Non-maturity shares are an essential funding source for most all credit unions and have at times comprised 70 percent or more of total deposits. By definition, NMS are accounts that have no contractual maturity date and may be withdrawn by a member immediately upon demand. NMS accounts exist in different forms and are most commonly grouped into checking, savings, and money market type accounts. The interest paid on NMS typically represents the majority of a credit union’s cost-of-funds expense. These various NMS accounts can respond quite differently in response to macroeconomic changes (most importantly, to the level of interest rates).
Thus, it is important for IRR modelers to differentiate NMS accounts into different behavioral segments based on their estimated sensitivities to interest rate changes and other market factors. Because NMS have no contractual maturity, their ultimate behavior is unknowable and this makes the need to dynamically model them especially critical to the effectiveness of the IRR measurement process.
The valuation of NMS accounts and the computation of NEV simulation results involve several key variables that include rate sensitivity factor, decay rate, and discount rates.
- The rate sensitivity factor, also known as Beta
, is a measure of the sensitivity of a credit union’s repricing of a particular NMS type given a change in market rates. Estimates of RSFs are typically supported by the credit union’s historical repricing behavior. For example, if market rates move by 100bps and the credit union changes the NMS account rate by 40bps, the RSF is .40 (40/100). RSFs are a key driver of NMS account values. - Decay rates measure the amount of monthly or annual run-off from a static pool of shares. The lower the observed decay rate, the longer will be the estimated average life of the NMS account. Because NMS accounts often pay a below market rate, longer average life assumptions will typically result in a greater NMS premium and a higher NEV.
- Discount rates - The NEV ratio, calculated by subtracting the net present value of liability cash flows from the net present value of asset cash flows, is highly dependent upon the choice of discount rates. For example, are liability cash flows discounted by an alternative cost of funds curve or a single point on the Treasury curve plus a pricing spread? Credit union staff and NCUA examiners should understand how the choice of alternative discount rates can significantly affect NMS premiums and the resulting NEV simulation results.
In 2001, the National Economic Research Associates prepared a study of NMS in response to a request from the NCUA that NERA evaluate the available methods for valuing non-maturity deposits of individual credit unions. The resulting study, The Evaluation of Credit Union Non-Maturity Deposits, includes a detailed discussion of valuation methodology of NMS.
The agency also provides guidance regarding sound practices for evaluating the behavior of non-maturity shares in the context of managing risk in its Letter to Credit Unions 03-CU-11, Non-Maturity Shares and Balance Sheet Risk.
Similar to prepayment behavior, the underlying assumptions (behaviors) of NMS are typically among the most material and complex assumptions that management must establish when modeling IRR exposure. The behaviors are influenced by potential actions of management and credit union members. These dynamics need to be considered when assessing the reasonableness of assumptions. Members have control over the amount and movements of their deposit accounts. Management has some discretion over the dividend rates paid on these accounts. In setting rates, management must take into account a wide array of factors, including local and national competitors, the credit union’s desired funding needs, and the relative costs of all funding sources.
The NMS model assumptions usually reflect both aspects of this relationship: management’s discretion to set rates and members’ immediate control over their funds. Consideration should be given not only to historical correlation analysis, but also to management’s strategic intentions regarding future rate movements. If the measurement system has the capacity to reflect different assumptions for rising and falling rates, management should establish rate sensitivity assumptions for both scenarios. While management does have discretion in setting the rates that are paid on NMS, they may have only limited influence over what they can pay because many depositors likely expect rates to remain competitive.
Credit unions can become captive to market pressure when setting rates on NMS balances in order to retain shares. This is especially true for any portion of NMS that is non-core and possesses inherently higher rate sensitivity. For example, during the period of extraordinarily low interest rates after 2007, many NMS accounts grew and exhibited lower sensitivity, which may have been a function of the desire for safety combined with the lack of yield opportunities elsewhere. For higher and/or rising rate scenarios, the NMS assumptions should consider a greater likelihood of non-core NMS being increasingly rate sensitive for those yield curve scenarios with a level(s) that corresponds to pre-2007 ranges.
