Mind Your Language: Market Responses to Central Bank Speeches

RePEc Short ID: pmc143
ORCID iD: 0000-0002-7220-4446

Ahrens, Maximilian; Erdemlioglu, Deniz; McMahon, Michael; Neely, Christopher J.; Yang, Xiye

doi:10.20955/wp.2023.013

Working Paper Revision

Mind Your Language: Market Responses to Central Bank Speeches

Abstract: Post-meeting central bank communication often moves markets, but researchers have paid less attention to the more frequent central bankers’ speeches. We create a novel dataset of U.S. Federal Reserve speeches and develop supervised multimodal natural language processing methods to identify how monetary policy news affect bond and stock market volatility and tail risk through implied changes in forecasts of GDP, inflation, and unemployment. We find that forecast revisions derived from FOMC member speeches can help explain volatility and tail risk in both equity and bond markets. Speeches from Chairs tend to produce larger forecast revisions and unconditionally raise volatility and tail risk. There is some evidence that a speaker’s monetary policy views, i.e, hawkishness vs. dovishness, may affect the impact of implied forecast revisions after conditioning on GDP growth. We show that central bank communication may calm markets, depending on the signals conveyed.

Keywords: central bank communication; multimodal machine learning; natural language processing; speech analysis; high-frequency data; volatility; tail risk;

JEL Classification: E52; C45; C53; G12; G14;

https://doi.org/10.20955/wp.2023.013