Home
Wednesday, 22 January, 2025
Master´s Thesis Defense Presentations
Master´s Thesis Defense Presentations
Room 3
Defense Committee: Jan Zápal (chair), Byeongju Jeong, Christian Ochsner
13:00 Bernhard Brunner: Predicting Policy Rate Changes from Central Bank Minutes Using Machine Learning: Evidence from the Czech Republic (1998-2024)
Chair: Sebastian Ottinger
Opponent: Paolo Zacchia
14:00 Klaus Hajdaraj: Causal Machine Learning for Heterogeneous Treatment Effects: An Empirical Application on Optimal Treatment Assignment
Chair: Paolo Zacchia
Opponent: Sebastian Ottinger
Bernhard Brunner
Abstract:
This thesis investigates how the sentiment of central bank minutes predicts upcoming policy rate changes in the Czech Republic from 1998 to 2024. I fine-tuned the RoBERTa language model on central bank communications to create a sentiment index of the Czech National Bank (CNB) Board meeting minutes. I then estimated an ordered probit model by regressing the upcoming policy rate change on this sentiment index while controlling for the current policy rate change and macroeconomic indicators. My results show that the sentiment of CNB minutes is a sizeable and statistically significant predictor of upcoming policy rate changes. Robustness checks using two alternative sentiment measures – a BERT sentiment index and RoBERTa sentiment dummies – and re-estimation of the model for subperiods further support my findings.
