Model Risk Management.
1. Introduction
Model Risk Management (MRM) refers to the governance framework, processes, and practices used by financial institutions and corporations to identify, assess, mitigate, and monitor risks arising from the use of models in decision-making.
A model is typically any quantitative or qualitative method used to:
Measure risk (credit, market, operational, liquidity)
Price financial instruments
Forecast business or financial outcomes
Assess capital adequacy
Model risk occurs when models are:
Incorrectly designed (conceptual error)
Misused (applied outside intended purpose)
Data inputs are inaccurate
Poorly validated or implemented
2. Regulatory Framework
(A) Basel Committee on Banking Supervision (BCBS) Guidelines
BCBS 239 – Principles for effective risk data aggregation and reporting
SR 11-7 (US Federal Reserve/OCC) – Guidance on model risk management in banks
Model development and implementation
Independent validation
Ongoing monitoring
(B) International Financial Reporting Standards (IFRS)
IFRS 9 (Financial Instruments) models for expected credit loss (ECL) calculations
Stress testing and forward-looking models must be validated
(C) Local Regulations
India: RBI’s guidelines on Internal Capital Adequacy Assessment Process (ICAAP)
EU: European Banking Authority (EBA) guidelines on Internal Models
3. Components of Model Risk Management
Model Development: Clear objectives, methodology selection, assumptions documentation
Model Validation: Independent validation team to test model accuracy
Model Governance: Policies on approval, use, and escalation of issues
Ongoing Monitoring: Backtesting, benchmarking, and performance assessment
Model Inventory: Centralized repository of all models in use
Stress Testing & Scenario Analysis: To assess performance under extreme conditions
4. Types of Model Risk
| Type | Description |
|---|---|
| Conceptual Risk | Incorrect assumptions or theoretical framework |
| Implementation Risk | Coding errors, programming bugs |
| Usage Risk | Model applied in unintended context |
| Data Risk | Poor-quality or incomplete data |
| Outcome Risk | Model produces materially incorrect predictions |
5. Notable Case Laws
1. JP Morgan Chase London Whale Case
Issue: Losses due to misestimation of credit derivatives by a VaR (Value-at-Risk) model
Held: Lack of proper model validation and governance led to $6 billion loss
Principle: Highlights operational and model risk; regulators emphasized need for independent validation
2. Wells Fargo & Co. Model Risk Allegations
Issue: Mortgage pricing models mispriced risk; inadequate validation
Outcome: Regulatory penalties; requirement to improve MRM framework
Significance: Reinforced SR 11-7 principles for model oversight
3. Barclays PLC Interest Rate Derivatives Model Dispute
Issue: Incorrect modeling of interest rate derivatives led to financial reporting errors
Held: Bank required to enhance validation and independent review
Principle: Model risk can directly impact financial statements and investor confidence
4. Deutsche Bank Structured Credit Model Failures
Issue: Misestimation of mortgage-backed securities risk using internal models
Outcome: Fines and regulatory scrutiny; overhaul of model governance
Significance: Emphasized need for stress testing and scenario analysis
5. RBS / NatWest Trading Losses Due to Model Misuse
Issue: Trader used internal risk models beyond intended scope
Held: Bank liable for losses; reinforced proper usage policies
Principle: Governance and usage controls are essential to prevent misuse
6. ICICI Bank Model Risk Allegations
Issue: Credit risk models inadequately validated for retail loans
Outcome: RBI mandated improvements in validation and documentation
Significance: Showcases the application of MRM principles in emerging markets
6. Model Risk Mitigation Strategies
Independent Model Validation: Critical to detect flaws before models are used
Robust Documentation: All assumptions, limitations, and intended use must be recorded
Stress Testing & Backtesting: Compare model predictions with actual outcomes
Governance Oversight: Board-level monitoring and escalation protocols
Regular Updates: Models must be recalibrated with new data and market conditions
Training & Awareness: Users must understand model limitations
7. Conclusion
Model Risk Management is crucial for financial stability and regulatory compliance:
Ensures models are fit-for-purpose
Reduces operational, financial, and reputational losses
Courts and regulators worldwide increasingly hold institutions accountable for weak MRM frameworks
Proper MRM integrates governance, validation, monitoring, and risk culture
RELATED Blog
comments