Information in central bank sentiment: An analysis of Fed and ECB communication
Abstract
This paper constructs a sentiment measure for published central bank (CB) communications using natural language processing and FinBERT techniques (machine learning models adapted to analyze financial news). We analyze sentiment of various Fed and ECB communications, but focus on FOMC minutes and ECB press conferences for our main results. We find that Fed and ECB sentiment move together. Using local projection analysis, we find that CB sentiment predicts policy interest rates and the Taylor rule several quarters into the future. Fed sentiment also leads ECB rates. In turn, stock market returns and deflation expectations tend to lead Central Bank sentiment.
Fields of Study & Subjects
Monetary Policy and Economic Impact
Market Dynamics and Volatility
Media Influence and Politics
Publication details
- Journal / source
- Journal of International Financial Markets Institutions and Money
- Publisher
- Elsevier BV
- Volume
- 110
- Pages
- 102335–102335
- Publication date
- 14 May 2026
- Language
- EN