Lombard Lending Risk Engine

A modular Python package for computing collateral haircuts, LTV limits, liquidity and concentration add‑ons, and stress‑scenario impacts for Lombard lending portfolios. The package provides a reproducible framework for evaluating credit exposure on portfolios pledged as collateral.

---

Features

• Portfolio returns aggregation
• EWMA volatility
• Parametric VaR‑based volatility add‑on
• Liquidity discounts based on ADV and position size
• Single‑name concentration add‑on
• Combined haircut model
• LTV and margin‑call evaluation
• Stress scenario utilities
• Centralized configuration parameters

---

Installation

Clone the repository and install in editable mode:

git clone https://github.com/yourusername/lombard-risk-engine.git
cd lombard-risk-engine
pip install -e .

Install development dependencies:

pip install -e .[test]

---

Package Structure

lombard_risk/
    config.py
    data.py
    collateral.py
    market_models.py
    liquidity.py
    concentration.py
    haircut.py
    ltv.py
    stress.py
    engine.py
scripts/
    run_lombard_eval.py
tests/
    test_engine.py

---

Basic Usage

1. Prepare returns and weights

import pandas as pd
import numpy as np

dates = pd.date_range("2020-01-01", periods=252)
returns_df = pd.DataFrame(
    np.random.normal(0, 0.01, size=(252, 3)),
    index=dates,
    columns=["A", "B", "C"],
)

weights = pd.Series([0.5, 0.3, 0.2], index=returns_df.columns)

2. Define liquidity profiles

from lombard_risk.liquidity import LiquidityProfile

liq_profiles = {
    "A": LiquidityProfile(isin="A", adv=1_000_000, position_value=300_000),
    "B": LiquidityProfile(isin="B", adv=500_000, position_value=400_000),
    "C": LiquidityProfile(isin="C", adv=200_000, position_value=300_000),
}

3. Run the engine

from lombard_risk.engine import LombardRiskEngine

engine = LombardRiskEngine()

result = engine.evaluate(
    portfolio_value=1_000_000,
    loan_amount=600_000,
    returns_df=returns_df,
    weights=weights,
    base_haircut=0.20,
    liq_profiles=liq_profiles,
)

print(result)

---

Configuration

All tunable parameters are defined in lombard_risk/config.py:

• VaR confidence level
• EWMA lambda
• Volatility buffer
• Liquidity thresholds and discounts
• Concentration threshold and penalty
• Margin‑call buffer
• Default base haircut

These values can be modified to match internal policy.

---

Testing

Run the test suite:

pytest -q

---

License

MIT License.