# 🎓 Three Essays on High-Frequency Return and Volatility Dynamics in Commodities and Financial Futures Markets
-orange?style=flat-square)
> **Doctoral thesis by articles (thèse par articles)** — Simon-Pierre Boucher,
> Faculté des sciences de l'administration (FSA), Université Laval.
> Co-authors: Marie-Hélène Gagnon & Gabriel J. Power.
---
## 🗂️ Repository Contents
| 📁 Folder | 📄 Content | 🏷️ Status |
|---|---|---|
| [`phd_chap1_20260731/`](phd_chap1_20260731/) | 🛢️ **Chapter 1** — Speculative trading in energy markets | 🟢 Revised for *The Energy Journal* |
| [`phd_chap2_20260731/`](phd_chap2_20260731/) | 📈 **Chapter 2** — iNAV & commodity volatility transmission | 🟡 Submission version, *Journal of Futures Markets* |
| [`PHD_chapitre3_theses_20260731/`](PHD_chapitre3_theses_20260731/) | 🏦 **Chapter 3** — FOMC tone & novelty (NLP) | 🟡 Manuscript (2026-03-06) |
| [`these-ulaval/`](these-ulaval/) | 📕 **Global thesis** — full ULaval document (`ulthese`/memoir) | 🟢 Compiles clean, 188 pp. |
---
## 🛢️ Chapter 1 — Speculative Trading in Energy Markets: Evidence from Macroeconomic Surprises
**Question:** Does speculative trading amplify or dampen the impact of macroeconomic news on commodity futures?
Using high-frequency data and 26 macroeconomic announcement releases, this essay measures how the intensity of speculation (an NLS proxy built from the CFTC disaggregated Commitments of Traders — money managers vs. swap dealers) conditions the reaction of energy (crude oil, natural gas) and metals (gold, silver, copper, palladium) futures to standardized macro surprises.
**Key findings** 🎯
- 🧯 Increased speculative trading **dampens** the impact of macro surprises on price drift, volatility, and bid-ask spreads.
- 💧 Speculators **improve liquidity and price discovery** while reducing volatility.
- ⚡ The damping effect is stronger for **procyclical commodities** (oil, gas) than for safe havens (gold).
- 👤 The beneficial effects are driven by **money managers**, not swap dealers.
📦 *In the folder:* `main.tex` (monolithic article), `tables.tex` (14 tables), `figures.tex` (6 figures), COVID & ZLB appendices, `master.bib`.
---
## 📈 Chapter 2 — Seeing Through the ETF: Indicative NAV and Commodity Volatility Transmission
**Question:** How does volatility flow between commodity ETFs and their underlying assets — and what does the intraday **indicative NAV (iNAV)** reveal that daily data cannot?
This essay builds a **novel minute-level iNAV dataset** for four single-commodity ETFs (gold, silver, oil, natural gas), decomposes realized variance into continuous and jump components (Barndorff-Nielsen–Shephard), and estimates HAR-X / HAR-CJ-X models at 1, 5 and 30 minutes plus a Bayesian VAR.
**Key findings** 🎯
- 🪞 The iNAV gives a **sharper image** of the ETF–underlying volatility relationship.
- 💥 Transmission runs primarily through **jumps**, not diffusion.
- ⏱️ **Sampling frequency matters**: 1-minute estimates are up to **2× larger** than 30-minute ones.
- 🥇 Precious metals: **unidirectional** (iNAV → ETF, passive arbitrage); ⛽ energy: **bidirectional & asymmetric**.
📦 *In the folder:* `main.tex` + `sections/` (intro, data, methods, results, conclusion), 13 tables, 8 figures (RV & IRF plots), `master.bib`, compiled `main.pdf`.
---
## 🏦 Chapter 3 — Returns and Volatility Around FOMC Announcements: A High-Frequency Analysis of Policy Tone and Novelty
%20%C2%B7%20Jord%C3%A0%20LPs%20%C2%B7%20panels-9cf?style=flat-square)
**Question:** When the Fed speaks, what moves markets — **what** it says (tone) or **how new** it is (novelty)?
FOMC statements (217 releases, 2000–2025) are decomposed into **policy tone** (hawkish/dovish) and **informational novelty** using a dual-model NLP ensemble (MiniLM + BERT, TSDAE+MNRL fine-tuning, PCA-based reference selection), then linked to 1-minute futures data via event regressions, minute-level panels and Jordà local projections, with placebo tests and five inference methods.
**Key findings** 🎯
- 🗣️ **Tone predicts directional returns**: a 1σ dovish shift → equity gains building to ≈ **+12 bps within two hours**.
- 🆕 **Novelty predicts volatility**: the stance × novelty interaction on VIX persists 5–120 min (*t* = −5.06).
- 📊 Stance moves realized volatility in **6 of 7 contracts** (*p* < 0.01).
- 🧪 Pre-announcement placebos are null — effects are announcement-driven.
📦 *In the folder:* `chapitre3.tex` + `sections/`, 24 tables, 19 figures, mathematical proofs appendix, `master.bib`, compiled `chapitre3.pdf`.
---
## 📕 Global Thesis — `these-ulaval/`
-008080?style=flat-square)
The full **thesis-by-articles document** assembled for the FESP (Faculté des études supérieures et postdoctorales) requirements:
- 🇫🇷 French front matter (résumé, remerciements, avant-propos) + 🇬🇧 English abstracts per chapter
- 📖 General introduction · 3 chapters · general conclusion · appendices A–B (Chapter 3 proofs & extras)
- 📚 Consolidated bibliography (`bib/these.bib`, 271 unique keys merged from the three articles)
- 🧾 [`INVENTAIRE.md`](these-ulaval/INVENTAIRE.md) — exhaustive audit of the sources, conflicts and every mechanical fix (labels prefixed `chN:`, package conflicts, BibTeX dedup…)
```bash
cd these-ulaval
latexmk # → main.pdf (0 errors, 0 undefined refs, 0 missing citations)
```
---
## 🧭 Thread of the Thesis
The three essays share one lens: **high-frequency data around information events**.
```
🛢️ Ch. 1 📈 Ch. 2 🏦 Ch. 3
macro announcements → ETF ↔ underlying → FOMC statements
× speculation volatility tone × novelty
(who trades matters) transmission (what & how new)
└──────────── returns · volatility · liquidity ────────────┘
```
---
## ✍️ Authors
| | |
|---|---|
| **Simon-Pierre Boucher** | PhD candidate in Finance, Université Laval — [simon-pierre.boucher.1@ulaval.ca](mailto:simon-pierre.boucher.1@ulaval.ca) |
| **Marie-Hélène Gagnon** | Professor of Finance, CRREP, Université Laval |
| **Gabriel J. Power** | IG Wealth Management Chairholder, Professor of Finance, CRREP & CRIB, Université Laval |
> 🙏 Funding: Social Sciences and Humanities Research Council (SSHRC) & Chaire Industrielle-Alliance Groupe financier.
---
📌 Folder suffix 20260731 = frozen snapshot of each article (July 31, 2026 versions). Sources are never edited in place — all thesis adaptations live in these-ulaval/.