← Fintech GlossaryRegulation & Compliance

SR 11-7 (Supervisory Guidance on Model Risk Management)

SR 11-7

SR 11-7 is a supervisory guidance issued by the Federal Reserve and the Office of the Comptroller of the Currency (OCC) in the United States that establishes standards for model risk management in financial institutions. It defines model risk as the potential for adverse consequences from decisions based on incorrect or misused model outputs. The guidance requires banks to establish comprehensive model risk management frameworks covering model development, validation, implementation, governance, and ongoing monitoring. Key requirements include independent model validation by qualified personnel not involved in model development, documentation of model limitations and assumptions, ongoing performance monitoring, and reporting to senior management and the board of directors. SR 11-7 applies to all models used by financial institutions, including credit risk models, market risk models, operational risk models, and increasingly, AI and machine learning models. The guidance has been highly influential globally, with similar frameworks adopted by regulators in other jurisdictions.

In Financial Services

SR 11-7 is the most important regulatory framework for model risk management in US banking. Banks must maintain a comprehensive inventory of all models, classify them by risk tier, and ensure each model undergoes independent validation before deployment. The validation process assesses model conceptual soundness, data quality, outcomes analysis, and documentation. AI and ML models present unique challenges for SR 11-7 compliance because their complex, non-linear nature makes traditional validation approaches difficult. The guidance requires ongoing monitoring of model performance, including backtesting against actual outcomes, and periodic revalidation when models are changed or when market conditions shift significantly. The model risk management function must be independent from model development, with clear reporting lines to senior management. Banks that fail to meet SR 11-7 standards may face regulatory action, including higher capital requirements or restrictions on model use.

Real-World Example

A large US bank maintains a model inventory of 2,000 models, including credit scoring models, fraud detection models, and AI-based trading algorithms. Each model is classified by risk tier and undergoes independent validation following SR 11-7 requirements. The bank's model risk management team of 50 professionals conducts annual validations of high-risk models, quarterly performance monitoring, and ongoing documentation updates. When the bank develops a new AI model for credit risk assessment, the validation team challenges the model's conceptual soundness, tests data quality, benchmarks against simpler models, and documents limitations. The validation report is reviewed by the model risk committee and presented to the board.

Why It Matters for Finance

SR 11-7 is the foundational framework for model risk management in banking. As financial institutions increasingly rely on complex models, including AI and ML, for critical decisions, the principles of SR 11-7 become even more important. The guidance provides a structured approach to identifying, measuring, and mitigating model risk that protects both the institution and the broader financial system. For AI governance, SR 11-7 provides a well-established framework that can be adapted for the unique challenges of AI models.

Related Terms

Model Risk Management (MRM)AI GovernanceExplainable AI (XAI)

Explore in Finatune

Contextual AITruLensArize Phoenix

Frequently Asked Questions

What is SR 11-7 in banking?

SR 11-7 is supervisory guidance from the Federal Reserve and OCC that establishes standards for model risk management. It requires banks to have comprehensive frameworks covering model development, validation, implementation, governance, and monitoring. The guidance applies to all models including AI and ML models.

How does SR 11-7 apply to AI models?

AI and ML models present unique challenges for SR 11-7 compliance due to their complexity and non-linear nature. Banks must validate AI models for conceptual soundness, test data quality, benchmark against simpler models, document limitations, and ensure ongoing monitoring. The guidance's principles-based approach is adaptable to AI.

What are the consequences of non-compliance with SR 11-7?

Banks that fail to meet SR 11-7 standards may face regulatory action including higher capital requirements, restrictions on model use, enforcement actions, and reputational damage. Regular regulatory examinations assess model risk management practices against SR 11-7 standards.

← Previous Term: SOC 2 (Service Organization Control 2)
View All Fintech Terms β†’