Emotion in Euro Area Monetary Policy Communication and Bond Yields: The Draghi Era CAMA Working Paper 25/2025 April 2025

Researchers combine speech emotion recognition and natural language processing with high-frequency financial data to demonstrate how ECB press conferences' vocal, verbal cues impact euro area bond yields and equity prices.

key_findings bullet 1 · key_findings · validation V0

Unexpectedly, positive vocal signals increase yields for Germany, France, and Spain, while negative cues markedly raise Italian yields, reflecting country-specific investor reactions that alter risk-free rates and market risk premiums.

key_findings bullet 2 · key_findings · validation V0

Authors introduce a synchronized dataset merging voice and text with deep learning, proposing VoiceAN and LanguageAN indicators, though audio noise and sentiment biases significantly constrain this innovative monetary policy analysis.

key_findings bullet 3 · key_findings · validation V0

This paper innovatively applies Speech Emotion Recognition and NLP to assess central bank communications impact on bond yields, offering novel insights for fixed income and quantitative finance research. Its distinctive methodology integrates vocal and verbal cues with asset pricing, creating perspectives and avenues in computational financial and bond market analysis.

key_findings bullet 4 · key_findings · validation V0