Non-maturity instruments present a unique challenge for NEV simulation modeling because they lack a contractual maturity date, which means their maturity (and average life) are not known and, therefore, must be estimated. Generally, an asset or liability instrument is valued using known or estimable cash flows (including a contractual or expected maturity date) based on the present value of all the expected cash flows over the life of the instrument discounted at an appropriate interest rate. Therefore, in order to model the intrinsic value of NMS accounts, an NEV model requires a maturity date input. This input is based upon management’s assumptions for the implied runoff over the life of the account (known as the decay rate) and its ultimate maturity.1 The most common assumptions for NMS accounts include decay rates, repricing coefficients (beta rates), and projected dividend rates.
NMS assumptions can either be developed internally by staff or by a third-party service provider. Reliance on third parties is common practice and acceptable provided the credit union’s board and management retain ultimate responsibility for approving final assumptions they deem to be reasonable and supportable. Management should use NMS assumptions that reflect member-specific factors and avoid overreliance on default settings that may be contained as predefined assumptions within off-the-shelf IRR models. Using historical regression to determine estimates of NMS behavior is challenging if the credit union doesn’t possess historical and detailed deposit information.
Credit unions with limited data and/or staff expertise will have more difficulty developing reasonable and supportable decay rates for NMS accounts. Industry averages can provide approximations when data is lacking, but may be inaccurate proxies because they are not uniquely tailored to the credit union’s membership, pricing strategies, market, and experience. However, using industry averages is better than arbitrary assumptions, and management can use them as a starting point until they develop adequate data sets. Management should consider modeling different decay rates under various rate scenarios and, when appropriate, should consider engaging third parties to assist in determining NMS assumptions. Examiners should recognize that, while NMS decay rates are often imprecise, they can be a significant factor in the IRR analysis when coupled with other drivers of NMS valuations.
Assumptions regarding NMS are particularly critical in market environments where credit union member behaviors may be atypical or where a high level of competition for such deposits exists. Generally, rate-sensitive and higher-cost deposits reflect higher decay rates than other types of deposits. Also, credit unions with lower capital levels may become more prone to reputation and event risks and should adjust deposit assumptions accordingly.
NMS (share drafts, regular shares, and money market accounts) are typically the largest segment of a credit union’s liabilities. The majority of NMS accounts are federally insured deposits and credit unions generally benefit in times of macroeconomic stress because these accounts are perceived as safe havens in which to store funds (often referred to as “flight to quality”). Liquid funds can move swiftly when financial markets are volatile as investors seek ultimate security of principal.
A significant influx of flight-to-quality deposits may remain for an extended period during uncertain times, as was observed in the years following 2007. This phenomenon makes it difficult for deposit modelers to distinguish between the portion of balances that is potentially rate sensitive and the “core” NMS that are more stable and long-term in nature. Ultimately, the portion of NMS balances that may be rate sensitive (and thereby more susceptible to withdrawal) is not knowable, but must be estimated just the same. A credit union can set attractive dividend rates to try to retain these shares, but if members leave their balances in NMS, they retain the option to move the funds into other higher-priced shares or out of the credit union altogether at any time.
As stated above, NMS are a large percentage of a credit union’s funding base, and represent a key driver of risk in a credit union’s balance sheet. NMS are, therefore, a critical component in determining a credit union’s exposure to IRR. For NII simulations, the correlation between changes in NMS dividends/interest, compared with changes in market interest rates will be a significant driver of the NMS rate sensitivity. If a credit union assumes its administered rates paid on NMS can significantly lag changes in market interest rates, and that members will not move or withdrawal their funds in response, then those NMS accounts are being modeled as less sensitive (or more “sticky”) relative to market interest rate moves.
In contrast, NMS dividends/interest rates that are assumed to reprice in lockstep with market interest rates are being modeled as more rate sensitive (or less “sticky”). Where management sets the repricing coefficient, also called the beta, is how the level of sensitivity between NMS and their reference market rate(s) is established within the model. A lower beta implies less sensitivity and a higher beta implies more.
For NEV modeling, assigning long implied maturities to NMS accounts typically indicates that management believes they have low rate sensitivity, whereas assigning short implied maturities indicates they believe the accounts have higher rate sensitivity. Assuming the dividend rate is fixed, a long maturity will result in a lower value (benefit) for NMS and higher NEV.2
Conversely, a shorter maturity will result in a higher NMS value and decreased NEV. A credit union that models its NMS with little or no sensitivity to changes in market interest rates (in its earnings simulation or NEV analysis) may significantly understate its actual IRR exposure by understating post-shock EAR and overstating the post-shock NEV.
