30. August 2007 by Bob Jarmak.

Accounting for MSR’s (Mortgage Servicing Rights)

Specialized accounting for MSR’s began in the late 1970’s with the creation of GNMA MBS, which was the beginning of the secondary mortgage market. Prior to this, loans were traded on an individual basis outside of any formal, standardized market. The emergence and subsequent growth of the secondary mortgage market remains one of the most important finance events of the past 50 years.

The ability to reallocate capital flows has had beneficial effects for all involved. The conventional MBS market followed, several years later. This led to the development of FASB 65, the controlling standard for many years. FASB 65 dealt with many aspects, including the treatment of fees and the treatment of purchased loans.

It was common for large originators and sellers to purchase block loans from other originators and subsequently securitize them.This led to the unequal treatment of purchased loans and originated loans. It also led to the abuses of the early 1990’s which caused numerous bank failures, where banks were assigning unreasonably high values for future MSR’s.

FASB 122 was introduced to combat some of these abuses, as well as treating originated loans the same as purchased loans, from an accounting standpoint. Various iterations of this have been introduced since, up to the current FASB 156.

The principle remains consistent; mortgage servicers have a contract to receive future income.The only requirement is that they perform their servicing duties. This is a legally binding contract from typically a quasi-governmental agency which may be sold or transferred to another servicer. The only real risk of non-payment is early prepayment. As such, the present value of these future rights should flow through current period results.The alternative? If accounted for as cash income when received, the equity position is understated as this future income is not recognized.

For some, it’s not material, but regardless, it would still be an understatement, which is wrong. Investors/regulators would be forced to make their own calculations of this value, to properly value the institution. The “cost” of this is increased volatility as MSR’s are subject to substantial revaluation as loans prepay early. But, this revaluation just represents reality. Some find this treatment objectionable for a variety of reasons, not the least of which is this increased volatility.

Other objections include the cash/non cash components and the comparison between this treatment and other assets.Parenthetically, some with the largest objections have no problem with goodwill or other similar assets. Perhaps the best way to view this accounting is as an introduction to the coming fair value transition.Some will continue to object, strongly in many cases, but the future direction appears pretty clear, and changes will continue in this direction.

Why does fair value make sense? All financial valuation is based on a very simple concept. It is the present value of the expected future income streams. For non-financial assets, value is replaced by utility, which allows items such as art to be valued consistently. Banks are really pretty simple businesses. They borrow funds and they reinvest them.To maximize returns, is to maximize the spread between cost and earnings. The specific levels are generally a function of the local or regional market, while the spreads are relatively consistent among similar institutions.

The base value of the institution is the present value of the spread cash flow for as long a period of time as the spread is expected to continue at that level (in practice, a perpetuity). Although the amount of the spread will move within a range based on the overall yield curve shape, over time it will be relatively stable. To this, we can add the present value of other income streams, such as servicing revenue, fee income, etc., and subtract the present value of expenses. We can also adjust for other assets such as real estate.When we’re done, we have an economic model for the value of the institution. It may bear no resemblance to book value.

This is why fair value is so important.To analyze values, we need to adjust the book value to fair value.

Doesn’t it make sense to obtain management’s view of these adjustments through the use of fair value accounting? Some would object to this, claiming an oversimplification.

What about the value added by excellent customer service? This will flow through in the growth rate of the balance sheet, and an increase in the margin, as customer service contributes (albeit slightly) to these factors.What about innovative programs and solid marketing? The answer is the same; it all comes out in the growth rate and the margin. What about controlling expenses? Again, it all comes out in the present value analysis.

Fair value also promotes good management. Growth is the critical component in valuation, due to the nature of compounding, which should not need to be emphasized to a banker! Yet, many continue to focus on expense control to the exclusion of growth, which has a much lower payoff than growth in the valuation process.

Just look at the valuation of Microsoft for a concrete example. The reason their stock trades at a lower multiple is that their expected growth continues to decline, even as they pile up cash.

There are no real disadvantages to using fair value accounting, other than inconvenience and change. And the advantages are numerous.

Posted in Servicing Valuation | Print | No Comments »

27. August 2007 by Bob Jarmak.

Fair Value Acconting

There are few topics that can cause stronger opinions from finance professionals than fair value (FV). After living with historical cost accounting (HCA) for so many years many see a myriad of issues when confronted with FV. Coupled with the valuation of future income streams, this is a guaranteed controversy.

