Albert Einstein may have said: “The important thing is not to stop questioning.” But constant questions hardly feel like a blessing when it comes to model management.
As scrutiny of analytic models increases, banking regulators are asking probing questions not only about how models affect credit policies and customer decisions, but also about the processes used for developing, validating, deploying and updating them. Banking executives and internal compliance officers, increasingly aware of the full dimensions of model risk, are also asking pointed questions.
Finding answers can add drag to the performance of analytics teams—often pulling them away from high-value work that adds to the company’s bottom line.
To demonstrate compliance and reduce response time to detailed questions, leading banks are implementing formal model management processes throughout the analytic lifecycle. But while best practices may be understood, they can be challenging to deploy consistently across analytic teams. It’s also difficult to know if they’re being followed at the right level of granularity, such that no matter where regulators probe—and even with analytic staff turnover—all questions can be readily answered.
What can banks do? We’ve been working with clients to implement model management infrastructures that include automated, configurable model workflow tools to promote process consistency and accountability. This approach improves model governance without creating extra work for analytic teams.
In fact, a bank devoting 80% of modeler time to regulatory requirements could reduce that to 20% or less. Analytics teams are freed to spend more time creating and refining models, leading to better model performance and more profitable decisions.
We cover these efforts with banks, along with lessons learned, in our latest Insights white paper, “Reducing Regulatory Drag on Analytics Teams” (No. 82). I invite you to take a read.