Why Model Governance Matters
Last month, my colleague Daniel Melo reported on a series of meetings he and IDC Financial Insights had held with European bankers on the subject of model governance. This is an increasingly vital issue for banks, insurers and other companies whose success depends on the quality of their analytics. Now you can read IDC’s thoughts in a new paper by Michael Versace, EMEA Banking – Model Governance Framework.
In this paper, Versace states that, “Developing and maintaining strong governance, policies, and controls over analytics and model risk is fundamental to the effectiveness of business strategies and financial decision making. A weak governance function reduces the value of modeling investments and quality of overall enterprise risk management programs.” As Versace points out, model governance unites the twin concerns of risk management and regulatory compliance.
IDC Financial Insights posits a set of model governance best practices, summarized here:
- Inventory and classify all models
- Develop and adopt standard production and deployment processes
- Use standard model validation frameworks and independent audit policies and practices
- Manage the risk associated with the data used in modeling operations
- Optimize model governance frameworks
This last point may warrant an explanation. What IDC means is that the processes and policies used for model governance need to be linked closely to the performance of decisions based on those models. Ultimately, model governance is not an end unto itself — it should be part of a comprehensive approach to decision management, to make sure you (as FICO’s tagline has it) “make every decision count.”
FICO is glad to see the renewed focus on model governance. Like IDC Financial Insights, we see this as an absolutely critical function to ensure model quality and operational integrity, which, as the white paper points out, “are as important as the complexity of the calculation of the model itself.”