The informative value of central banks talks: a topic model application to sentiment analysis

Abstract

<abstract><p>Central banks communication has lately become an important tool to guide expectations and its impact on the economy has been acknowledged by the literature. Nowadays central banks speeches face an increasing variety of topics, which are not discriminated by text analysis. In this paper we build a topic-weighted central bank sentiment index as a combination of machine learning and text analysis techniques to investigate large datasets. First, we develop a methodological framework to grid search the best Latent Dirichlet Allocation (LDA) model to uncover the latent topics in central banks' speeches and releases published between 2000 and 2021. Then, we build a topic-specific sentiment index based on dictionary techniques. Next, we summarise the results in a topic-weighted Central Bank Sentiment Index (CBSIw) for the Bank of Canada (BoC), the Bank of England (BoE), the European Central Bank (ECB) and the Federal Reserve (Fed). We find that the main common driver of the CBSIw is the monetary policy topic, followed by macroprudential policy and payments and settlements. We also uncover bank-specific topics and topics related to new challenges, for example innovation and climate change. Moreover, we find that the CBSIw decreases after the Great Recession, signalling a worsening in sentiment, as well as during the COVID-19 crisis. Finally, we employ a probit regression to further assess the predictive power of our monetary policy topic-specific index. We find that the indicator helps predicting future changes in policy rate, corroborating the evidence that central banks communication signals future monetary policy decisions.</p></abstract>

The Predictive Power of Central Bank Communication: Evidence from Mexico

“…Moreover, the analysis revealed that a rate hike (cut) and a hawkish (dovish) monetary policy statement are associated with higher odds of observing a restrictive (accommodative) decision at Banxico's next meeting. These results are consistent with recent research empirically testing the same hypothesis in other cbs (see Astuti et al, 2022;and Priola et al, 2022).…”

“…These findings are in line with other relevant papers (see Picault & Renault, 2017;Baranowski et al, 2021;Priola et al, 2022;and Astuti et al, 2022) that evaluate the predict effects of cb communication on future policy decisions. This study contributes to the existing literature by focusing on these aspects for the Mexican cb, utilizing recent nlp techniques outlined in the literature.…”

“…The dictionaries categorize the words used by the cbs as either hawkish or dovish with the goal of capturing the policy bias of central bankers, thus, they allow to assign a numerical value and create a measurement that accurately reflects this bias. Using this type of indicators, there is growing evidence for different cbs that support the idea that communication can help anticipate future interest rate decisions (Tobback et al, 2017;Baranowski et al, 2021;Astuti et al, 2022;and Priola et al, 2022).…”

“…We estimated the predictive power of cb communication following a similar methodology as in Tobback et al (2017), Baranowski et al (2021), Astuti et al (2022), andPriola et al (2022). First, using a Latent Dirichlet Allocation (lda) model, we derived a series of latent topics that were the focus of Banxico's language between 2008 and 2022.…”

“…Cermeño & Navarrete (2011), Téllez-León & Venegas-Martínez (2013),Picault & Renault (2017),Baranowski et al (2021), andPriola et al (2022) apply a similar methodology, choosing only three values for the discrete change in the policy rate (cuts, holds and hikes). FollowingAstuti et al (2022), we employ various categories at different levels of analysis to accurately assess the influence of communication on specific monetary policy changes.…”