The Differential Market Impact of ECB Narrative Tone Across Communication Channels

Author: Shreyas Urgunde
Institution: Warwick Business School
Supervisor: Philippe Mueller
Date: September 2025

📄 Paper:
👉 View the dissertation

---

📌 Overview

The objective of this thesis is to quantify how the tone of European Central Bank communications influences euro-area markets across policy statements, press conferences, Monetary Policy Accounts, and Executive Board speeches.

It primarily constitues:

• Channel-specific tone & Comparison: Using ECB-specific lexicons to build sentence-level hawk–dove scores and assess whether tone adds incremental information beyond policy news across all channels.
• Event timing & high-frequency identification: Intraday reactions are measured in tightly defined windows taken from the EA-MPD and EA-CED datasets to precisely time announcements, use high-frequency returns to isolate immediate market moves, and extract policy-news factors via PCA.
• Isolating pure tone effects: Separate narrative tone from contemporaneous policy surprises (Target/FG/QE) to recover the tone-only impact.
• Daily persistence: Building on intraday analysis using EA-MPD and EA-CED, I re-estimate on daily changes to test whether tone effects persist beyond the announcement window and survive broader market noise.
• State dependence: Interact tone with the ECB CISS to gauge how tone’s impact varies with financial stress.
• Robustness & reproducibility: Verify results with alternative tone dictionaries, orthogonalized (“policy-free”) tone, and subsample analysis (pre/post-COVID).

---

💡Key findings

Topic	Finding
Intraday drivers	Policy surprises dominate high-frequency reactions; tone effects are generally small and statistically weak.
Q&A equities	During the press-conference Q&A, equities show a modest positive response to tone; interest rates do not co-move with tone in this window.
Daily persistence	By the daily close, most tone effects attenuate or vanish; daily variation is largely absorbed by broad risk and idiosyncratic factors.
Accounts × CISS	In Monetary Policy Accounts, tone is positive at average stress but its impact dampens as systemic stress (CISS) rises.
Term-structure pattern	Loadings align with theoretical intuition: Target → short rates, Forward-Guidance → belly, QE → long end & spreads; equity/FX effects are smaller and more heterogeneous.
Bottom line	Narrative tone matters in specific contexts, but hard policy signals carry most of the market moves.

---

📁 Project Structure

• Financial Datasets/ — Curated market and rates datasets required to run the analysis (CSV/XLSX).
• Research Paper/ — Final paper (Dissertation.pdf) and LaTeX source (dissertation.tex).
• Dissertation.ipynb — Core analysis notebook with fully reproducible code (tables, figures, regressions).
• all_ECB_speeches.csv — ECB communications dataset (speeches only); other channels were web-scraped directly from the ECB website.

---

🔧 Replication guide

Requirements

• Python 3.10+
• Jupyter (or VS Code with the Python + Jupyter extensions)
• Key Packages: pandas, numpy, statsmodels, scikit-learn, matplotlib, nltk, regex, unidecode

Setup

1. Clone or download the repo. Keep the folder structure unchanged (the notebook expects relative paths like Financial Datasets/..., all_ECB_speeches.csv, etc).

   git clone https://github.com/shreyasxi/ecb-narrative-tone-market-impact-thesis.git
   cd ecb-narrative-tone-market-impact-thesis
   

2. Open Dissertation.ipynb (in Jupyter or VS Code).

3. Run All (VS Code: Run ▸ Run All) to:

   • load the ECB text + tone data,
   • build the daily asset panels and controls,
   • estimate baseline and CISS-interaction regressions,

Outputs: regression summaries, coefficient tables, and figures in results/ (also referenced in the paper).

---

📚 Citation

If you use this work, please cite the thesis as:

Urgunde, S. (2025). The Differential Market Impact of ECB Narrative Tone Across Communication Channels. Warwick Business School.
Available at: https://shreyasxi.github.io/ecb-narrative-tone-market-impact-thesis/

BibTeX

@mastersthesis{Urgunde2025_ECBTone,
  author       = {Shreyas Urgunde},
  title        = {The Differential Market Impact of ECB Narrative Tone Across Communication Channels},
  school       = {Warwick Business School, University of Warwick},
  year         = {2025},
  type         = {Master's Thesis},
  url          = {https://shreyasxi.github.io/ecb-narrative-tone-market-impact-thesis/}
}