Market Value – It’s Not Just a Regulatory Compliance Tool!
By: Tom Farin
I can hear your comments already. “MissFed is a thrift case. All that market value measurement stuff is irrelevant to me. I’m a bank (or credit union). My regulator doesn’t ask me to evaluate market value risk. If they don’t think I need to evaluate market value risk, I don’t see any need to worry about that stuff.” “The only measure of interest rate risk I’m going to worry about is what a rate shock will do to my income. That’s what my board worries about. It’s what my examiners worry about. And it’s what my stockholders worry about.”
Evaluating Market Value Risk – A Historical Perspective
There is no question that the adoption of market value oriented measurement systems have come about as a result of regulatory initiatives. OTS led the charge in the late 1980’s with the original TB-13. It required (and continues to require) that thrifts evaluate the effect of immediate and permanent rate shocks on the Market Value of Portfolio Equity (MVPE). FASB followed with FAS 107 and FAS 115. FAS 107 requires a footnote on financial statements divulging the true market value of a financial institution’s assets and liabilities. FAS 115 requires institutions to mark-to-market investments designated as ‘available for sale’ and reflect the effect of market value changes in its financial statements. FFIEC waded in with stress test requirements that measured whether an investment was low or high risk based on the effect of the shocks on their market value. And the bank and thrift regulators have encouraged (and in some cases required) banks and credit unions that are large enough and risky enough to use market value measurement systems to evaluate interest rate risk.
The Alphabet Soup of Market Value Terminology
Along the way, regulators have used a variety of terms to describe what they are asking you to measure. It seems like every time they hire a new Ph.D. as their interest rate risk guru, the new holder of the interest rate risk mantle feels compelled to invent a new term to describe market value risk. OTS began with MVPE then changed the term to Net Portfolio Value (NPV). The bank regulators countered by calling it Economic Value of Equity (EVE). The present hot term in academic circles is Value At Risk (VAR). And they spend a great deal of time and energy explaining the merits of their market value brand. Sort of like regular gasoline. I wonder whether the car can tell the difference. They all cringe when we refer to their hot technique as ‘liquidation value’. We all learned in accounting class that capital is the difference between an organization’s assets and its liabilities.
If so, the difference between the market value of an institution’s assets and liabilities is the market value of its capital account. In other words, its what a regulator would expect to have left over after it liquidated an institution’s assets and used the proceeds to divest its liabilities. Call it what you want. If it walks like a duck and quacks like a duck …
Why Market Value Measurement Systems Were Imposed
It’s no coincidence that the strongest market value risk regulatory requirements evolved in the thrift industry. Thrift balance sheets are heavily laden with mortgages and mortgage-backed securities. And mortgage products generally have longer durations and more option risk than virtually any other asset commonly appearing on financial institution balance sheets. While the fact that an instrument has a long duration is not necessarily bad, it creates the potential for a significant mismatch between the duration of an institution’s assets and its funding. This is a particular concern when the vast majority of funding provided by consumers is short-term in nature. Duration mismatches can create considerable interest rate risk, resulting in significant changes in income and liquidation value with changes in rates. The duration problem is compounded by the fact that options imbedded in mortgage products can result in significant cash flow changes with changes in rates. Durations of fixed-rate mortgages increase in rising rate environments as prepayments slow. Durations of fixed-rate mortgages shorten in falling rate environments as consumers prepay their loans to take advantage of lower rates. Adjustable-rate mortgages, have changes in principal and cash flows resulting from the effect of changes in rates on prepayments. Holders are also faced with the effect of annual and lifetime caps. These additional imbedded options may limit how quickly ARM rates may respond to changes in interest rates. As banks and credit unions became more heavily involved in mortgage products (loans and securities) during the 1990’s, regulatory agencies became more concerned over the levels of interest rate risk and option risk. While bank and credit union regulators haven’t generally imposed OTS TB-13 style interest rate risk requirements, shops with larger concentrations of mortgage product are being pushed in that direction. “But Farin, it makes no sense to manage an ongoing institution around changes brought on by rate shocks to its liquidation value. The concept is only relevant to regulators. If my regulator doesn’t require that I run this kind of analysis, it makes no sense to waste my time doing it. I’m going to spend my time learning to do income simulation effectively!”
What Income Simulation Won’t Tell You
I’ve laid out the regulatory framework, but haven’t really dealt with whether market value measurement systems are a legitimate management tool. For reasons listed earlier in the article, it makes sense for management to worry about the effect of changes in rates on income.