As previously stated, assigning values to NMS is complex because the actual future behavior of the accounts is not knowable, yet the risk measurement calculations require management to define principal cash flows for NMS accounts over a discrete time horizon and set an expected maturity, so they can simulate expected earnings (for EAR) and a present value (for NEV). A credit union’s NMS behaviors are sensitive to circumstances at the credit union, in the credit union’s competitive market, and in the general economy. Risk modelers responsible for developing assumptions should take such factors into consideration. These factors can include, but are not necessarily limited to:
- The credit union’s need for funds and its ability to use alternative funding sources. In a period of low loan demand, a credit union may change its deposit pricing policies to disincentivize shares and allow some of its NMS to run off. As loan demand increases, a credit union seeking to increase liquidity may raise its rates in order to attract more shares.
- The credit union’s pricing structure and member base. Using a variety of implicit and explicit pricing structures for its core shares/deposits, a credit union can tailor pricing for certain parts of its member base. For example, a credit union may waive certain account fees for members who maintain minimum balance requirements. Applying tiered pricing strategies will divide a credit union’s member base between high-balance, rate-sensitive members and low balance, rate-insensitive members. The demographics of the credit union’s member base may help a credit union to determine the rate sensitivity of their shares/deposits.
- The credit union’s marketing and strategic plans for its share and deposit products. In developing and planning marketing strategies, credit unions may view individual products as having life cycles or market niches that influence how the credit union will position and price them in the future. For example, credit union management may decide to let a certain share or deposit product become less rate sensitive over time, and to introduce a new, more rate sensitive deposit product to members who are more likely to move balances when interest rates change.
- The number and type of competitors within the credit union’s market. The pricing behavior of competitors will likely influence the degree and speed with which a credit union will respond to changes in market interest rates. As consumers become more knowledgeable and comfortable with alternative investments, a credit union’s competitive market expands. Increasingly, a credit union competes not only with other credit unions or financial institutions for shares and deposits, but also with investment houses, mutual funds, and even entities that advertise their rates and services over the internet.
- The general level and trends of market interest rates. Market interest rates (such as the rate that a member could earn by investing in other credit unions, financial institutions, etc.) help determine members’ “opportunity cost” of maintaining a balance at the credit union. The opportunity cost of holding NMS is relatively low when market interest rates are low. As market interest rates rise, so does the opportunity cost of holding these types of shares and deposits. As the spread widens between market interest rates and the rates a credit union pays on its NMS, there may be increasing incentives for credit union members to move balances to other options outside of the credit union.
- Product development and changes in credit union regulations. The development of new financial products and changes in the credit union industry can have a dramatic effect on the structure of a credit union’s share base and member behavior. When analyzing NMS, credit unions should consider any impending legislation or new products that would significantly alter the credit union industry and/or force a credit union to change its assumptions about member behavior.
To identify the appropriate assumptions for NMS, management should:
- Analyze the credit union’s share base and the demographics of its local market.
- Assess how the circumstances described above and any other relevant factors will influence the level of shares/deposits and the rates offered.
- Consider how those circumstances could affect the credit union’s share base differently in alternative interest rate scenarios.
In summary, the tools used to analyze NMS should vary with the size and sophistication of the credit union. Credit unions with more complex balance sheets and cash flows will typically use tools such as share/deposit studies to help create and support assumptions about their shares and deposits. The methods of smaller credit unions may be similar, but less robust.
It’s important for examiners to keep in mind that modeling NMS behavior is subjective and nobody actually knows how the accounts will react to macroeconomic factors (including changes in market rates). Examiners should not debate whether a credit union should obtain a deposit study or take a position on whether the individual underlying assumptions are accurate. Instead, the primary supervisory focus should be upon:
- Whether the scope and rigor of the analysis is commensurate with the size, complexity and risk of the credit union
- Whether the sensitivity analysis is sufficiently broad and thoughtful in its scope
- How management is using the information in conjunction with its IRR management decisions
Additional guidance on NMS is included in 03-CU-11, Non-Maturity Shares and Balance Sheet Risk.
Workpapers & Resources
- 03-CU-11, Non-Maturity Shares and Balance Sheet Risk
- National Economic Research Associates, The Evaluation of Credit Union Non-Maturity Deposits (September 10, 2001)
Last updated September 14, 2016
