Type and Duration
Preproposal PhD-Thesis, since September 2023Coordinator
Liechtenstein Business SchoolDescription
Introduction and BackgroundThe Federal Reserve (Fed) significantly influences U.S. economic policy. Fed speeches, pivotal for market participants, provide insights into monetary policy, economic forecasts, and regulations. Over time, the Fed has shifted towards transparent communication to enhance policy effectiveness. Analyzing these speeches with machine learning (ML) and natural language processing (NLP) offers valuable insights into their impact on financial stability.
Research Objectives
This research aims to automate Fed speech classification and measure their effects on financial markets. Key questions include:
1. How can ML classify Fed speeches by relevance to financial stability?
2. What are the short-term and long-term market impacts of these speeches?
3. How can we distinguish between anticipated and actual speech effects?
Methodology Data Collection and Preparation
A comprehensive database of Fed speeches will be created, preprocessed, and enriched with metadata like date, speaker, and context.
Text Classification Using Machine Learning
Various ML and NLP techniques, including Bag-of-Words, TF-IDF, Word Embeddings, and Transformer Models (e.g., BERT, GPT-3), will classify the speeches.
Model Training and Evaluation
Using annotated data, models such as SVMs, random forests, and neural networks will be trained and evaluated with accuracy, precision, recall, and F1-score metrics.
Analysis of Impact on Financial Markets
The classified speeches' impact on stock prices, bond yields, and exchange rates will be analyzed using statistical and econometric methods, including event studies and time-series analysis. The goal is to separate anticipated market effects from actual impacts post-speech.
Relevant Literature and Theoretical Foundations
This research builds on the work of Jurafsky and Martin (2020) on NLP, Goldberg (2017) on neural networks, Blinder et al. (2008) on central bank communication, and Hansen and McMahon (2016) on transparency.
Expected Outcomes and Implications
The classification method is expected to differentiate speeches by their relevance to financial stability reliably. Analyzing speech impacts will provide insights into monetary policy communication. This research could enhance Fed communication strategies and contribute to financial stability. It will offer a real-time tool for market participants and policymakers, aiding in risk management and decision-making.
Conclusion
This interdisciplinary approach combines ML with economic analysis to study Fed speeches' impact on financial markets. The anticipated outcomes will advance research and practical applications in financial economics, enhancing central bank communication strategies.