This is probably due to a number of factors, both real and psychological (behavioral). There is the natural and normal reluctance to change and discard the large historical knowledge base acquired through years of experience. There is also the tendency to view the current method as best, to validate it’s use, much like the purchase of a new good.

But, when we look at the current system of accounting, we see a confusing, troubling mix of HCA and FV, leading to the current problematic environment. On many levels, it is difficult to support the continued reliance on HCA.

What is the purpose of presenting financial information? Who is the intended audience? as we all well know, these questions form part of the problem. The purpose likely depends on the audience. And there are many audiences, or stakeholders in the process. We have government regulators, whose interests may transcend the stated purposes. There may be multiple regulators, each with their own vested interests. And investors, some sophisticated and some not. For mutual institutions we have customers. Then, there is management, who may have vested interests and purposes.

If truly the purpose of publishing information is to permit the valid comparison of organizations, the method should not matter, as long as it is applied consistently. But with a large number of confusing and conflicting regulations, this is not possible.

There is also a movement to principle based accounting, from rule based accounting. Unfortunately this requires the presence of complete confidence in all involved in applying these principles, and based on a long history of observing human behavior, it’s highly unlikely we should believe in the consistent presence of principles where money and personal gain are involved. Is the support for the continuation of HCA rational and reasonable? On most levels, the answer has to be no. In a financial environment where the value of everything changes every day, how can we support the contention that the continued reliance on HCA presents the best view of an organization’s financial status and value?

How can you rationally argue that charging below market rates for loans and /or paying above market rates for deposits has no effect on the value of a financial institution? How can you rationally argue that because it is difficult to establish an exact value for less liquid assets and liabilities, it is better to ignore changes in values than to develop reasonable estimates or ranges of estimates? How can you rationally argue that because current HCA allows or forces certain transactions off the balance sheet, this is the correct way to present the information? How can you support a system that presents financial information in a way that is basically meaningless and useless without substantial further analysis?

While acknowledging there are potential issues with FV, it is difficult to ignore the value of this change.

For many years, market based firms, such as Wall St. dealers have used FV. With the value of much of their balance sheet items readily available through published market quotes, the task appears relatively simple. However, appearances are sometimes deceiving. Even a market as large and public as the MBS market can be “shaded” by dealers. If the market is dominated by the 5 or 6 largest dealers, and their natural position is “long”, it’s quite amazing how prices have a tendency to rise at year end. Similarly, the growth in derivative instruments has been dramatic over the past 15 years, and many of these instruments have no liquid markets. Clearly there is an incentive to establish market prices favoring dealer positions.

Yet, even with numerous hedge fund collapses, FV has remained an accurate way to measure balance sheets of market makers. Compared with these complex balance sheet valuations, community institutions have trivial valuation issues. If it were very difficult to establish FV, sales and mergers would occur at levels with very large price differences. But we know this is not the case. That’s because the normal balance sheet of a community institution is heavily concentrated in loans on the asset side, and deposits on the liability side. Since both are interest rate based, valuations are comparatively simple, with adjustments for credit quality, local market variations, etc. So, if an institution competes in a market where deposit rates are low and loan rates are low, the net will still be accurate. Similarly, competition in a highly competitive market will result in a narrowing of spreads and a resulting decline in value. Perhaps the class with the most significant effect on value will be real estate, as HCA does not take this into effect. However, institutions are very familiar with appraisal techniques to adjust for this value difference.

What’s the single largest advantage of FV? Accountability. It remains simple to move a purchase of securities or a loan acquisition program to “portfolio.” This still generally means the purchase was made or the program was made at higher market levels than the current levels, so the mark-to market is negative. It is often justified to “match” A/L positions. In some cases this is accurate, however, a loss is a loss, whether recognized today, or over time. Bleeding a slow death is just as fatal as a quick one; the only difference is time.

What if a deposit or loan program is priced to attract business? Of course, it may be a great business decision, but this is a marketing expense and should probably be recognized immediately and not amortized over time. Marking to market using FV will properly account for this, where traditional HCA will not. In fact, there are very few examples where HCA presents a fairer, clearer picture than FV. It is very difficult to face the constant truth of FV; it also leads to better long term decisions and more disciplined management.

Posted in Servicing Valuation | Print | No Comments »

27. August 2007 by Bob Jarmak.

ARMS as Investments

CW (conventional wisdom) holds that (residential) Adjustable Rate Mortgages (ARM) are good bank portfolio investments. Yet, numerous analyses show that there is virtually no period in the past 30 years that favors ARM investments over numerous alternatives.

