Derivatives Model Validation
Saudi bank
The bank required an independent review of its derivative pricing models against market practice and applicable regulatory guidance, covering the concept behind each model, its testing, its implementation in the trading system and the monitoring that follows.
Scope of work
Conceptual soundness assessed for each model against its stated purpose and the bank's pricing and risk strategy:
- Theoretical foundations evaluated against alternative methodologies, with limitations and their implications set out
- Key assumptions and variable definitions checked against industry practice and market consistency
- Data sources, governance and calibration datasets checked for accuracy, completeness and alignment with local market characteristics
Calibration review covering calibration approach, parameter selection, data usage and model stability, to confirm robust performance across conditions.
Stability and sensitivity testing, validating accuracy and scalability under varied conditions, reviewing bias in data and development, and running stress scenarios to identify where each model responds sharply to input movement.
Implementation verification, checking that the build in the trading system aligns with the theoretical design and operational requirements, and identifying integration issues and process inefficiencies.
Back-testing and overrides, verifying model predictions against historical data, documenting recalibration processes, and assessing manual overrides for justification and impact with a view to reducing dependency on them.
Documentation and training, recording model purpose, structure, assumptions and data sources, with targeted training designed using the bank's own examples and a risk-based continuous monitoring framework established.
Outcome and value
- A documented, independently reviewed model set across the derivative book.
- Deficiencies identified with remediation and recalibration set out.
- Internal expertise built, so the bank can maintain and monitor the models itself.
- A monitoring framework sized to the bank's resources rather than to an ideal.