What’s that Noise? Analysing Sentiment-based Variation in Central Bank Communication

To which degree can variation in sentiment-based indicators of central bank communication be attributed to changes in macroeconomic, financial, and monetary variables; idiosyncratic speaker effects; sentiment persistence; and random ‘noise’? Using the Loughran and McDonald (2011) dictionary on a tex...

Description complète

Enregistré dans:
Détails bibliographiques
Publié dans:MAGKS - Joint Discussion Paper Series in Economics (Band 41-2022)
Auteurs principaux: Hayo, Bernd, Zahner, Johannes
Format: Article
Langue:anglais
Publié: Philipps-Universität Marburg 2022
Sujets:
Accès en ligne:Texte intégral en PDF
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
Description
Résumé:To which degree can variation in sentiment-based indicators of central bank communication be attributed to changes in macroeconomic, financial, and monetary variables; idiosyncratic speaker effects; sentiment persistence; and random ‘noise’? Using the Loughran and McDonald (2011) dictionary on a text corpus containing more than 10,000 speeches and press statements, we construct sentiment-based indicators for the ECB and the Fed. An analysis of variance (ANOVA) shows that sentiment is strongly persistent and influenced by speaker-specific effects. With about 80% of the variation in sentiment being due to noise, our findings cast doubt on the reliability of conclusions based on variation in dictionary-based indicators.
Description matérielle:12 Seiten
ISSN:1867-3678
DOI:10.17192/es2024.0748