The ECB Press Conference Statement: Deriving a New Sentiment Indicator for the Euro Area

Kanelis, Dimitrios; Siklos, Pierre L.

Abstract

We analyse the introductory statements of the ECB president and derive new sentiment indicators for the euro area based on a novel approach. To evaluate sentiment, we utilize a Large Language Model, namely FinBERT, which classifies the verbal sentiment of economics and finance-related textual data. We find that the ECB's conveyed sentiment about monetary policy, which is influenced by the economic outlook and the state of the euro area macroeconomy as expressed in speeches, plays a significant role in shaping the content of press conferences following a governing council decision. In contrast, speech sentiment regarding financial stability does not significantly influence introductory statements.