Historical Simulation Value at Risk (VaR)

Implementation of a Historical Simulation Value at Risk (VaR) framework in Python, including Expected Shortfall, portfolio P&L estimation, backtesting analysis, and Basel Traffic Light classification.

Overview

This project implements an end-to-end Historical Simulation Value at Risk (VaR) framework using Python.
The model estimates portfolio risk using empirical return distributions and evaluates model performance through backtesting and Basel Traffic Light classification.

Key Features

• Historical Simulation VaR (99%)
• Expected Shortfall (ES)
• Portfolio-level P&L computation
• VaR backtesting and breach analysis
• Basel Traffic Light regulatory classification
• Risk visualisation (time-series and distribution-based)

Methodology

1. Load historical market price data
2. Compute log returns
3. Construct a portfolio using predefined weights
4. Generate daily P&L
5. Estimate VaR and ES using rolling historical windows
6. Perform backtesting to identify VaR breaches
7. Classify model performance using the Basel Traffic Light framework

Project Structure

• data/ - market price data
• src/ - risk engine modules
• notebooks/ - execution and analysis
• requirements.txt - dependencies

How to Run

1. Install dependencies: pip install -r requirements.txt
2. Open the notebook notebooks/run_model.ipynb
3. Run cells sequentially to:

• compute VaR & ES
• perform backtesting
• visualize breaches
• generate Basel classification