But here’s the rub. Many institutions simulate the effect of rate shocks on income over a oneyear planning horizon. Some may model over as many as three years. Others may go out as far as five years. But the mortgage product we are placing in our portfolio generally has a term structure much longer than the longest income simulation horizon we are willing or able to run. In some cases, imbedded options won’t be triggered during the planning horizon we are modeling. The cap risk in a 5-year lock ARM cannot be felt for five years. A 10-year FHLB Advance with a 5-year call option will not be called for at least five years. Tom Parliment, in the companion article relating to the MissFed case, is using the 10/5 callable advance as a funding source. If we model his strategy over a three-year horizon, the call option will never be in the money. At its 6.46% rate, the callable advance will outperform both the 10year bullet advance (7.36%) and the 5 year bullet advance (7.25%) regardless of the rate environment for planning horizons of up to five years.
A ‘Simple’ Example
To illustrate the point, let’s assume MissFed decides to fund their growth in 5/1 ARMS with 10/5 callable advances. The ARMs have a 2% annual and 6% lifetime cap and a fully indexed rate of 8.70% at origination. To keep the analysis simple, we’ll assume all originations occur immediately and total $50 million. Origination rate is 7.25%. Figure 1 illustrates the results of modeling this transaction over a variety of planning horizons.
Term & Environment 0-5 Years +200 0 -200 5-10 Years +200 0 -200 10+ Years +200 0 -200
5/1 ARM 7.25% 7.25% 5.25% 9.25% 7.25% 5.25% 9.25% 7.25% 5.25%
10/5 Advance 6.46% 6.46% 6.46% 8.46% 6.46% 6.46% 8.46% 6.46% 4.46%
Spread 0.79% 0.79% -1.21% 0.79% 0.79% -1.21% 0.79% 0.79% 0.79%
Option Action
Customer Finds Better Teaser FHLB Calls Debt Customer Re-Teasers Customer Re-Teasers Customer Re-Teasers Customer Re-Teasers, Advance Matures Customer Re-Teasers, Advance matures
For planning horizons of up to five years in the flat and rising rate environmenst, with an asset yield of 7.25% and a funding cost of 6.46%, this transaction has a 79 bp positive spread. In this example, rate environments of +200 bp, 0 bp, and –200 bp are occurring in the initial 5-year period. In the falling rate environment, the customer refinances to take advantage of a much lower teaser (now 5.25%). The advance remains at 6.46%, a spread of –1.21%. Note that depending on the timing of the rate drop, this refinance could happen very early in the five-year period. Until then the spread is a positive 0.79%.
Between years 5 and 10, a variety of results occur. In the rising rate environment, the institution resets the mortgage rate to 9.25% after five years, the full increase allowed by the annual cap. The FHLB calls its debt after five years. The resulting 0.79% spread is dependent on the assumption that callable 10/5 advances will continue to be available in five years at the same spread to bullets we saw at the time this transaction was consummated. In the flat rate environment, the customer refinances to obtain a new 5/1 ARM to avoid paying the fully indexed rate of 8.70%. Spread is also 0.79%. In the falling rate environment, the negative spread (-1.21%) results from the resulting refinance activity – which continues for an additional five years. Beyond year 10, both the ARM and the advance have adjusted to the new level of prevailing market rates. Assuming the same products remain available at the same relative level of rates, the 79 bp spread will exist for the remaining life of the products. However, because the ARMS are amortizing and prepaying throughout their life, less than 40% of their original principal will be left after 10 years. This minimizes the chance that after 10 years we will make up enough income to offset what was lost in the first 10 years in the falling rate environment. Simple? Try going through this example with your board. Then give them a quiz to see if they understood anything that you said! At 79 bp, the initial spread on this transaction is very marginal, especially considering the credit risk and servicing costs on the product. But needing income, MissFed might go forward with this transaction because the alternative may be zero loan growth. And 79 bp is better than nothing. They are more likely to go forward with this transaction if they use income simulation as their sole interest rate risk management tool. Depending on the length of their planning horizon and the timing of the rate shock, they will not feel the full effect of the falling rate environment’s negative spread. If their planning horizon is short enough (1 year) and the rate decline is gradual (200 bp over 1 year), they will see virtually no effect of the rate drop on income.
