Articles

ALM modeling seeks to project and measure the impact of different interest rate environments on a financial institution’s profitability and capital levels. This interest rate risk can be understood as the difference in repricing speeds between the different asset and liability products that the credit union holds. However, not all products have contractual maturities or repricing schedules. Such is the case for many deposit products, often referred to as Non-Maturity Deposits (NMDs), which also happen to be the primary funding source for most credit unions. To appropriately assess the interest rate risk of the balance sheet, assumptions must be made as to the rate sensitivity (betas) and expected maturity (WAL/decay rates) of these funds.

First, we will explore betas on Non-Maturity Deposits, which target to measure the how management will set dividend rates as wholesale market funding changes in cost. Betas are expressed as the percentage of a market rate move that would be reflected in the new product price. For example, a 25% beta on a deposit that is modeled against Fed Funds would indicate that the dividend rate would increase by 25 bps for every 100 bps increase in Fed Funds. Generally speaking, a higher beta assumption would be considered more conservative, because it would imply that the credit union would pass through higher dividend rates in rising rate environments as well as aggressively cut deposit dividends when rates are predominantly low. Let’s consider an example:

Share Drafts have a 25% beta assumption to Fed Funds. In this example, Fed Funds is currently at 1% and the dividend rate is 0.25%. If Fed Funds increases to 2%, the CU would increase the dividend rate to 0.50%.

One of the pitfalls of betas is that the rate relationship is described as linear, which may not be reflective of management’s rate setting decisions. For example, after the long low-rate environment following the Great Recession, many financial institutions were very slow to increase deposit dividends when the Fed started the process of hiking rates. It can also be difficult to assess what the market rate required for a particular product in different rate environments due to the dynamics of local competition in your area. This highlights the need for management to periodically review and adjust betas to ensure that the ALM model provides the most useful information for strategic decisions.

The other primary components of ALM assumptions are the expected maturities and decay rates, two different approaches to address the same question: How long will a deposit remain at the institution? Expected maturity may also be stated as a weighted average life, or the expected the midpoint of an amortization schedule (e.g., when 50% of the account has been withdrawn). A decay rate would be the speed at which the account is withdrawn, and therefore can be used to calculate the expected maturity.

These applied assumptions are used to find the point on the pricing curve that is used to calculate the economic value of those liabilities. For example, if a 5-year assumption is applied to a Share Draft, an ALM modeler may model the value of the dividend paid on the deposit compared to 5-year funding from the FHLB, LIBOR contracts, or other pricing curves. Because credit union funding is so concentrated in non-maturity deposits, the expected maturity assumptions will be the primary driver of determining the interest rate risk in the reports.

A deposit study is one method to determine the appropriate average life assumptions for non-maturity deposits This involves analyzing the change in deposit mix and average balances over time. Unfortunately, past performance is not necessarily indicative of future deposit behavior. To the contrary, shares have grown across the credit union industry every year with few exceptions for a few decades. In 2001, the National Economic Research Associates (NERA) released a broad study on the average lives of non-maturity deposits at credit unions per the NCUA’s request. Their base case average lives are summarized in the table below[1]. While the study is admittedly a bit stale, we find that most credit union average life assumptions tend to fall into the ranges below.

ALM modeling is a critical component of both credit union strategy and regulatory compliance. Periodic review and adjustments to model assumptions will be necessary as member behavior continues to evolve in the coming years. This is particularly relevant due to the rapid expansion of credit union balance sheets over the past year and a half, fueled by government stimulus programs.  As the dust continues to settle, it will be important for credit union managers to reassess their deposit behavior as it may have materially changed.

By Phil Lucas

Senior Balance Sheet Adviser