Monetary Policy Performance and the Accuracy of Observations

This paper evaluates the consequences for monetary policy performance of acquiring more accurate real time data. A forward looking model is set up and calibrated to fit the broad stylized facts of the U.S. economy. Two different assumptions about the information structure of the economy are made. Under the first, both policy makers and the public cannot observe potential output, but have to estimate it by applying the Kalman filter to noisy observations. Under the second structure, the public knows the true state of the economy, while the policy makers still have to estimate it. Evaluation of a standard loss function gives the counterintuitive result that less accuracy in real time data can lead to small, but positive, changes in welfare.