Summary
A global investment bank headquartered in Japan encountered regulatory issues with its model monitoring. Together, the bank and Evalueserve created a model monitoring program. Within a couple of years, it brought 80 different models into the process, built model monitoring templates, assigned monitoring frequency, determined thresholds, and set a testing strategy. Evalueserve also built micro-automations to streamline repetitive manual tasks related to model monitoring. The bank is now in good standing with regulators and has improved, more efficient model monitoring processes to prevent future issues.
The Problem
The investment bank encountered regulatory issues in countries including the United States. The U.S. Office of the Comptroller of Currency (OCC) expressed concern about the bank’s ongoing model monitoring program. The OCC issued some Matters Requiring Attention notifications (MRAs) and noted the bank was in violation of the SR 11-7 regulation. The investment bank’s team was overstretched and recognized they did not have the capacity to resolve the issues internally. Evalueserve was brought on to help them meet model monitoring and validation requirements.
Our Solution
Evalueserve partnered with the global investment bank to set up a new model monitoring program. Our experts evaluated existing model monitoring processes, which highlighted the deficiencies and limitations of their previous system. The team then compared their existing model monitoring framework with industry-wide best practices and proposed changes to the old process. Upon receiving the clearance from the client, we implemented those changes.
To set up the model monitoring program, our experts:
- Built templates,
- Assigned monitoring frequency,
- Determined thresholds, and
- Set a testing strategy.
The revamped process for model monitoring was used on 80 different models in a variety of model classes, including pricing, credit risk, AML, and counterparty credit risk.
Our approach was two-pronged with risk analysts and programmers. The risk analysts offered invaluable domain expertise, and the programmers built micro-automations to streamline repetitive manual tasks.
The team also developed accelerators and use case-specific automations that helped speed up model monitoring even from the early days of our engagement, resulting in a 40-50% faster process. The team works closely and has relationships with the bank’s model owners and developers, helping address any monitoring or validation recommendations.
Business Impact
Evalueserve’s work streamlined the model monitoring process. Now, when a new model is developed, there is an approach document outlining how it needs to be brought into the monitoring program. The MRAs were addressed, and the global investment bank was brought back into compliance with SR 11-7.
At the beginning of this engagement, the model monitoring process took six to eight weeks to complete for one model; now, it takes four to five weeks to complete model monitoring for one model, and the effort typically marks a 40-50% increase in efficiency.
