Integrated Risk App (Python)

Validation-Grade Market & Credit Risk Engine

---

🎯 Project Objective (North Star)

This project is a model-risk–oriented risk analytics engine designed to measure, validate, and document market and credit risk models in a manner consistent with institutional risk management and model validation practices.

The objective is not to build a trading system or a dashboard-centric application, but to demonstrate:

• Sound quantitative risk methodology
• Explicit assumptions and loss conventions
• Clear separation between model logic, configuration, and presentation
• Standard validation and backtesting diagnostics

This repository functions as a model risk validation sandbox.

---

📌 Scope (What This Project Covers)

Market Risk

• Value-at-Risk (VaR) and Expected Shortfall (ES)
• Supported methodologies:
  • Historical Simulation
  • Parametric (Normal)
  • Monte Carlo (multivariate normal with covariance shrinkage)
  • Filtered Historical Simulation (GARCH-lite)
• Multi-confidence-level analysis (e.g. 95%, 99%)
• Rolling-window estimation
• Out-of-sample backtesting using Kupiec Proportion-of-Failures (POF)

Credit Risk

• Expected Loss (EL) framework:
  • Probability of Default (PD)
  • Loss Given Default (LGD)
  • Exposure at Default (EAD)
• Portfolio-level aggregation
• Segment-level loss decomposition

Stress & Scenario Analysis

• Deterministic equity shocks
• Interest-rate shocks via duration approximation
• Correlation stress
• Historical window replay

---

🧱 Repository Structure & Governance

This repository is intentionally structured to reflect institutional model governance and review practices.

risklib/
market/
market_risk_model.py     # MarketRiskModel + configuration objects
market.py                # Risk measures, simulation, backtesting primitives
credit/
credit_risk_model.py
app/
app.py                     # Streamlit viewer only (no modeling logic)
docs/
model_report.md            # Model methodology & validation notes

Design Principles

• risklib/ is the source of truth
  All modeling, estimation, and validation logic lives here.
• app/ is presentation-only
  The UI visualizes results but does not implement risk logic.
• Consistent loss-based convention
  All risk measures are reported as positive loss amounts.

---

🔍 Market Risk Model (Measurement & Validation)

Market risk is implemented as a validation-grade model object, separating configuration, estimation, and diagnostics.

from risklib.market.market_risk_model import MarketRiskModel, MarketRiskConfig

cfg = MarketRiskConfig(
    alpha=0.99,
    method="historical",     # historical | parametric | monte_carlo | fhs
    horizon_days=1,
    exposure=1_000_000,
)

model = MarketRiskModel(returns, weights, cfg)
model.fit()

var = model.compute_var()
es  = model.compute_es()
summary = model.summary()

This design mirrors institutional practice:

• Explicit configuration objects capture modeling assumptions
• Model objects encapsulate estimation logic
• Outputs include metadata required for validation review
• Backtesting is conducted separately and out-of-sample

---

📈 Market Risk Validation

Backtesting Framework

• Rolling 1-day VaR backtests
• Trailing-window estimation
• Exception tracking
• Kupiec Proportion-of-Failures (POF) test for unconditional coverage

Interpretation The Kupiec test evaluates whether the observed exception frequency is statistically consistent with the expected (1 − α) rate implied by the model.

---

🧠 Methodology Overview

• VaR: loss exceeded with probability (1 − α)

• Expected Shortfall: average loss conditional on VaR exceedance

• Filtered Historical Simulation:

  • Volatility estimated using fixed-parameter GARCH(1,1)
  • Returns standardized and rescaled

• Expected Loss (Credit): EL = PD × LGD × EAD

All models follow a consistent loss-based convention.

---

🧪 Validation & Testing Philosophy

This project emphasizes model validation, not just model output.

• Unit tests verify:

  • VaR monotonicity across confidence levels
  • ES ≥ VaR under consistent loss conventions
  • Correct exception counting in backtests
  • Credit loss aggregation consistency

• Backtesting results are interpretable and reproducible

• Assumptions and limitations are explicitly documented

---

🚫 Out of Scope (By Design)

To preserve clarity and rigor, this project does not attempt to be:

• A trading or portfolio optimization system
• A real-time production risk engine
• A regulatory-approved model

Advanced extensions (e.g. factor models, ALM, optimization) are intentionally deferred.

---

🖥 Application Interface

A lightweight Streamlit interface is provided strictly as a viewer to:

• Run models with selected parameters
• Visualize backtests and stress results
• Export validation series

All modeling logic remains outside the UI.

---

⚙️ Tech Stack

• Python
• Pandas / NumPy
• SciPy / Statsmodels
• Plotly
• Streamlit
• yfinance (demo data only)

---

⚡ Quickstart

git clone https://github.com/sensor-aae/Integrated-Risk-App.git
cd Integrated-Risk-App

python -m venv .venv
source .venv/bin/activate   # Windows: .venv\Scripts\activate

pip install -r requirements.txt
streamlit run app/app.py

---

📄 Documentation

A Model Risk Report (methodology, assumptions, validation results, and limitations) is maintained in the docs/ directory and mirrors institutional model documentation standards.

---

⚠️ Disclaimer

This project is for educational and demonstrative purposes only. It is not intended for production use or investment decision-making.

---

⭐ Why This Project Exists

This repository is designed as a work sample for roles in:

• Market Risk
• Model Risk / Model Validation
• Credit Risk
• Pension & Institutional Investment Risk
• Risk Consulting

It reflects how quantitative risk models are built, tested, challenged, and reviewed — not just how they are computed.