Forecasting in central banks

A survey of central bank forecasting

In this paper, Binder and Sekkel provide an overview of the forecasting frameworks of the Federal Reserve (Fed), the European Central Bank (ECB), the Bank of England (BoE), and the Bank of Canada (BoC) and how they have evolved in recent years. Some of their conclusions:

• Central bank forecasts are generally of high quality, at times outperforming those of professional forecasters. However, central banks’ information advantage over private forecasters has varied over time.
• Central banks’ informational advantage over private sector forecasters appears to have weakened over time.
• Asymmetric (and time varying) loss functions can help reconcile apparent departures from rationality in central banks’ forecasts when assuming a quadratic loss function.
• The COVID-19 shock has led to large forecast errors. The authors argue that going forward “central bank models should allow for more state-dependent, asymmetric and accelerated inflation dynamics, such as nonlinear Phillips curves and time-varying volatility.”

Central Bank Forecasting: A Survey
Authors: Carola Conces Binder, Rodrigo Sekkel
From: Haverford College, Bank of Canada

The Fed’s recent bad forecasts

This is a critique of the Fed’s forecasting from September 2020 to December 2022 and its failure to forecast the sharp rise in inflation:

The Fed’s forecasting failures stemmed from a confluence of modeling and analytical errors, human and institutional errors, and a striking absence of risk management. The Fed’s FRB-US macro model failed to forecast the stimulative impact of expansive fiscal and monetary policies. The Fed’s ongoing reliance on the Phillips Curve and its perception that it could control inflation by credibly managing inflationary expectations contributed to poor forecasts. Human and institutional errors contributed to poor projections and misguided policies. The Fed’s perception of its credibility may have been overestimated based on the sustained low inflation of the prior expansion. The Fed used poor judgment in interpreting the data, and it tended to rely on the “best outcome” projection, with insufficient attention paid to alternative outcomes.

The Fed: Bad Forecasts and Misguided Monetary Policy
Author: Mickey D. Levy
From: Hoover Institution, Stanford University