses, although some parameters cannot be uniquely identified. They calibrate two of the unidentified parameters based on independent evidence, allowing them to estimate the remaining parameters from the impulse responses. The calibrated parameters include the frequency of price changes and the elasticity of the marginal disutility of producing output with respect to an increase in output.