Taking Inventory of Model Risk: Day Four of RMA’s Governance, Compliance, & Operational Resiliency Conference

Thanks to all who attended the 16th RMA Governance, Compliance & Operational Resiliency Conference. Throughout the week, GCOR XVI virtually convened senior regulatory and banking executives to discuss climate risk, innovations in operational risk, and the technologies and trends transforming model risk management practices.

On its final day, GCOR XVI was capped off by the Model Risk Management Summit. After an incisive opening keynote tracking the evolution of operational risk measurement, panels highlighted key challenges facing modelers and risk managers, and how to maximize the model risk management resources available.  

Here are some key takeaways from GCOR XVI, Day Four.

Be open to adjusting your models. Panelists at “Understanding Model Overlays and Underlays” discussed what risk managers should consider when making adjustments inside and outside of a model, and why validating calibrated models does matter.

Nikolai Kukharkin, head of quantitative risk control at MUFG, stressed being flexible on how long a new model change will last, and to see if a model even needs to be adjusted. “Don't throw away your model when there are new aspects coming into play,” added Kukharkin.

Forecasting models often require use of overlays and other adjustments in periods of increased uncertainty. Liming Brotcke, model validation executive at Ally, urged model risk managers to quickly identify the root cause of changes seen in a model’s output.

It is also critically important to know when to pull back on model overlays. “Ongoing performance monitoring is critical so that the model can be constantly evaluated,” said Yu Pan, chief model risk officer at U.S. Bank. Panelists recommended having a process to record additional changes, and to also document when and why a model or its overlays was retired.

Making uncertainty transparent. Panelists at the “Taking Inventory: Model Identification and Ownership” session shared their views on how to navigate the vagueness of model definitions. The discussion included key considerations for risk managers when assessing model inventory of both quantitative and qualitative models, and the governance of this process. “If there is uncertainty in the output that is produced by that tool, then it’s definitely a model,” noted Aaron Benson, director of model risk management at Zions Bancorp.

To ensure proper risk management, model risk managers will benefit from having a holistic approach when assessing their organization’s models, tools, and the governance frameworks, according to Vinit Jagdish, director of model risk management at Western Alliance Bank. “It does come down to you to see where people try and actually change their development approach to get out of these different [regulatory] requirements,” said Jagdish.

PNC Bank Director of Model Validation Evan Sekeris concluded that model risk managers can foster positive interactions between the regulators and the modelers by seeking greater transparency for both groups. Sekeris added that model risk managers must “have strong governance, clear processes, and be as comprehensive as possible, and then execute correctly.”

If you missed a live session, don’t worry – all sessions from GCOR XVI will be available to watch at your convenience, from June 10 - 30, 2022. 

We look forward to seeing you at GCOR XVII!

Related Links:

- GCOR XVI, DAY 1: Climate Risk and Opportunities 

- GCOR XVI, DAY 2: Assessing Emerging Risks; Managing Non-Financial Risks 

- GCOR XVI, DAY 3: Retaining Talent; Training Boards

- GCOR XVI, DAY 4: Taking Inventory of Model Risk