Measuring Central Banks’ Sentiment and its Spillover Effects with a Network Approach

Abstract

Applying text analysis to speeches and press releases we construct a Sentiment Index (CBSI) of four central banks to (i) investigate spillovers generated by CBSIs on financial variables with GIRF; (ii) analyze the time-varying and statistically significant spillovers among CBs’ communication with a new a two-step network approach. We find that after the Great Recession the sentiment conveyed by CBs became more negative. Second, CBSIs influence financial variables with the right sign. Finally, CBSIs affect each other and these effects are time-varying: the Fed seems to be the most influential institution, especially in the aftermath of the Great Recession and the ECB appears to have increased its communication spillovers after the European Debt Crisis.