Are Banking Supervisory Data Useful
for Macroeconomic Forecasts?

Banking and Policy Working Paper 1-02

Working paper 1-02 has been superseded by the paper available on
Contributions to Macroeconomics: Vol. 3: No. 1, Article 3.

Abstract

Some argue that central banks can improve monetary policy by including confidential supervisory assessments of banking organizations in their forecasts of inflation and unemployment. In this study we examine the extent to which forecasts of these variables would have been improved with the inclusion of supervisory data. We begin by reproducing the earlier results used to support the claim. We critically examine them and extend the analysis from in-sample to out-of-sample testing. Finally, we check the robustness of our findings by extending the analysis period, using a different methodology to determine the contribution to forecasts, and substituting a different measure of supervisory information. Our analysis does not support claims that confidential supervisory information would have improved forecasts of inflation. Confidential supervisory information improved forecasts of unemployment in some periods. It is unclear, however, if the frequency or level of improvement would have altered monetary policy in a nontrivial way.

All the authors are in the special studies and policy section of the Federal Reserve Bank of Minneapolis. Feldman is an assistant vice president, Kim is an economic analyst, Miller is a vice president, and Schmidt is a senior financial analyst. We thank Joe Peek, Eric Rosengren and Geoffrey Tootell for use of their data and Robert Avery, Michael Keane, Jose Lopez, Larry Ozanne, Julie Randall, Art Rolnick and Dick Todd for their comments. The views expressed are those of the authors and do not necessarily reflect the views of the Federal Reserve Bank of Minneapolis, the Board of Governors, or the Federal Reserve System.

The authors welcome your comments on this paper.