By Mark A. Haberland, Darling Consulting Group
There’s nothing new about banks’ use of asset-liability management models to manage risk. But after six years mired in the worst economic crisis of our generation, examiners have renewed their focus on how banks measure and manage interest rate risk and how they support the assumptions that drive the results.
How strong is your bank’s approach? The ultimate utility of the ALM model depends on three factors:
• Data: How strong is the quality of the data inputs?
• Assumptions: How reasonable are the assumptions behind your models?
• Decision making: How do the data and the assumptions drive the direction the bank takes?
As I will discuss, banks must be careful how they let ALM tests and simulations influence their decisions. Some efforts, driven by regulatory requests, may skew results unnecessarily.
Washington steps up ALM emphasis
At the end of 2013, FDIC (FIL-46-2013) and OCC (on a conference call) made it clear that their exams would be focusing on banks’ processes for identifying and managing risk, especially in anticipation of the eventual rise in interest rates, and the regulators’ preference for “bank-specific” model assumptions.
The quality of data inputs is an important aspect of the modeling process. It sets a proper foundation for account behavior in both normal and stressful rate environments. A review of historical data is the best way to support the very assumptions that drive the model results: cash flow reinvestment, prepayment activity, and deposit sensitivity.
Yet, it is critical for banks to continually review and challenge these assumptions to ensure they represent each institution’s current operating environment.
Developing and supporting key assumptions
“Euclid taught me that without assumptions there is no proof. Therefore, in any argument, examine the assumptions.”—Eric Temple Bell (Scottish mathematician)
Hundreds of assumptions go into every ALM model—yet only a small percentage of those assumptions have any material impact on results.
For the majority of community banks, the assumptions that apply to the deposit base (sensitivity betas and average lives) have the most profound influence on the risk profile—both earnings and value at risk.
The methods banks use to develop these assumptions have been in the regulatory crosshairs in recent exams. The comment has changed from, “Are these assumptions supported?” to, “Show us how you support these assumptions.”
If you haven’t been asked to prepare a deposit analysis, odds are very high that regulators will do so during your next exam. The most effective way to provide the support for these most critical assumptions is to have a historical deposit study prepared. But a deposit study taken at face value is worth little. What is most often missing is the logical—yet often overlooked—next step of applying qualitative analyses to these results.
Numbers can lie
It is critical to look at the results and ask: “Do these numbers make sense? Knowing what we know, is this how we will manage our deposit base in the years to come?”
Then the bank must determine whether the study or model assumptions (or some combination) are truly most reflective of their business plan.
And it is critical that management and ALCO understand how those ultimate results were derived, as they will need to explain them to the examiners and provide documentation regarding how they were developed.
Deposit studies are not an exercise in futility—not just something to be done simply to appease the examiners. We have generally found that absent a study (or other documented support), banks tend to apply more conservative assumptions which, by definition, shows an increased exposure to rising rates.
As we continue to wallow in this low-rate environment, and margins come under pressure, banks are looking for ways to survive. They want the ability to put on longer-term assets to get some spread, yet a liability-sensitive profile makes that a difficult argument to win with the examiners.
A study that supports lower betas and longer average lives allows for the ability to extend assets and get more yield. Should the study support the conservative assumptions, then at least you are confident in the assumptions and can look elsewhere for income opportunities.
Stress testing provides confidence
Once you have the support you need for your deposit assumptions, you will have more confidence in your model results and, therefore, the decisions you make at ALCO. But these assumptions impact more than just the net interest income (NII) and economic value of equity (EVE) models: Don’t get interest rate risk management confused with balance sheet management.
The assumptions that drive NII and EVE model results also impact your liquidity management and contingency planning processes. Stress testing provides guidelines for potential exposure levels and back-testing the model provides additional confidence in the assumptions and model results.
The importance of accurate and supportable assumptions cannot be overstated when developing an ALM model. However, assumptions are just that—something estimated, and not definite.
Introduce scenario/stress testing to your ALM process. Taking your key assumptions and running alternative simulations that show the impact of different or stressful factors will provide the bank with useful information regarding potential exposure to the balance sheet and what levels of stress have the most impact on results.
