Hamza Bennani (University Paris Nanterre)

Titre: Does a Sentiment Shock Help to Predict Monetary Policy: Evidence from the ECB

Abstract: This paper aims to assess whether sentiment shock is helpful in predicting ECB monetary policy decisions. We use a bag-of-words approach and several dictionar- ies on ECB’s President introductory statements to derive a measure of sentiment. Next, we orthogonalize the sentiment measure on a set of macroeconomic and finan- cial variables to compute sentiment shock. Finally, we test whether sentiment shock is useful to predict ECB monetary policy decisions. We find that sentiment shock is significantly and positively related to future ECB policy decisions, even when controlling for future economic conditions and market expectations about mone- tary policy. Further extensions show that the predictive power of sentiment shock is robust to (i) the scoring method chosen to compute sentiment, (ii) alternative expectations about monetary policy and (iii) the macroeconomic forecasts used in the monetary policy Taylor rule. However, we find that the predictive power of sentiment shock is sensitive to the dictionaries used to compute the measure of sentiment. These findings highlight an additional channel by which ECB communication improves monetary policy predictability.