Model Risk: Establishing a model governance system for a tier 1 G-SIB bank

Challenge and Regulatory Context

  • Our client’s external auditors had identified the absence of an overarching model governance framework.
  • The goal was to help the client set up a model governance system to monitor the model life cycle.
  • Model risk had become an important regulatory topic with the publication of the OCC’s SR 11-7.
  • Monitoring and mitigating model risk were therefore central to the new function.

Success

  • We provided an initial set of model control reviews for various model types.
  • We were able to leverage much of the work done by model developers and validators in order to avoid repeat testing.
  • Our model governance tool proved both suitable for, and adaptable to, a broad range of model types. It was subsequently employed at several tier 1 banks.

Approach

  • We developed a new methodology to break model risk into three separate types: statistical risk, assumption risk and procedural risk.
  • We further developed our “model map” concept to decompose a model into its individual input, process and reporting elements and trace the connections between them.
  • We created a model risk management tool which works alongside the model map. This tool is comprehensive in that it compares the controls around each model element with a self-consistent master list.
  • The tool then generates a semi-automated “control review” document to highlight high risk areas of the model, as well as any gaps in the control regime throughout the model lifecycle.

Click here for more on Fintegral’s services in model risk management, model validation and model governance.