The intricacies of institutional communication - Text mining policy documents

The goal of the project is to systematically examine the impact of central bank communication in the Central East European region and to see whether central banks can affect inflation expectations and sovereign bond yields. The underlying data would come from central bank press releases and governor speeches and the main text mining tools would include both supervised and unsupervised machine learning approaches.

The other direction of the research would look at the IMF’s policy advices and their actual implementation. The analysis would use the complete original database of Article IV reports between 2004 and 2019 and the applied AI techniques would be similar as above.

The project would yield actionable policy insights regarding the communication of central banks and how the CEE region handles IMF policy advice.

Keywords: machine learning, monetary policy, communication, text mining