Annual Conference
|
International Macroeconomics, Money & Banking
|
May 2017
More than Words: A Textual Analysis of Monetary Policy Communication
This paper employs various tools from computational linguistics to monetary policy statements to gain exploratory insights into the nature of central bank communication. The sample was taken from a wide array of central banks, covering major central banks and others under the inflation-targeting (IT) regime, from 2000 to 2015. Three major aspects of communication were examined in this study, namely (i) readability – the ease with which a reader can understand a written text, (ii) topics – the key themes that are discussed in the policy statements, and (iii) tones – how positive/negative the outlook is in the central bank’s language assessment. We find that understanding central bank language generally required an advanced reader, although the statements varied substantially across banks in terms of both syntax and structure. The use of academic and unfamiliar words was among main contributors to such complexity. By employing hierarchical clustering analysis to objectively construct the structure of central bank rhetoric, monetary policy deliberation was indeed a complicated decision making in which the authorities must strike a balance between a fairly large number of critical issues. Our topic modelling analysis reveals that IT central banks had generally communicated more about inflation topic than economic growth topic. But discussions of global economic developments become more prominent in policy deliberation since the breakout of the global financial crisis in 2008. We also find that tone matters; when the net tone is informative about economic outlook, it significantly relates to change in policy rate decision. Lastly, our econometric evidence indicates that central bank communication complements well with the policy interest rate in enhancing the effectiveness of monetary transmission mechanism
Keywords:
monetary policy, Communication, Computational linguistics