The digital way to model risk—and pass a tough exam

As one of the top U.S. retail banks by deposits, this financial institution offers a full suite of products and focuses on a simplified, personal approach that sets it apart. That's the good news. The bad news: It was failing the Fed's Comprehensive Capital Analysis and Review (CCAR) exam, which measures the fitness of the country's banks. Facing increased regulatory scrutiny, the company turned to Genpact to get its MRM standards and policies up to speed and to strengthen its model risk governance capabilities.

Challenge

Turn failing model risk management into a best-in-class process

Banks create scenarios—models—to understand their risks when making investments, or issuing credit. But sometimes the models themselves are wrong: As well, sometimes, they no longer perform the way they should and don't flag risks.

That happened to one of the top retail banks in the U.S.—and it failed the Federal Reserve's annual CCAR exam as a result. The Fed uses the CCAR exam to determine if banks have enough capital to withstand economic stress, and to assess whether their plans for dealing with risk are strong enough.

Clearly, this bank had to have a better risk management solution, so it turned to us to craft one. We found the following evident problems right away:

• No single central repository of models
• No realistic schedule of upcoming validation events
• No consistent standards, templates, or processes to meet regulatory requirements in a timely and efficient manner

No question—independent validation of its credit, capital, and market risk models was critical.

Solution

Automate and integrate

After analyzing the bank's current MRM process, we identified a number of issues, including:

• 400+ legacy models, many not validated
• Skyrocketing costs for regulatory consultants—more than $30 million annually
• Excel-based inventory management that depended heavily on manual inputs
• Based on our analysis, we chose our artificial intelligence powered solution Cora RiskMonitor to replace the bank's existing MRM framework.

Cora RiskMonitor, built on our AI-based Cora platform, integrates model risk management and allows machines and humans to work together for more accurate and efficient validation and governance. Using Cora RiskMonitor, our team created a customized MRM solution that included:

• Centralized model inventory - a document organization system for inventory management
• Integrated model risk workflow - incorporating and automating 30+ activities and processes, from generating activity requests to providing new validations and getting approvals
• Streamlined model validation - seamlessly integrating the MRM framework for end-to-end access across the entire model validation process
• Harmonious MRM integration - with the bank's internal systems for easy data retrieval and display
• Structured model risk reporting - that produces up to 25 customized reports

Impact

Truly functional models. Clear-headed decisions. Higher grades.

Our solution allowed the bank to make models that work, so it could make better decisions and do better all around. Here's what it got:

• 100% validation of tier-one models and 80% validation of tier-two models
• 70% of backlogged models validated
• Improved oversight and control
• Support across all model classes
• Key regulatory matters requiring attention remediated

The bank strengthened its credentials, built strong client support, and achieved operating excellence. Delighted with the impact, it decided to go in for a multi-year managed partnership with us. By doing so, the bank to not only achieved a scalable delivery model, it also strengthened its overall MRM standards.