Despite the many challenges faced this past year, many financial services companies have made substantial progress in their transition away from the family of Interbank Offered Rates (IBORs). Much of the work in 2020 targeted critical areas of foundational infrastructure and processes supporting the discounting transition from the London Inter-bank Offered Rate (LIBOR) to the Secured Overnight Financing Rate (SOFR) as well as from the Euro Overnight Index Average (EONIA) to the Euro Short-Term Rate €STR). A core piece of this transition work included models and analytical platforms supporting the products impacted by the discounting switch. The work related to model risk management (MRM), however, is far from over with the second and third lines of defense just getting started on their assessments.
At this stage, companies should have a comprehensive list of impacted models along with a timeline for when each model is needed in production. High risk and highly used models should be identified and resourced effectively as delays in validation are likely to cause issues for effective business management. Models that forecast interest rate curves and support product enablement should be given prominence based on the timing needs and materiality to the respective businesses or functional areas that utilize the outputs to conduct their business.
After establishing the impacted model inventory, a survey of model owners and developers to identify their respective concerns with the transition can provide key insights into thematic issues across the model portfolio. Data quality is likely to be a pervasive issue when considering the infancy of markets utilizing the Alternative Reference Rates (ARRs). A recognition of model shortcomings by model development teams, and a thorough vetting of its impacts by the business, provides MRM and Audit with a good foundation for which to conduct their respective reviews. These teams should also provide feedback on whether proposed solutions to these challenges adhere to the spirit of the relevant regulatory guidance, internal model risk management policies and industry best practices.
With the identification of data, systems, technology, or other challenges such as resource capacity, the execution plans should address how to mitigate the risks faced by the transition. First-generation models using ARRs need stringency in the vetting of assumptions and sensitivity analysis. Additionally, a more aggressive model monitoring schedule is likely appropriate as the data used in production may change considerably as market using ARRs mature. The maturity is likely to impact the dynamics of the data, including potential reduction in rate volatility, requiring a more frequent model recalibration.
Many of the IBOR-based models have had years, if not decades, to refine, recalibrate, and improve model accuracy. This is not a luxury companies have for first generation ARR models and a deep analysis should be undertaken by model developers, model validators and model auditors to ensure that materiality and risk rating classifications remain relevant. Moreover, some of the automated processes used to execute LIBOR models may not be adapted in time for the first production runs of ARR models which may create additional operational risk. In some instances, compensating controls may be necessary to help mitigate some of these risks which only creates additional work for MRM, Audit, and business model owners as they will have yet another set of artifacts to review.
Model validation and audit teams require specialized skill sets and competencies to provide effective challenge and align to the process rigor supporting strong MRM outlined by supervisors and internal policies. Talent is expected to be in short supply this year as LIBOR is not the only large-scale exercise to be undertaken by these respective groups. Continued management of the ongoing COVID-19 pandemic, the Comprehensive Capital Analysis and Review (CCAR) cycle, Recovery and Resolution Planning (RRP) submissions, and new models (e.g., machine learning, climate risk) are expected to tax the capacity of these teams.
To address the continuously increasing burden on these validation and audit teams, heavy recruitment of talent is always an option to increase the number of available resources. This is a zero-sum game for the industry as one bank’s gain is another bank’s loss. Searching outside the industry for similar skill sets may partly resolve this issue, but the time consumed in bringing a model validator or model auditor up to speed in a new industry will likely limit the contribution to the overall workload at least immediately. Additionally, there is the challenge of determining what to do with the excess resources once the transition is complete.
To overcome these issues, banks need to be efficient and develop execution plans that provide insight into the highest priorities. Given the nature and extent of the challenges, raising an army of model validators or auditors is not a viable option. Banks should review their execution plans to ensure that they are utilizing their resources across the various programs and teams are producing the right artifacts to enable the right prioritization of deliverables. Finally, banks should ensure that there is discussion around accelerating strategic solutions in lieu of temporary solutions to save limited time, resources and money in the future.
How DHG Can Help
DHG’s Enterprise Risk and Quantitative Advisory team provides benchmark transition solutions with the appropriately skilled model risk resources to augment financial institutions’ MRM and audit teams, as well as support relevant model owners. Our global platform is designed to address the predictable surge of newly developed and legacy models requiring coding, calibration, and validation resulting from replacing IBORs with alternative reference rates. Our team of enterprise risk and regulatory professionals, quantitative analysts and thought leaders have delivered MRM services across multiple industries and has the knowledge needed to deliver and communicate on complex topics such as benchmark transition.