Market Value Risk – An Estimate Of The Effect of Rate Shocks On Income Over The Asset’s Life
While a detailed discussion of the methodologies in market value calculations is beyond the scope of this article, it is instructive to summarize these calculations. To value a financial instrument using discounted cash flow methodology, we (1) project its principal and interest cash flows, then (2) use a discount rate to mark those cash flows to market. To determine the effect of rate shocks on this value, we (1) re-project the cash flows for the new rate environment, then (2) use the discount rate (plus or minus the rate shock) to mark the cash flows to market in the new rate environment. The change in market value caused by the rate shock is the market value sensitivity of the instrument. A good simulation model will take the effect of imbedded options into consideration when projecting (and re-projecting) cash flows. The discount rate is the fair market rate for instruments with similar levels of risk and cost. Note that unlike a short-term income simulation, market value calculations consider all the cash flows thrown off by an instrument during its life! Let’s take a look at the effect of modeling this transaction using market value methodology. In this example, the book value of both sides of the transaction is $50 million. On a mark-to-market
basis in the current (0 bp) rate environment, the ARM’s value is $48.1 million. Why is it below the book value of $50 million? For reasons we’ll get into in a companion article, this ARM is poorly priced relative to its risks and costs. MissFed would pay for its poor pricing with a loss if they decided to sell this product in the market. The advance also has a market value in the current (0 bp) environment of less than par ($47.16 million). It enjoys the fact it is being discounted using the 5-year bullet advance rate of 7.25%. If you book this transaction today, your institution’s liquidation value on a mark-to-market basis increases by $980 thousand. This reflects the model’s estimated net present value of the transaction’s potential contribution to real capital over the life of the transaction.
Book Value 5/1 ARM 10/5 Callable Portfolio Equity 50,000 50,000 -200 bp 50,647 54,658 (4,011) 0 bp 48,142 47,162 980 +200 bp 45,286 44,414 872
The option risk in this transaction is clearly illustrated in the rate shock results. In a rising rate environment, an asset with a duration of nearly five years (the lock period on the ARM) shows about the same drop in market value shown by the nearly 5 year duration funding source. Keep in mind that the FHLB will call the advance at the end of 5 years because they will be able to reinvest the funds at a significantly higher rate. The effect of the rate shock on portfolio equity is nominal ($108 thousand or 0.22% of the transaction.) In the falling rate environment, the effect of prepayments on the mortgage shortens its duration. As a result, the gain in market value from the favorable rate shock ($2.5 million) is less than the pain from the unfavorable rate shock (-$2.8 million). And my falling rate environment results use an OTS prepayment model that fails to anticipate the ARM teaser rate jumping that we feel is likely to occur in the next falling rate environment. See the companion article for details. Frankly, I don’t see these ARMs ever going above par in falling rate environments. Keep in mind that when mortgages prepay, they prepay at par! The ARMS are funded by a nearly 10-year duration liability in the falling rate environment. No, you can bet the FHLB won’t call your 10-year debt five years early to take advantage of the opportunity to reinvest at lower rates! As a result, the gain in the value of the liability exceeds the gain in the value of the asset by nearly $5 million. Portfolio Equity declines from a positive $980 thousand to negative $4.01 million! Here’s my question to your board. Would you trade the potential for an increase in your capital of just under $1 million in a flat or rising rate environment for a potential loss of $4 million in a falling rate environment? Your board would need to consider the probability of a drop in rates to be pretty low before taking this bet.
Portfolio Equity – A long-Term Income Effect?
I can hear my partner Tom Parliment now. “Farin, you can’t do that -- assume changes in portfolio equity are equivalent to the long-term income effect of a strategy.”
He’s right. My assumption is heroic for a number of reasons: • The tax effect is not considered. • Alternative investment options are not considered. • Additional operating expenses associated with the strategy are not considered. • Assumptions as to reinvestment of cash flows were not taken into consideration. • Yada, yada, yada, yada. Income simulators, do it your way if you wish. Prepare your one-year plans that fail to consider the long-term effects of your decisions. What the heck! With staff turnover rates in the financial services industry lately, you may not be there to see the long-term effect of your decisions anyway. Or if you prefer, develop a long-term plan modeling your strategy with your simulation model. Of course, to fully model the 5/1 ARMs, you’d need to simulate their performance over 30 years. Oh, to heck with the tail you say? To properly model the 10/5 Advance you’d need to go at least 10 years. That’s still a $50 million tail out there in a flat or falling rate environment between years 5 and 10. And please agonize over how you plan on reinvesting cash flows coming off your ARM portfolio – 7, 8, 9, and 10 years from now. Then go ahead and project your marginal costs of servicing mortgages 8 years from now. Let’s see … we’re in a declining cost industry … And be sure to consider all your reinvestment alternatives and all your current alternative investment and loan options. In about five years, you’ll have your plan and a whole boatload of paper reports to back it up. Then watch me tear your assumptions to shreds, like I’m sure my partner is doing to mine. And of course, by the time you are done, you’ll be well past the time frame needed to execute this decision. No, I’ll take my rough and tough way of estimating the long-term risk-return tradeoffs of a strategy using my market value tool. I turned around the analysis used in this article in a few minutes. And I have a report showing the potential long-term effect of the strategy to supplement my short-term income simulation modeling. I’ll use both in presenting my recommendation to the board. And I’ll tuck it all away in my ALCO minutes so if anyone questions the analysis later, I can show the documentation behind my decisions. And I’ll win the majority of those arguments because market value is: • The primary tool used by the security markets in valuing and stress testing financial instruments. • The primary tool used by the secondary markets in purchasing your loans. • The primary tool used by those deciding how to value the branch you are attempting to buy or sell. • The primary tool used by regulators in evaluating interest rate risk in complex financial instruments. • The primary tool used to value your franchise when your stockholders decide to sell your butt!
