REQUIRED READING: In recent years, the majority of servicers that hedged the interest-rate risk of their mortgage servicing right (MSR) portfolios have done exceedingly well. The combination of hedging in a declining-interest-rate environment and positive model error, in the form of slower-than-expected prepayment speeds, has been a great benefit to servicers deploying a hedge strategy since 2007.
In fact, in the wake of the financial and mortgage market crisis, even while regulatory changes and litigation issues are rattling servicing operations and derivative markets are being overhauled, financial risk managers in the servicing business have continued producing better results than have been seen in decades.
So where does this put the servicing industry as a whole? It bifurcates the industry into those who actively manage MSR risk and those who wish they had.
However, this favorable environment will not continue indefinitely. The cost of running a risk management program on the MSR asset eventually will return to negative levels. Hedging does cost money in the long run, but the trade-off is stability of the asset value and an opportunity to increase the odds of earning the return that a servicer sets to make when the servicing was created.
The concept is straightforward: When this hedging cost is built into the production and capitalization return requirements, firms should be able to run a vastly more stable business over a much larger variety of economic and market cycles. This cost typically is between 1% and 2% of the return on the asset. The question is whether this is attainable for the small to midsize servicers, rather than for only the largest servicers.
The fact of the matter is, many servicers outside of the top 20, whether a privately held mortgage company or a regional Federal Deposit Insurance Corp.-backed bank, long ago threw in the towel regarding financial risk management of the MSR asset. Yet the asset can easily lose 30% or more of its value for a 1% change in market rates.
In light of this financial risk, and given the progress of the financial services industry in terms of advances in technology and business sophistication, is it still a realistic assumption that a servicer can ignore the financial implications of the asset and remain a viable long-term player in this volatile business? Have servicers not advanced enough to contemplate this issue with sufficient resources and tackle the risk in a sound, cost-effective manner?
Those servicers that do not actively manage this risk have various reasons for not doing so, most of which stem from the perceived cost and complexity of the process, which is historically understandable. But as the financial markets have continued to refine risk management processes and improve the breadth and depth of the services offered, the reality is that taking hold of a servicer's financial stability is not as far-fetched as many might imagine.
With that said, in today's markets, most risk management programs in place remains at the level of the top-20 servicers. These firms have amassed such significant amounts of MSRs that there clearly is a need to have risk management programs in place. The most obvious reasons are to manage the impact to their quarterly earnings, add resolve to investor concerns over their financial stability, and reassure regulators and agencies that their businesses are well protected against financial hardship on the borrowers being serviced.
For whichever reason might have spurred these institutions to create MSR risk management programs, the underlying purpose remains the same: to create a stable business model, better understand risks to that model in terms of profitability and lock in returns of the asset across a wider array of possible markets.
With this in mind, the concept of MSR risk management is as crucial for small servicers as it is for the top-20 servicers. Even when not actively managing the risk through hedging processes, small servicers must still understand the risk. If a servicer claims the natural business hedge of the mortgage model is working, the servicer should be able to support the claim with verifiable analysis and results.
Looking across the entire financial services sector, there are virtually no markets where such a significant financial risk is not at least measured and understood by market participants. If it measures the financial risk of the MSR asset and then chooses to not hedge the risk, the servicer has taken a stance on the risk, and this position should be clearly communicated to management and stakeholders rather than go unreported or poorly understood.
This concept is also relevant to the production pricing process for small and midsize servicers. When the servicer retains the ability to hold additional units of MSRs without being constrained by concerns of the size of the risk, it opens up the best-execution process to allow for stronger internal retention numbers and increase the profitability of the origination engine.
MSR risk management is often derived from two main functions: measuring interest-rate risk and offsetting that risk, or hedging. Historically, defining MSR risk was so cumbersome and difficult that only the largest servicers would contemplate the resources necessary for the task – Ph.D.s in back rooms devising statistical models and postulating economic theories that would attempt to isolate the manageable interest-rate risk.
In many ways, the issue of understanding MSR risk is as important and complicated as always. The difference today is that technology is beginning to catch up with modeling needs to support sound risk management practices. Now, servicers can use desktop computer systems to run thousands of iterations of risk, providing a wealth of information that was not available to smaller servicers in the past. This has allowed all servicers to better understand the risk of their MSR portfolios.
