PhD Research: Narrative Dynamics in Financial Markets
Gabin Taibi | University of Twente & Bern University of Applied Sciences
Supervised by Prof. Dr. Joerg Osterrieder
"How do financial narratives influence volatility dynamics and regime changes in financial markets?"
This research bridges Robert Shiller's narrative economics with quantitative financial analysis, developing novel methods to detect, measure, and predict how narratives propagate through markets and influence price dynamics.
GabinTB/PhD-Narrative-Finance
Central repository for all PhD research code, notebooks, and publications
Recent Activity
Repository Structure
PhD-Narrative-Finance/ ├── content/ │ ├── notebooks/ # Jupyter notebooks │ ├── scripts/ # Python scripts │ └── publications/ # Paper drafts ├── data/ │ ├── processed/ # Clean datasets │ └── raw/ # Source data links ├── .env.example └── README.md
Title: Modeling Narrative Dynamics for Volatility Regime Detection in Financial Markets
Chapter 1: Systematic Literature Review
Under ReviewAI-enhanced systematic review mapping narrative concepts in financial research using PRISMA methodology.
Expected Contribution:
First AI-enhanced SLR comprehensively mapping narrative finance literature with 200+ papers analyzed.
Chapter 2: Key Financial Market Narratives
In ProgressNarrative detection methods: supervised embeddings, LLM tagging, and unsupervised clustering with BERTopic.
Expected Contribution:
Novel taxonomy of financial market narratives with detection methodology.
Chapter 3: Market Microstructure & Volatility
SubmittedHFT analysis using nanosecond-level Deutsche Borse data; realized, implied, and rough volatility estimation.
Expected Contribution:
Nanosecond-level volatility signatures during narrative events.
Chapter 4: Do Narratives Drive Markets?
In ProgressExtending Sadka et al. framework to evaluate narrative explanatory power for market movements.
Expected Contribution:
Quantified explanatory power of narratives vs. traditional factors.
Chapter 5: Narrative-Driven Structural Breaks
PlanningIntegrating textual and volatility features for regime detection using PELT and ML classifiers.
Expected Contribution:
Integrated narrative-volatility regime detection framework.
Chapter 6: Conclusion & Implications
PlanningSynthesis of findings and implications for risk management and regulatory policy.
Expected Contribution:
Practical implications for risk managers and policymakers.
Deutsche Borse HFT
Nanosecond-level market microstructure analysis using Eurex and Xetra data
Narrative Modeling
NLP and transformer-based narrative detection from central bank speeches
TOPol Framework
Transformer Narrative Polarity Fields for semantic shift detection
Bubble Detection
NFT/DeFi bubble detection with 48 visualizations (Lennart Baals)
Deutsche Borse HFT Data
Nanosecond-level Xetra/Eurex trading data via academic NDA
BIS Central Bank Speeches
Worldwide central bank speeches corpus (Gigando project)
FOMC Minutes (NER-tagged)
Preprocessed Federal Reserve communications with entity tags
RavenPack News Headlines
Financial news headlines for sentiment analysis
St. Louis Fed FRED
Macroeconomic indicators (CPI, GDP, Unemployment)
S&P 500 Volatility
Market volatility series via Yahoo Finance
Reproducibility Statement
All analysis code is version-controlled on GitHub. For reproducibility, each notebook includes:
- Random seed initialization
- Environment requirements (requirements.txt)
- Data preprocessing steps documented
- Model hyperparameters logged
ORCID
Stefan Schlamp
IndustryDeutsche Borse AG
Head of Quantitative Analytics - HFT data collaboration
Axel Gross-Klussmann
IndustryQuoniam Asset Management
Narrative modeling from financial news collaboration
Collaboration network: PhD (orange), PI (blue), Industry (green), Academic (purple)