What is Model Risk Management?

What is Model Risk Management and How Financial Institutions Strengthen Governance, Compliance, and Decision-Making with Evolving Models

Model Risk Management (MRM) is a process followed by financial institutions such as banks, insurance companies and asset managers to compensate for the potential risk of failures of the quantitative models used as part of their core functions. A model can be a simple algorithm or a complex quantitative process to assist stakeholders in the financial institution to take decisions prudently. As part of the lifecycle of models, the MRM leverages the experience of experts like model developers and validators to perform the development, validation, implementation, monitoring and documentation of the models.

Why Does it Matter?

Models and quantitative analysis play a crucial role, especially in decision-making in financial institutions. For example, let's say banks use models to decide who gets a loan, what the origination risks are and how to measure and mitigate it, and how to create the capital reserves. Therefore, part of the risk of underwriting the loan goes to the model. The number of models used by a financial institution goes from hundreds to thousands, and MRM has become more complex due to the evolving nature of the models, including Gen AI. With this complexity, more MRM guidelines issued by central banks are more detailed in terms of expectations to be followed by financial institutions. Without proper MRM, financial institutions may face serious consequences like financial losses and might even sufferreputational damage. 

Case Study: How the adoption of an Inventory and Governance solution helped a Bank comply with the MRM expectations

Scenario: A Bank was interested in creating a strong MRM practice. They were looking for solutions that could automate and enforce the governance of their Model Inventory, while adding efficiencies on the model validation and monitoring.

Evaluation: After reviewing the capabilities of several vendors’ solution on Model Governance and Inventory Platforms, the financial institution selected MRM Vault from Solytics Partners due to the road map vision and ability to capture emerging models, such as Gen AI. Also, the MRM Vault product had most of the features required with the ability to configure at the front-end. A product purpose built to help financial institutions streamline workflows, consequences like financial losses and might even lead to reputational damage. 

Client outcome:

With MRM Vault, Solytics Partners made it possible for the institution to launch a fully operational MRM platform efficiently within Weeks. Resolving the following:

• Replaced completely the Manual tracking of models to an Automated Governance System which aligned with the MRM policy.
• The client was able to get quick visibility through custom dashboards for Model owners, validators and PMOs.
• Internal MRM efficiency was scaled without additional effort or resources.
• Governance system became more transparent and Audit ready.

Added Value: The financial institutions’ transformation went beyond, and MRM Vault redefined the organization's model governance culture. It acted as a tool which brings accountability and innovation across the business. Consequences like financial losses and might even lead to reputational damage. 

Key Model Risk Management Practices:

In addition to the regulatory guidelines, a MRM program typically includes at least the following components: ‍

1. Governance: Governance means having the right roles and responsibility's structure, and policies defined to manage model risk efficiently. It can include structures like setting up clear policies, adding procedures, assigning tasks, and regular reviews of the policies. Each model and stakeholder of the organization plays an important role.

2. Model Inventory: A detailed inventory of all models is expected to be maintained by financial institutions as a pre-requisite. These models can be in use, or in development, or retired. Inventory might possibly contain a varying level of information (model metadata or model attributes) with a definitive purpose, thereby defining how risky, material or complex the models are, when and how each model can be used, etc. The model inventory acts as a central source to monitor model performance, risk ownership and compensating controls. 

3. Model Validation: In today’s evolving risk environment, models play a huge role in the prediction and management of financial information. That is why we don’t want models to fail or contain vulnerabilities when producing its outcomes. This is where the independent validation of models comes into picture where it helps structuring thereby checking if the model is working correctly. This way the process is objectively driven and keeps the review fair from model developers. 

4. Ongoing Monitoring: Models are required to be monitored to make sure they are working as intended based on the methodology they were trained in and performing as expected to meet regulatory requirements and evolving industry practices. Model owners need to monitor model performance, and the inventory platforms provide support to identify model breaches and create processes to mitigate the problem, have a model changed or decommission the model. 

5. Documentation: Evidence of the model information as well as its lifecycle captured as part of the governance process. These documents are intended to help a third party to understand how the model works, how they are used and how decisions are made from its outcomes. The Model Documentation includes details like when this model was built, when it was validated, etc. ensuring transparency across all the stakeholders in the MRM team

Looking Beyond MRM

As businesses increasingly rely on models to drive strategy, operations, and customer engagement, the risk posed by those models cannot be ignored.

By establishing a sound MRM program, financial institutions can improve model reliability, reduce compliance exposure, make smarter, safer decisions in an increasingly complex business environment. 

_‍‍References

1. Supervisory Letter SR 11-7: Guidance on Model Risk Management
Link: https://www.federalreserve.gov/supervisionreg/srletters/sr1107.htm

2.  Comptroller’s Handbook – Model Risk Management
Link: https://www.occ.gov/publications-and-resources/publications/comptrollers-handbook/files/model-risk-management/pub-ch-model-risk.pdfmodel-risk-management/pub-ch-model-risk.pdf

3. Guideline E-23 – Model Risk Management
Link: https://www.osfi-bsif.gc.ca/en/guidance/guidance-library/draft-guideline-e-23-model-risk-management

4. Model Management Standards (Attached to Notice 5052/2022) – Central Bank of the UAE
Link: https://www.centralbank.ae/media/0oaarr3a/model-management-standards-attach-to-notice-5052-2022.pdf

5. Prudential Practice Guide CPG 220 – Risk Management – Australian Prudential Regulation Authority (APRA)
Link:  https://www.apra.gov.au/sites/default/files/cpg_220_april_2018_version.pdf