Model Risk Manager

The workflow approach drives and enforces the operating model, roles and responsibilities between the various participants (model sponsor, developer, owner, implementer and reviewer) capturing the key data at each stage, and recording the required approvals at each stage.

Proactive prompts via e-mails or alerts will highlight any delays in the process and reporting indicates either up or downticks in progress.

Issues can be raised against a model in any phase of its lifecycle and issues can be raised against more than one model if it is a regional, product, country issue. The issue workflow requires the issue owner to acknowledge and accept, at which point actions to resolve the issue are tracked.

Risks follow a similar workflow with the inclusion of mitigating actions and similar pro-active oversight of these actions to ensure nothing falls out of sight.

Regulatory submissions are all tracked and approved through a list of required artefacts for each regulator including standard submission templates, formal approvals, and evidence as examples. The submission process will only flag as ready once the organisations’ mandatory artefacts are complete and approved. Each model can be flagged as a part of a regulatory submission, such as ICAAP or ILAAP, and the individual execution runs and feedback recorded and tracked in the system

Model review against policy and market environment can be dramatically simplified with the ability to link models to the appropriate policy parts and/or market and product data. A change in policy can quickly be traced back to all models that require reviewing and a change in the market environment can quickly be resolved back to the models that would be impacted.

Model Risk Manager automates the process of governing the model life-cycle eliminating the manual co-ordination, reconciliation, tracking and tracing that exists today. Efficiency of the broader organisation is uplifted through the automated hand-offs and controls ensuring tasks are not sitting in in-boxes, are automatically escalated on delay and gaps in submissions are eliminated at source.

Pro-active monitoring of model execution is provided through a model execution server where models post their execution, input market data, parameters and results depending on the organisations’ appetite.

This allows the detection of models running that are stale and out of approval, models that should not yet be running due to lack of approval, result sets that are out of bounds for the given materiality or model execution errors that are not being reported through the primary system in a timely way. This control provides assurance that the models used to drive an organisations decision making and reporting are the ones that were formally approved.

Model Risk Manager can integrate with existing golden sources of information with low-friction data integration options allowing automated updates to audit logs, issue repositories or to hand off sub-work flows into other management tools such as Jira avoiding rekeying and duplication.

To provide overall control and insight, reporting is available at an individual owner or an aggregate department level to track the most important KPI’s such as issues, risks or actions over-due and by how much and what level of materiality,  number of days taken to get through each stage of approval. Key metrics can be monitored both to drive compliance and efficiency.