Algorithm and Model Risk Governance Review

The obligations and definitions for model risk management set out in the Fed’s SR11-7 have caused a high level of confusion in US institutions active in the algorithmic (algo) space. The definition of a model, in particular, is so broad that in some circumstances it may cover not only quantitative financial models but also algorithmic trading tools and components.

In mapping this definition onto their current algo and model governance frameworks, the client saw both overlaps and gaps, creating risk of both a lack of governance in some areas and a constricting level of governance in others.

A senior GreySpark team of SMEs and analysts partnered with the client to understand their approach and triangulate it with that of peers, best practices and the regulatory guidance in order to deliver a comprehensive, achievable and robust set of policies and definitions.

Detailed analysis closely aligned to the client’s ask:

• Current State: Understand the full scope of the client’s issue, teardown of the current processes including all versions of policies and procedures in order to understand how the current state was formulated. This process included multiple senior stakeholder interviews.
• Regulatory Landscape: A breakdown of relevant regulations in regions applicable to the client, looking at current and future rules to give the client a 360 view.
• Peer Review: Leveraging our relationships across the street we performed 13 interviews with e-trading Heads in order to understand the industry’s approach and problem sets.

GreySpark provided the client with continuous updates throughout the project allowing the client view of the process and a sense of collaboration where possible. The ultimate delivery was a 50+ page document covering the following areas in forensic detail:

• Client Current State
• Regulatory Landscape
• Peer Review
• Recommendations

GreySpark was able to establish new, unambiguous definitions, clarify new processes through decision-tree illustration as well as documentation and eliminate the gaps, large and small, which were the result of two parallel policies being developed at different times by different groups.

GreySpark also worked with the client post-engagement to assist with a small number of live examples in order to display our work and demonstrate its robustness to our stakeholders.

The client has now fully tested the recommendations since deployment and is particularly pleased with the ability to articulate and illustrate the solution set in a way that can be consumed throughout the necessary business units without translation.