Q&A with Emily Hollis

Emily Hollis
November 24, 2009 | COMMENTS

Question: Should I use par values for all non-maturing deposits?

Par values recognize no economic gain or loss in the value of non-maturing deposits. This means that the model results are unaffected whether a credit union pays zero percent or five percent on its share draft accounts. Obviously, net interest margins (and subsequently net economic values) are grossly affected by such decisions.

Industry theory generally holds that if a financial institution attracts funds at lower rates than its borrowing costs it creates economic value in its balance sheet. Par values also embrace the assumption that your dividend rate will correlate 100 percent to overnight funds. In those rare instances of aggressive, high-dollar money market accounts, this indeed is applicable. Otherwise, this method is erroneous.

The greatest weakness of this method is that by grossly understating liability duration it causes the analysis to miss the danger of spread compression in a downward rate environment. We know that, in practice, non-maturing liabilities do have duration as accounts remain with an institution over time, regardless of the prevailing interest-rate environment and dividends paid. We also know that banks have paid significant premiums for low yielding funds.

We believe that the most critical determinant of value for a non-maturing deposit is the projected dividend rate paid by the credit union. Consider a fixed-rate bullet (a cash flow that has a single maturity and is void of any amortizations, puts, calls, etc., such as a CD) instrument that matures in 10 years. It has significant interest-rate risk (or price volatility). If we changed the fixed rate to a rate that adjusts monthly with the market, it has no interest-rate risk and its effective duration is zero. Therefore, if your institution has historically changed the interest rate on a share account in lockstep with the market, and the market is represented by the effective Fed funds rates, the analysis will show little sensitivity in economic value in different interest-rate scenarios, regardless of maturity or decay rate.

Therefore, the maturity used in the analysis is secondary as detailed in the explanation above. Using a decay rate of 15 percent, the remaining balances of share accounts at the end of five years would be 36 percent and at the end of 10 years remaining balances would be 14 percent. So extending the maturity to 15 years produces only minimal differences of economic value. This is because the present value of the balances is such a small percentage of the original and the maturity is so far out that the premium value is insignificant.

Although a statistical analysis gives verification to values, a simple analytical method is to model the dividends in an income simulation for each interest-rate scenario and then present value the cash flows to calculate the economic premiums.

Placing par values on all non-maturing liabilities provides equal modeling results between having 100 percent of funds in money market accounts paying six percent, or 100 percent in share drafts paying zero percent. Not only is this erroneous but such an assumption could be dangerous.

Emily Hollis is president of ALM First Financial Advisors in Dallas, Texas. Contact Hollis at 800-752-4628 or ehollis@almfirst.com.