Quant Risk Parity

A risk parity portfolio construction and backtesting framework implementing both the ERC (Equal Risk Contribution) and S&P 3-step methodologies with target volatility, realistic cost modeling, and robustness testing.

Features

• Two weight methods: ERC iterative optimizer and S&P 3-step inverse-vol method
• Target volatility: Dynamic leverage to hit 10% (configurable) annualized vol
• Realistic costs: Transaction costs, slippage, and leverage financing costs
• Two universes: Original 8-ETF and expanded 14-ETF (mapping to paper's 26 futures)
• 9-panel dashboard: Equity curve, drawdowns, rolling Sharpe, weights, risk contribution, turnover, returns distribution, rolling vol
• IS/OOS evaluation: In-sample / out-of-sample performance split
• Robustness testing: Regime stress tests, parameter sensitivity, bootstrap CIs (in analysis/)

Quickstart

pip install -e .
qrp fetch --flag all
qrp run
qrp report

Weight Methods

ERC (Equal Risk Contribution)

Iterative optimizer that finds weights where each asset contributes equally to portfolio risk. Applied via the covariance matrix (EWMA or rolling window).

S&P 3-Step (from the paper)

Based on "Indexing Risk Parity Strategies" (S&P Dow Jones Indices, 2018):

1. Inverse-vol weight each instrument to target vol
2. Asset-class multiplier to equalize risk across equity, fixed income, commodities
3. Portfolio multiplier to hit overall target volatility

Asset Universes

Universe	Tickers	Description
Original (8)	SPY, IEFA, EEM, IEF, TLT, LQD, DBC, GLD	Default
Expanded (14)	SPY, FEZ, EWJ, IEI, IEF, TLT, BWX, USO, UNG, UGA, DBC, GLD, SLV, DBA	Maps to paper's 26 futures

Project Structure

src/qrp/
├── __init__.py
├── paths.py          # Centralized file paths and constants
├── config.py         # Pydantic configuration with defaults
├── data.py           # Data fetching (Stooq) and caching (parquet)
├── weights.py        # ERC optimizer + S&P 3-step + risk contributions
├── backtest.py       # Portfolio simulation with cost modeling
├── metrics.py        # Sharpe, Sortino, Calmar, max drawdown, IS/OOS
├── visualize.py      # 9-panel dashboard generator
└── cli.py            # Typer CLI (fetch, run, report)

analysis/             # Standalone research scripts from paper replication
├── risk_parity_backtest_v2.py      # Cost modeling + variant comparison
├── risk_parity_v3_robustness.py    # Regime stress tests + sensitivity + bootstrap
├── risk_parity_v5_expanded.py      # 14-ETF expanded universe
├── risk_parity_pie_charts.py       # Risk contribution visualization
└── risk_parity_paper_period.py     # Paper-period direct comparison

tests/
├── __init__.py
└── test_core.py      # Unit tests for weights, metrics

References

• Liu, B. et al. (2018). Indexing Risk Parity Strategies. S&P Dow Jones Indices.
• Dalio, R. et al. (2015). Our Thoughts about Risk Parity and All Weather. Bridgewater.
• Hurst, B. et al. (2010). Understanding Risk Parity. AQR Capital Management.