Moving Toward Dynamic Analysis
One of the major shortcomings of regulatory interest rate risk analysis is the regulator is limited to modeling interest rate risk in existing balance sheets. Much of the analysis performed by outside firms providing interest rate risk analysis services is vulnerable to the same criticism. No matter how fancy their methodology, they are limited to telling you how much interest rate risk existed in an old balance sheet. Static interest rate risk analysis tells you how much trouble you are in, but is of limited use in helping you identify the most cost-effective method to solve the problem. You may have noticed that the interest rate risk analysis presented in Tom Parliament’s article evaluates the interest rate risk in a proposed strategy. In fact, interest rate risk analysis tools are used to look at multiple proposed strategies. The highest and best use of simulation technology is in the hands of a manager using the model to evaluate the risk/return tradeoffs associated with proposed interest rate risk solutions. Dynamic interest rate risk analysis can include both net income simulation and the analysis of the effect of rate shocks on liquidation value. In the short-term solution to the MissFed case, TP looks at two alternative funding strategies. We evaluated those strategies in the following way.
Income Simulation
We implemented the funding solution in the first month of the three-year forecast. We then ran income simulations under three different rate environments (-200bp, 0bp, +200bp). Rate shocks unwound gradually over the first year of the forecast. When unwinding gradual rate shocks, you need to model strategies over multiple years. We ran ours out over three years. By modeling a base strategy (no funding changes) against our two alternative funding solutions we were able to see the relative effects of the two alternatives in comparison to the base strategy over a three-year horizon. We would choose the strategy that has the least negative effect on return (in the flat rate environment) while keeping the risk (income variances in the other two rate environments) within our policy limits.
Market Value Testing
The long-term effect was analyzed using market value testing. A primary concern in the market value area was getting MissFed’s liquidation value variance within OTS tolerance levels by 3/31/00. Our as-of date for our input data was 2/28/00. We ran a one-month simulation, implementing our funding changes in the first month of the forecast (March). We then told the model to apply stress tests to the ending balance sheet (3/31/00). In doing so, we could test the potential effect of the strategy on market value risk before implementing the strategy in the real world. Assuming our model accurately emulates the calculations done by the OTS model, we would see our 3/31 results before OTS and before we implement the strategy in the real world.
In addition to comparing our two alternatives, we were also able to use the model to determine the optimum size of the transactions. We were looking for the solution that would get us within the OTS tolerance limits at the lowest cost to income. So we fiddled with different sized transactions until we achieved those goals. As TP’s article indicates, it required $75 million of 10/5 advances, or $30 million of the 30 year amortizing advances with 5-year prepayment options to meet these objectives. Once we had settled on transaction sizes, comparing the results of the two strategies allowed us to identify the strategy with the best risk/return tradeoffs. As soon as we identified the strategy with the best risk/return tradeoffs, all it would have taken is a phone call to FHLB Dallas to implement the strategy we had chosen. Were the results worth the effort? The net income difference between the two proposed solutions was in the low sixfigure range per year. With that kind of money you can pay for care and feeding of a pretty fancy simulation model. And this is the impact of just a single decision.
Putting It All To Work
And those of you that own a competent simulation model have the ability to evaluate the effect of rate shocks on EVE, MVPE, NPV, or whatever you want to call your liquidation value. If you’re not sure how to go about doing it, you may want to check out the resources available through our Web site, or consider attending one of our consulting workshops.