Across the financial services industry, increased oversight has led to significant increases in expenses related to assessing and monitoring risk. We see over and over that institutions are weighted with significantly higher regulatory standards, but are not given commensurate financial resources. Model validation is an area where banks are incurring significant expenses to meet regulatory and internal requirements.
In response to client demand, RiskSpan makes the following recommendations to institutions that are looking to maintain the quality of the model validation process while reducing the associated costs.
Model Governance Policy
The first step is devising a model governance policy that is aligned with regulations and the institution’s approach to risk management. Fundamental to the policy is the identification of the models themselves. Once the models are identified, model owners must be notified that their respective models (or tools or applications) are defined as a model, and as such, are expected to adhere to the institutions’ model governance policy. Model owners will need to fully understand the expectations of the model validation regulatory guidance in order to prepare their business units for successful model validation reviews.
Once the policy is created and expectations are communicated to model owners, a risk ranking will need to be performed (for example, High, Medium and Low), which will shape the scope and prioritization of model validation activities. Risk managers within the organization may want to consider different standards for model documentation and the detail of model validation reports based on the risk ranking of the model.
Model documentation is one area where there are significant cost-saving opportunities. Model validation is less costly when business owners have been given easy-to-follow documentation templates based on model governance policies. Template-building is an up-front activity that guides business units to produce quality documentation.
Inversely, the lack of proper model documentation exposes the business to risk and a drag on financial resources. When there is limited communication of the model’s capabilities, purpose and limitations, the workload of a model owner ends up being transferred to a model validation team as the validators attempt to gain a basic understanding of the model. This can end up being a costly activity, and consume valuable resources.
Scheduling of the actual validation itself should take place after model documentation is complete. In fact, the price of admission into a model validation program should be a robust set of model documents. Risk managers that coordinate the validation must be sensitive to business cycle of model owners and times of validators and regulator expectations. At the same time, validation activities should not be pushed to the very end of the year.
Validation Test Plans
An additional element that can be developed early-on is a validation test plan which increases transparency (particularly when third-party validators are used) and allows for testing to be run on a periodic basis more efficiently. Test plans can be used for years after they are first developed, and may be modified to account for market changes that could impact model performance.
Buy versus Build?
Bank executives are faced with a choice: outsource model validation or maintain an internal staff to perform model validation activities. Depending on the complexity and required technical expertise to understand a model, a bank may not have the specific expertise contained within an internal model validation department, therefore outsourcing all or part of the validation may be necessary. Alternatively, banks that prefer to keep validation resources “in-house” may consider periodic staffing support with subject matter expertise to make it through periods of high volume of validation activities (for many banks, validations end up occurring in the 2nd half of the year, and end up being in a crunch at year-end).
About the Author:
Pat Greene currently supports strategic and tactical initiatives by RiskSpan to enhance a suite of valuation tools that provide pricing, analytics, and risk reporting for multiple asset classes, including mortgages and structured securities. He has delivered technology solutions and provided financial model validation support to multiple RiskSpan clients whose business practices rely on credit models, interest-rate models, prepayment models, income simulation models, counter-party risk models, whole loan valuation models, and bond redemption forecasting models. Pat is an experienced executive who has been responsible for the management of a multi-billion dollar asset securitization program for a national financial institution. He has experience in the development and implementation of business unit objectives, management of a $4 million operating budget, and the oversight and monitoring of service levels with legal resources, accountants, and other financial institutions that supported an industry leading asset sales program. He is a skilled manager experienced in the development of business strategy that leads to business process change and technology implementation. Pat is a graduate of the United States Naval Academy and received a M.B.A. from Loyola College in Baltimore, Maryland.