A clash is brewing between US bank model risk executives and the Bank Policy Institute (BPI) over the future of a key Federal Reserve guidance, SR 11-7, which governs model risk management. While the BPI called for the repeal of the guidance last April, senior model risk executives at banks are pushing to maintain it, according to a report from Risk.net published Tuesday.
SR 11-7, issued in 2011, establishes a comprehensive framework for managing the risks associated with using quantitative models in banking decisions. It applies to all Federal Reserve-supervised institutions, including bank holding companies, savings and loan holding companies, state member banks and foreign banking organizations operating in the US, as well as those overseen by the Office of the Comptroller of the Currency (OCC). The guidance covers models used for credit risk assessment, market risk management, and regulatory compliance, among other areas.
The BPI argues that SR 11-7 is outdated and hinders innovation, particularly in the rapidly evolving field of artificial intelligence (AI). However, model risk executives believe the guidance provides essential oversight and consistency in a sector increasingly reliant on complex algorithms. They fear that repealing SR 11-7 would lead to a weakening of risk management standards and potentially expose banks to greater financial losses.
The debate comes as banks are increasingly adopting AI to enhance operations and improve risk management. SR 11-7’s principles-based approach allows banks to adapt their model risk management frameworks to accommodate new technologies like AI, according to a 2024 report by the BPI itself. However, the guidance as well raises the complexity of AI implementation, prompting banks to invest in stronger governance, policies, and controls.
Failure to comply with SR 11-7 can have significant consequences for banks, though the specifics of those consequences were not detailed in available reports. The guidance defines a “model” broadly, encompassing not only traditional statistical models but also modern machine learning systems and generative AI applications. Any system that processes data through mathematical transformations to produce quantitative outputs used for business decisions falls within the scope of SR 11-7.
The disagreement highlights the tension between industry efforts to reduce regulatory burdens and the desire of risk professionals to maintain robust oversight of increasingly complex financial models. As of Tuesday, the Federal Reserve had not publicly responded to the BPI’s call for repeal, leaving the future of SR 11-7 uncertain.
