We are using cookies, in order to improve the site usage.
You can change your cookie settings at any time. Learn more
bg-arrow-down icon-arrow-up icon-back-to-top icon-linkedin icon-menu icon-search icon-twitter logo-white slider-arrow-left-gray slider-arrow-left slider-arrow-right-gray slider-arrow-right

Part 1/5: Testing and Validating Asset Liability Management Model Beyond Checking the Compliance Box


Given risks in the current banking environment, regulators are more aggressively supervising the interest rate risk models. There are two approaches to ALM models validations: i) A full replication of the model being used; and ii) A rigorous testing of the assumptions, calculations and methodology of the subject models. Typically, ALM models are considered “black boxes” where the institution inputs the data, parameterizes the assumptions, and the model outputs the results. The outputs from ALM models are then used to drive major business decisions. If a financial institution does not have full confidence in their model and its ability to produce accurate, reasonable and stable results for their organization, this could result in lost earnings opportunities, potential risk exposures and regulatory exceptions. Banking examiners typically focus on the following three interest rate risk areas:

  • Effective Interest Rate Risk (IRR) measurement - Models, metrics, and risk limits, including:
    • Key assumptions: monitoring and updating
    • Deposit duration/valuation assumption validation
    • IRR model effectiveness to evaluate risks tied to the fair value of investment securities
  • Stress testing (scenario and sensitivity analysis) to identify/quantify IRR exposure and excessive risk
  • Management action plan: Procedures for taking risk-mitigating actions when risk limits are breached

In order to validate, BankingBook Analytics (BBA) offers both options to its clients, i.e., perform full replication of subject models based on given data or perform a rigorous professional examination of calculations to help our clients not only better understand the fundamental aspects of their models but also to provide a complete assurance in the outputs.

In this first of the five part series, we will outline the tests for key modelling assumptions. Modelling assumptions are not constant but need to be reviewed periodically to ascertain their validity. When reviewing the use and reasonableness of assumptions, it is important to benchmark them across an array of similar institutions. Some notable assumptions that need to be tested at regular intervals are:

  • Repricing rate paths: A bank or a credit union needs to use rate assumptions to reprice the run-in volumes of assets and liabilities. Typically, these rate assumptions are based on the market conditions. A bank or credit union then generates the income simulation of interest margin making. These expected base case repricing rate paths are then shocked higher or lower on a parallel basis and sustained at the shocked level for the full 12 month period. These rate shocks cause the base NIM forecast to change. The change in interest margin becomes the interest exposure for the shock scenario.
  • Interest rate sensitivity: Each business line needs to be assigned interest rate sensitivity. Can a 1% rate shock result in a corresponding income reduction? For example, all things being equal, $100 million of fixed rate loans shocked down by 1% should see an income reduction of $1 million over the next 12 months. Based on the funding mix, it needs to be investigated whether a 1% rate shock will result in an income reduction by a similar delta or a part thereof. Loans are usually supported by liabilities which could be fixed term, demand, and some may argue even by equity. The key question that needs to be tested is what funding support exists, how frequently it is repriced and at what corresponding rate shock, i.e., sensitivity?
  • Interest income and expense offset: Typically, Net Interest Income at Risk calculation should not include the assumed growth rate in relation to the rate shock. By removing this ‘white noise’ from the data, shock test results can be compared on an apples-to-apples basis. This approach is also considered acceptable from the regulatory perspective as it streamlines cash flow for repricing purposes.

Stay tuned for our next article in which we will talk about the treatment of deposit accounts in the context of the validation and testing of Asset Liability Management Models.

For more information on this topic, or to learn how the Team BBA can assist your organization,

For more information, contact BBA Marketing

+1 (905) 499-3618

Related Materials


Alternative Rating Development Approaches: Shadow Bond Approach

Where good-bad analysis cannot be used due to lack of default data, the ‘shadow-bond method’ offers a less robust but statistically valid alternative. Here the ability of financial factors to predict default is modelled by measuring their ability to predict external rating agency default rates.


Regulatory Approval of Internal Ratings Based Approach

More than ten years after the roll-out of Basel 2, many lending intuitions in Canada are still using the Standardized Approach (SA) for regulagtory reporting. As a consequence, reporting institutions are either setting aside disproportionately higher capital for their loan book, are engaged in regulatory arbitrage by issuing residential real estate loans or are involved in "originate-to-distribute" lending. All of these aofre-mentioned consequcnes contrinute to higher systemic risk.


Distribution Analysis for Information Risk (DAIR): A Cyber Quantification Framework

We know that cyber threats continue to evolve and pose increasingly significant risks to organizations. We also know that the impact of cyber-attacks extends beyond direct financial consequences. Cyber incidents can lead to serious service disruptions, reputational damage and share price deterioration, along with potential for fines and litigation. We also know that the impact of cyber-attacks extends beyond direct financial consequences. Cyber incidents can lead to serious service disruptions, reputational damage and share price deterioration, along with potential for fines and litigation.