Once the risks are quantified, the hedging process begins. Historically, teams of interest-rate strategists and traders would enter into securities and derivative transactions, offsetting the risk through cross-hedging. This is the strategy of hedging with a different
type of instrument than defines the underlying risk being managed, and relying on an expectation of the future relationships to minimize ongoing asset volatility.
The issue is that when these instruments move in unanticipated ways relative to each other, poor results can materialize, even with risk management programs in place. The industry has seen numerous cases where servicing market participants have either exited the business outright or even suffered an untimely demise at the hands of faulty correlation models and difficulty managing this risk. Today, MSR risk managers have become more astute in the modeling process and have eliminated much of this cross-hedging risk through better assessments of the asset risk. Running the asset models relative to each specific model input allows the hedgers to use such instruments and avoid many of the cross-hedging pitfalls suffered in the past.
The most widely cited reason for ignoring the financial risk is, of course, the natural hedge of the mortgage business model: Expected increases in production have an inverse financial relationship to the value of the servicing rights portfolio. This effect has been referenced often but quantified rarely, and for good reason. The relationship does exist, but it is much less dependable than is typically presumed. The timing of the increasing production business and the impact of the servicing value change are not aligned, which, for publicly traded companies in particular, can cause significant problems that a risk management program would contemplate.
A more significant issue for smaller servicers is the balanced exposures between the two business lines. Because each MSR has a particular risk profile that is not linear (more downside risk than upside benefit) and is specific to characteristics of the underlying loan, such as the note rate relative to a par market rate, the inflection point of net financial risk tends to match up very poorly. This relationship also fluctuates over time, as the origination business will tend to spike up much higher when new lows in interest rates are reached versus a simple retracement of prior lows.
In both instances, the MSR asset will fall in value. Furthermore, if a firm were to quantify this natural hedge and remain risk-neutral, it would likely require the acquisition and disposition of servicing at less-than-optimal price levels. For a typical servicer, the benefit of increased production in a lower-rate environment might offset a portion of the financial risk of an MSR portfolio but almost surely will not be well balanced to leave the servicer at an economic break-even.
Studying the relationship of new production volume relative to rates can add insight to this issue. Looking at the last 10 years of market data on the refinance index, which is a good proxy for applications and subsequent originations, we find that the change in rate only explains approximately 20% of the change in the refinance index, whereas the change in the value of a servicing right is more than 85% related to changes in market rates. We see that at periods, this relationship coincides well, but these periods are short-lived and typically are seen when rates are reaching new lows rather than fluctuating at random, as we would typically presume.
If the need for stability of the servicing asset is desired, a fundamental set of principles should adhere to any well-rounded risk management program. From the aspect of management of the firm, financial goals of the program should be defined prior to engaging in risk management. Without this set of goals, the program can migrate into unintended strategies and results over time. Goals should be in the form of loss tolerances and related to the underlying health of the financial statements – what the firm can afford to lose.
A system of checks should be put in place that verifies that what the asset model is projecting is reasonably correct versus actual experience over time. This is most importantly tracked on prepayment speeds. Because prepayment speeds determine upwards of 80% of the financial risk of servicing, it is vital that the modeled speeds across interest rates be reasonably accurate.
Next, a commonsense approach to the asset-modeling process should always be used. This does not mean modelers should not strive for accurate economic representations of reality. When the models become too complex to be intuitively explained, something is bound to go wrong.
Last, the hedging decision should use instruments that define the risk of the asset value.
In summary, until the agencies and other interested parties can decide on an alternate compensation mechanism – one that does not expose the MSR asset to market rates – the financial risk of servicing is real and should not be taken for granted. If servicers decide active management of the risk is not preferred, they should do it with full knowledge of the risks being taken and clearly communicate those risks. But if an active approach is preferred, they must budget for appropriate returns net of the cost of a risk management program and maintain a commonsense approach to the risk management strategy.
Brian Stewart is a managing director with MountainView Risk Advisors. He is based in the firm's Denver office and can be reached at (303) 633-4715 or bestewart@mvrisk.com.