The reason is that there are inherent limitations on the structure and operation of ARMs. Let’s look at a few:

• ARMS are always discounted at inception. The amount varies, but is usually a minimum of the margin over the index;
• There are no prepayment penalties on most conforming ARMS;
• There are rate caps on ARMS. So, initially, the lender earns the index until the first adjustment. The lender could purchase the equivalent UST and earn the same yield, but higher return on capital since the capital requirements are lower for UST’s.

Plus, this ignores the origination and loan servicing setup costs, which are quite high, causing the net lender return to fall.

If rates are flat, the loan will adjust up to the margin, subject to the cap. However, mostborrowers will simply refinance, thus the lender will lose all additional yield, and be forced to reinvest the funds. Or, incur new origination and loan servicing setup costs to acquire a new loan. borrowers a lender wants in the portfolio.

If rates increase, the exact same thing will happen.

If rates fall, the loan may or may not adjust to the then current market. If not, the loan will again refinance.

How can the lender win? It’s virtually impossible. Although there is a level of rates which will provide a slight improvement over the index investment, this occurs very infrequently, and quickly moves away. Why do lenders invest in ARMS? Some probably do make some sense. If there is no initial discount from the index and the property is non-conforming, ARMS are the only realistic alternative to meet customer demand. Plus, a prepayment penalty can reduce prepayment risk. The key is that the initial rate should never be discounted, The main reason many do so is for asset/liability matching. But, there are many better alternatives to ARMS, though these can
be somewhat more complex.

Posted in Investments | Print | No Comments »

27. August 2007 by Bob Jarmak.

MBS (Mortgage Backed Securities) Pricing

The MBS market remains a mystery to many, due to the complexity of the financial instruments. At the risk of “insulting” the knowledgeable readers, it might be a good time to review some of the common questions that arise from apparent pricing anomolies in the MBS markets.

1. How can 2 different coupons trade at the same price? Common sense would seem to tell us that if one security passes through a higher coupon, it should have a higher price. This is obviously true for securities trading at a price around par (100). However, as the price increases, that nasty prepayment factor comes in.

Mortgage market prices are generally driven by:

A. Coupon Rate
B. Spread to comparable US Treasury Instruments
C. Anticipated prepayment rate
D. Supply and demand

So, the first step to determine the indicated price is to run the estimated prepayment numbers to derive a duration, then compare the duration of the MBS to the US Treasury yield curve, adding the market derived spread to finally determine the indicated price.

For example, assume a 9% coupon has a prepayment rate assumption of 1000 PSA, or 60% per year. This might lead to a duration of around 1 year. Then assume the 1 Y UST yields 1.2%, and the indicated spread is 75 BP for this point in the yield curve, this might then lead to a price of 108 or so. Using the same analysis for 8.5% coupons, might show a prepayment assumption of 900 PSA, 54% per year, and a duration of 1.5 Years. If the 1.5 Yr. UST rate is 1.5%, and the required spread is 85 BP, the price might then be the same as the 9% coupon!

Finally, consider supply and demand. Although the market is large and liquid, temporary shortages and surpluses do occur. If a wall street trader has sold a large position of 8.5% securities, and is having difficulty finding them in the market, it will cause the market price to rise well above the indicated price. Before the markets were as large and liquid as they are today, in the 1980’s, this type of shortage would cause prices to rise or fall as much as 1% over/under the indicated value.

More than a few traders lost their jobs because they forgot that the trade was more than an electronic blip on the screen, and they had to deliver real securities at some point!

2. Why do FNMA and FHLMC MBS trade at different prices for the same coupon? This one is pretty simple; there are two main reasons. First, the payment delay and streams are slightly different for the two issuers, which is relatively simple to calculate. Next, there can be a perceived difference in credit risk between the two issuers, such as we saw with FHLMC, causing the credit spread to change.

3. As an investor, should you pay a premium for specified pools vs. TBA’s? While it may appear that specified pools offer greater information and subsequently lower risk, the purchaser pays a higher price. As a veteran of wall street, I can assure you that they have analysed the specified pools, probably a lot better than the normal investor, and have assigned a price well above the fair value, depending on the salesperson to “talk a good sale.” Bottom line, is that realized returns will almost always be higher on TBA’s, since they are priced lower to reflect the risk, even though the street will always deliver the “worst” pools on TBA’s. One of the most effective strategies is simply to roll the TBA’s each month, since the return will undoubtedly exceed taking delivery. (For shorter term time horizons).

Posted in Servicing Valuation | Print | No Comments »