Inflation Trading - IOTA Monitor

End-to-end inflation basis trading framework: data pipeline, signal generation, backtesting, and interactive dashboard - built on live ECB data.

Live Dashboard

Launch the interactive dashboard

No installation required, it runs directly in your browser !!

Overview

This project models the IOTA spread : the basis between euro area ZC inflation swap rates and bond-implied breakeven inflation. The IOTA captures a persistent wedge between two instruments that theoretically price the same inflation expectations, but diverge due to liquidity premia, indexation lags, and technical flows.

$$IOTA_t = \text{Swap}_{proxy}(t) - BE_{bond}(t) = \text{Swap}_{proxy}(t) - (y_{nominal} - y_{real})$$

Trading logic: The IOTA is mean-reverting over medium-term horizons. A z-score above +2σ signals the swap market over-prices inflation relative to the bond market → sell basis. Below -2σ → buy basis.

Architecture

inflation-trading/
├── notebooks/
│   └── iota_trading_strategy.ipynb   # Full research pipeline
│   └── zc-inflation-swap-model.ipynb # Enhanced zcis curve model
├── src/
│   ├── data_loader.py               # ECB API : nominal, real, HICP
│   ├── basis.py                     # Swap proxy, IOTA, z-score
│   └── backtest.py                  # Backtest engine & metrics
├── dashboard/
│   └── inflation-dashboard.py       # Streamlit interactive monitor
├── data/                            # Generated charts & outputs
├── requirements.txt
└── README.md

---

Methodology

Data sources (ECB Statistical Data Warehouse)

Series	Dataflow	Frequency	Description
Nominal 10Y	YC/B.U2.EUR.4F.G_N_A.SV_C_YM.SR_10Y	Daily	Euro area AAA govt bonds (Svensson)
Real 10Y	FM/M.U2.EUR.4F.BB.R_U2_10Y.YLDA	Monthly → interpolated	Inflation-linked benchmark
HICP YoY	ICP/M.U2.N.000000.4.ANR	Monthly → interpolated	Euro area realized inflation

Breakeven construction

$$BE_{bond} = y_{nominal} - y_{real}$$

Swap proxy

Since euro area ZC swap rates are not publicly available, we construct a synthetic proxy blending realized inflation momentum and the ECB long-term anchor:

$$\text{Swap}_{proxy}(t) = (1-\lambda) \cdot \bar{\pi}_{252d}(t) + \lambda \cdot \pi^*_{ECB}$$

with $\lambda = 0.6$ and $\pi^*_{ECB} = 2%$.

Signal

$$z_t = \frac{IOTA_t - \mu_{252d}}{\sigma_{252d}}$$

Z-score	Signal	Action
$z > +2\sigma$	SHORT basis	Sell swap, buy breakeven
$z < -2\sigma$	LONG basis	Buy swap, sell breakeven
$|z| < 0.5\sigma$	EXIT	Close position

---

Notebook vs Dashboard : Two Complementary Approaches

This project is split into two distinct layers that serve different purposes and use data differently.

Research Notebook (iota_trading_strategy.ipynb)

The notebook is the full research pipeline : it documents every methodological choice, assumption, and limitation in detail. It is designed to be read like a quantitative research paper.

What it covers:

• Raw data fetching, cleaning and anomaly detection (525 aberrant observations identified and removed from the ECB real yield series over 2021–2023)
• Step-by-step construction of the breakeven, swap proxy and IOTA
• Stationarity analysis and regime identification (4 distinct macro regimes identified over 2004–2025)
• Full parameter optimization across 27 configurations (window × entry × exit)
• Regime-by-regime P&L attribution
• Full diagnostic dashboard: rolling Sharpe, P&L distribution, z-score vs next-day P&L scatter (R² = 0.000)
• Critical analysis of model limitations

Data used: full historical series 2004–2025, static, loaded once.

Interactive Dashboard (dashboard/inflation-dashboard.py)

The dashboard is a live monitoring and simulation tool : it fetches fresh data from the ECB API at each session (cached for 1 hour) and focuses on actionability rather than research depth.

What it covers:

• Live Monitor: current breakeven, IOTA level, z-score and active signal updated with latest ECB data
• Backtest tab: fully interactive : the user can adjust the rolling window (60–504 days), entry threshold (1–3σ) and exit band (0–1σ) and see the P&L and drawdown recompute in real time
• Stress Test tab: scenario analysis : simulates the impact of a ±100 bps breakeven shock on the IOTA level, z-score and signal across 11 shock scenarios

Data used: live ECB API calls, refreshed every session.

Key difference: the notebook uses cleaned, anomaly-filtered data with documented methodological choices. The dashboard uses raw interpolated data for real-time responsiveness : a deliberate trade-off between research rigour and operational usability.

Results

Metric	Value
Period	2005 – 2025
Best Sharpe (optimized)	0.15
Total P&L	+8 bps/year
Max Drawdown	-420 bps
Nb Trades	29

The strategy is regime-dependent: it performs well during smooth mean-reversion environments (2010–2019, post-2023) but suffers during acute stress episodes (GFC 2008, COVID 2020) where the basis diverges persistently before snapping back.

The Sharpe of 0.15 reflects the fundamental constraint of this study: the unavailability of public ZC swap quotes. With real-time swap rates (Bloomberg/Tullett), the signal would be sharper and the basis more precisely measured. This model is best interpreted as a directional framework rather than a production trading strategy.

ZCIS Curve Notebook (zc-inflation-swap-model.ipynb)

Focuses on a more complex curve construction model from CPI projections:

• seasonal decomposition and forward CPI path
• annualized ZCIS proxy curve
• term‑structure diagnostics and limitations

Data used: HICP NSA monthly series.

Getting Started

git clone https://github.com/mb69-code/inflation-trading
cd inflation-trading
pip install -r requirements.txt

# Run the dashboard locally
streamlit run dashboard/inflation-dashboard.py

# Or explore the full research notebooks
jupyter notebook notebooks/iota_trading_strategy.ipynb
jupyter notebook notebooks/zc-inflation-swap.ipynb

---

References

• Kerkhof, J. (2005). Inflation Derivatives Explained. Lehman Brothers Fixed Income. (https://the.earth.li/~jon/junk/kerkhof.pdf)
• Schulz, A. & Stapf, J. (2009). Price discovery on traded inflation expectations. BIS.
• ECB Statistical Data Warehouse : data-api.ecb.europa.eu
• Barclays Paper : Global Inflation‑Linked Products: A User’s Guide (https://www.docdroid.net/PsAYm5Z/barclays-global-inflation-linked-products-a-users-guide-pdf)
• Wanningen, C.F.A.R. (2007) : Inflation Derivatives, Blue Sky Group (Thesis : https://essay.utwente.nl/fileshare/file/58033/scriptie_C_Wanningen.pdf)