Key Factors to Consider When Valuing and Modeling Non-Maturity Deposits

January 2019

Matthew Goldberg, Principal & Sudip Chatterjee, Managing Director, Valuation & Business Analytics Practice
 
Over the last decade, bankers have been enjoying the benefits of low funding costs with virtually no changes in the Federal Reserve’s monetary policies. As a result, all the complexities with modeling deposits in a changing interest rate environment were not on anyone’s mind. However, the environment has changed; understanding the value of non-maturity deposits is essential in calculating a financial institution’s value. Banks that understand the dramatic forces that are impacting business take appropriate steps to combat them and emerge as market leaders. 
 
Non-maturity deposits (‘NMDs’), such as retail savings, interest and non-interest-bearing checking and money-market accounts have no stated maturities. Thus, depositors can withdraw their funds at any time without any penalty. Modeling this early redemption option can be very challenging, but very rewarding, since NMDs are typically viewed as one of the most stable sources of funding for banks’ assets. Banks have other means of funding at their disposal, such as FHLB advances or the repo market but holding a substantial proportion of core deposits allows easy access to a stable and cheaper source of funding. Due to their short-term maturity and repricing nature, these instruments have relatively low interest rate risk which makes them more attractive.
 
The following five factors should be considered while modeling the behavior of deposits as market rates rise:
  • New Basel III Liquidity Coverage Ratio (LCR) and the Net Stable Funding Ratio (NSFR) require a detailed understanding of NMD behavior;
  • The uncertainty of balance growth in the system since the beginning of quantitative easing and how much of this growth will migrate to other investments as the economy improves;
  • Since 2008, there has been no significant change in short-term interest rates. As a result, deposit rates remained very low creating a data constraint over this period. Using this data to do any regression-based analysis will be greatly improvised;
  • At times, stress testing assumptions about the stability of a deposit base can be a bit harsh.  To mitigate such model risks, it is imperative to run several sensitivity and scenario analyses;
  • Technological invention in the market such as the advent of digital banking.
 
Deposit Pricing
In terms of modeling deposit rates and balances for NMDs there are three primary components to consider. 
  1. Interest rate models for interest-bearing NMDs
  2. Expected future trends on the balances for NMDs
  3. Servicing costs associated with NMDs
 
It has been proven over time that a change in deposit rates have a strong relationship to the changes in the short-term market rates such as Fed Funds or Treasury-bill rates, which is defined as beta:
 = change in deposit rate/change in market rate.
 
Different types of accounts are likely to show different repricing behavior and therefore should be analyzed separately. Historically, business accounts tend to be more rate sensitive than personal savings accounts, and money-market accounts are more sensitive than ordinary checking accounts. In order to model the deposit rates, some models use a lag where changes in deposit rates tend to occur after a certain amount of time following the short-term market rates change. It should be noted that deposit rate modeling can include other factors beyond a single rate beta. Apart from using lag adjusted betas, some of the sophisticated models use various other factors. These additional factors include multiple betas under rising and falling rate environments and peer group deposit rate setting behavior data. Modelers should also consider factors that might cause future repricing patterns to differ from historical observations, such as competing rates that make depositors switch accounts using online banking.
 
The second factor to consider is estimating the decay rate of existing deposit balances and the growth rate of new relationships. 
 
Decay Rate t+1 = Deposit Balance t0 + Projected growth – Runoff t0
 
Predicting deposit runoff amount depends on a wide range of variables such as market segments and individual customer behavior. It also varies between larger financial institutions where Treasury functions and other borrowing transactions are an important factor; whereas in a smaller institution, factors such as interest payment and various credit related payments are more important to consider while forecasting runoffs.  
 
A time-series data from a pool of accounts can be used to estimate decay rate of an account type.  Where the historical relationship between bank’s deposit rates and balances can be used as an indicator to estimate decay rates. Assuming, if a bank pays an above market rate, it will attract more deposits but at a thinner margin where as if it pays below market rate, it will widen margin at the expense of shorter life.
 
The third and final factor to consider is the cost of servicing the deposit base, which is a non-interest operational cost associated with these accounts. Typical examples of the servicing cost are not limited to the customer-facing activities, but also back-office activities such as maintaining controls of these accounts or transaction processing costs. 
 
The next step involves estimating the deposit cash flows under a forecasted path of short-term interest rates and then calculating the present value of this cash flow by discounting back to the present value using corresponding monthly Treasury rates. The sum of all the discounted cash flows is the present value of the NMDs.
 
Conclusion
Valuing non-maturity deposits is a culmination of multiple sub-models and the assumptions made around calculating betas, decay rates, servicing costs, and discount rates. It is recommended in order to gain accuracy with valuing these instruments as well as advised to incorporate multiple future interest rate paths since a single rate path may give biased results. To avoid any erroneous deposit value, some models consider using a rate floor and other asymmetric events while projecting rate paths. To understand the behavior of the weighted average life (WAL) of these instruments, it is important to run a sensitivity analysis around each of the primary assumptions. This will not only help with improving the quality of the assumptions, it will also help with performing attribution analysis of the banks’ net interest margin between market risk, credit risk, and Treasury to deposit spread.
 
How Can BDO Help
BDO understands these implications and recognizes both the opportunities and challenges of the current environment. Our years of experience with Asset/Liability Management and use of several third-party vendor tools puts us in a powerful position to advise our clients throughout the entire lifecycle of deposit modeling. Unlike the 2001 environment when rates rose from a recessionary period, the current market conditions are far more complex, and it requires proper planning, forecasting, and pricing mechanisms in place to be ahead of the competition.