Banks could examine the impact of a 15%-20% runoff of non-maturity deposits as rates rise, with those deposits either migrating into higher-cost accounts or leaving the bank altogether. This could have a profound impact on NII sensitivity as well as raise potential liquidity concerns—the importance of a strong ALCO process that examines all balance sheet risks is vital to the development and implementation of key strategies.
Knowing where we are today relative to policy and what our potential exposures are under current and stressful scenarios helps bring into focus the direction in which ALCO needs to move and how best to get to the destination.
Armed with the confidence of reliable model results, the ALCO has the tools it needs to drive critical strategic decision making. And these tools are continually sharpened by adding the knowledge of potential exposures facing the bank through additional scenario and stress testing, as well as the updating of prior studies.
As is often the case, there can be many tools that would accomplish the task but, as an ALM manager, which one(s) should you focus on when the time comes to execute on a transaction or to make your case with the examiners?
Right tool for the job
As ALM practitioners, Darling Consulting Group has long been proponents of the NII simulation as the primary tool for measuring and managing interest rate risk. By examining a set of core scenarios (rate ramps) and alternative simulation runs (twists, etc.), bankers can obtain the most accurate interpretation of earnings at risk under reasonable rate scenarios.
By focusing too much attention on instantaneous and permanent rate shocks or Economic Value of Equity (EVE) results, ALCOs can inadvertently manage to overstated or, worse yet, non-existent exposure.
Shocks and EVE introduce unrealistic aspects into the modeling process that can result in poor decision making.
For example, rates do not “shock” up or down 300 or 400 basis points and remain at those levels for an extended period of time. That remains true even going back to the most volatile period, when Fed Chairman Paul Volcker was running the central bank.
And the presumption that EVE is a “liquidation” of the bank is contrary to the belief that it will continue operating and cash flow will be replaced at current market rates.
And while we are on the subject of the EVE analysis, this is the tool where inaccurate model assumptions can cause serious unintended consequences. The EVE is a primary risk management tool reviewed by field examiners today, and its results are dramatically impacted by the assumptions on non-maturity deposits.
However, an EVE built on erroneous or unsupportable assumptions runs the risk of providing an inaccurate representation of rising rate exposures. This can limit the bank’s ability to most effectively manage ongoing margin/earnings compression in the current rate environment through extending assets, shortening liabilities or using derivatives; especially if the EVE suggests an overly conservative exposure to rising rates.
Sifting through data to find information
Yet, even when focusing ALCO’s attention on the NII results, there are often so many scenarios that are run. How does the bank go about determining which scenarios to use for decision making?
There is no one simple answer. Much will hinge upon where we find ourselves in the rate cycle.
For example, it is always prudent to examine the results assuming a static environment (balance sheet and rates remain constant). This provides a benchmark for earnings projections upon which to base all other exposures.
Today, even as we continue to be mired in this historically low rate environment, we find ourselves saying, “It can’t possibly get worse, can it?”
Truth be told, should the bond market rally and long rates fall, that is a worst-case scenario for many banks that will see increased margin compression with very little, if any, room for additional funding cost relief.
Therefore, looking at a strategy’s impact should the long end fall is also good practice.
The expectation, and regulatory concern, is that rates will, at some point, rise.
So it only makes sense for ALCOs to look at the impact potential strategies will have during the eventual rise in rates. There are an infinite number of scenarios banks can look at for rates increasing—but which ones carry the most value?
It would make sense to look at the impact of a more aggressive increase in rate coupled with a more “realistic” movement (i.e. a twisting of the curve, similar to what was seen back in 2004, in anticipation of rising rates).
As banks struggle to maintain earnings and margins in today’s environment, the role of ALCO has never been more important. The importance of an accurate model and assumptions is at the forefront of regulatory exams and bankers’ minds.
Now is the time to ensure your process will provide you with the information you need to effectively run the bank and make the right decisions at ALCO at the right time. Do not underestimate the importance that assumptions have on results.
Poor assumptions lead to poor decision making.
But strong assumptions—and the use of the right tools to promote decision making—can help position your bank for success.
About the author
Mark Haberland is a Managing Director at Darling Consulting Group. In this role, Haberland works directly with financial institutions to help provide solutions for asset-liability management challenges. He provides support in the areas of liquidity risk management, ALM modeling, regulatory compliance and executive level education.