ALM Model Validation: What You Should Know Right Now

With rates on the rise, interest rate risk (IRR) will become a growing threat for community banks for the foreseeable future. In particular, the regulators have intensified their scrutiny on asset liability management (ALM) model validation as it relates to IRR.

In this environment, even small and low-risk community banks are being held to high modeling standards, with a formal interest rate risk ALM validation becoming standard on most bank examiner requests. The regulators are concerned about community banks going out too far on the yield curve in search of earnings and thus increasing their interest rate risk.

Definition of ALM Validation

Regulatory guidance defines ALM validation as “the set of processes and activities intended to verify that models are performing as expected, in line with their design objectives and business uses.” Or in other words, validation tests how well your ALM model is performing in relation to its intended use.

To be credible and effective, your ALM model validation process must meet two main criteria:

  1. The validation must be independent of the modeling process and whoever is responsible for the modeling.
  2. The process must meet requirements detailed in the regulatory guidance.

Also keep in mind that your ALM model validation process should be commensurate with the size, complexity and level of risk of your bank.

One of the main things regulators will be looking at is the assumptions that are used in your ALM model. In the FDIC’s Supervisory Insights (Vol. 11, Issue 2), the FDIC states that “a systematic approach to developing common-sense assumptions for use in interest rate risk management systems is an important part of a bank’s strategic planning.”

Assumption Weaknesses

The use of unsupported or stale assumptions is one of the most common issues identified by examiners, Supervisory Insights adds. Common weaknesses found during regulatory review of assumptions include:

  • Using peer averages without consideration of bank-specific factors.
  • Not differentiating between rising- and falling-rate scenarios.
  • Oversimplifying balance sheet categories, which can lead to potentially faulty analysis.
  • A lack of qualitative adjustment factors to historic data, such as not considering a higher run-off factor for surge deposits.

Also, banks often don’t attempt to evaluate how the results of their IRR measurements would change in response to changes in assumptions, according to Supervisory Insights. If the results would be significantly impacted by changes in critical assumptions, you should plan for a range of values for these assumptions.

Contact us if you have more questions about ALM model validation.