Over the past several years, financial institutions have embraced the increasing use and reliance on technology. Automated predictive, economic, and financial models have assisted them in making faster and better business decisions. Many institutions are also in the process of developing or implementing credit loss models to address the Financial Accounting Standards Board’s new current expected credit loss (CECL) standard.

But how should organizations manage risks? A robust model risk management (MRM) framework is critical.

Increasing model use, increasing risks

The proliferation of data and the increasing complexity of financial analyses have caused many financial institutions to turn to models to increase performance, reduce mundane and repeatable tasks, and save time and resources. However, the use of models also presents significant risks if a strong MRM framework is not in place to govern usage.

The challenge is that few small and medium-sized financial institutions have robust model risk management processes to govern their models. While financial institutions above $10 billion are subject to model risk management regulatory guidance, smaller financial institutions do not have the same obligations — although MRM is encouraged. This has led many to approach model implementation on an ad hoc basis, with functional areas developing models to enhance their specific decision-making processes. The issue with this provisional approach is that it opens an organization to a wide range of risks, including those associated with input accuracy, data completeness and alignment of bank-specific assumptions and strategic goals.

Making model risk management a priority

Smaller institutions might not be subject to the same regulations as their larger counterparts, but they should not ignore such requirements altogether as they may be subject to such MRM requirements in the future. Additionally, if they are going to spend the time and resources developing and implementing models, financial institutions should make sure those models work as intended. The last thing any financial institution wants to do is rely on inaccurate models for making critical business decisions.

Where to start?

Financial institutions using predictive, financial, or economic models should consider enhancing their approach to MRM. As a starting point, this could include undertaking the following key activities:

• Create an inventory of existing models — It is essential to generate a list of any current or in-development models. Be clear about the difference between a model and a tool so all stakeholders understand how to use and contribute to the inventory. In connection with documenting the inventory, include each model’s purpose, owner, data sources, and significant assumptions.
• Understand regulatory requirements related to model use and verification — Financial institutions should take time to understand the regulatory requirements related to model development, implementation, and use, including validation, even if compliance is not currently required. This understanding will help the organization manage its entity-wide risk and help establish MRM processes aligned to comply with regulations they may be subject to in the future.
• Test and validate models — Institutions should test and validate any significant or complex models before and after implementation so management can be confident in model outputs. For example, before implementing a new model, running parallel with the existing process will ensure the new model is operating as intended and in line with expectations. Ongoing, the model should be tested for accuracy to determine if use is still appropriate given potential changes in facts and circumstances. As recommended in the regulatory guidance, individuals or a third party (independent from the models’ users and developers) should conduct the testing and validation. Based on the testing process results, institutions can identify model errors, track corrective actions, and ensure appropriate use.

o Note: Financial institutions should validate the usage of third-party models to determine whether a model is appropriate for its intended use and that any customizable model assumptions are accurate and relevant.

• Involve the right stakeholders — MRM should be an entity-wide activity. The board should be responsible for providing governance of the entire MRM process, while management should develop the MRM framework and related strategies. Leaders with insight across the organization should be engaged in the MRM process to ensure assumptions are appropriate, model documentation is robust, and data sources are valid and accurate.

Knowing you are making the right decisions

Models can be instrumental in driving better business decisions or your financial reporting process — but only if you are able to rely on the outputs. If you would like more information on our model validation services or how we can enhance your MRM framework, please contact your local Plante Moran business advisor.