Central bank and media sentiment on central bank digital currency: an international perspective
Summary
Focus
We examine the sentiments of central banks and the media regarding central bank digital currencies across 15 major global economies between 2016 and 2022.
Contribution
Leveraging large language models, we develop jurisdiction-level indices of central bank and media sentiment towards central bank digital currencies, based on central bank publications and news articles.
Findings
We reveal significant divergence between central bank and media sentiments, with notable variations over time and across jurisdictions. Analysing the interplay between these sentiments, we observe that central bank sentiment tends to have more influence on media sentiment than the reverse. Additionally, we identify substantial cross-border sentiment spillovers, where sentiment in leading economies shapes sentiment in other regions. We demonstrate that cryptocurrency and equity markets primarily respond to shifts in central bank sentiment.
Abstract
This paper examines the sentiments of central banks and the media regarding central bank digital currencies across 15 major global economies. Leveraging large language models, we develop jurisdiction-level central bank digital currency sentiment indices derived from central bank publications and news articles on a daily basis. Our findings reveal significant divergences between central bank and media sentiments, with notable variations over time and across jurisdictions. Analyzing the interplay between these sentiments, we observe that central bank sentiment tends to exert a stronger influence on media sentiment than the reverse. Additionally, we identify substantial cross-border sentiment spillovers, where sentiment in leading economies shapes sentiment in other regions. Through an event study approach, we demonstrate that cryptocurrency and equity markets primarily respond to shifts in central bank sentiments. Specifically, more positive central bank sentiments on central bank digital currency are associated with negative impacts on cryptocurrency market returns and the stock performance of banking and payment-related firms.
JEL classification: E58, G12, G18
Keywords: Central bank digital currency (CBDC), central bank communication, media sentiment, large language model (LLM), financial market