Description

I am recommending the short sale of FMD as a result of an unsustainable business model and aggressive accounting that overstates what I believe to be the true economic value of their earnings stream and book value. I believe that FMD is worth $12-$15 per share and is currently trading at $28 per share.

FMD operates in the nascent private student loan industry which is experiencing enormous growth and increasing competition and FMD has generated a few years of 30%+ ROEs as a result of strong industry headwinds and is currently priced (at 5.80x economic book value) as if these returns are sustainable in the long run. SLM, which has the largest market share in the student lending market and economies of scale in both student lending and servicing maintains an excellent business model that generates an ROE of 30-50%+, is valued at 7x book. I believe that FMD deserves a large discount to SLM.

FMD will probably lose some or all of its 3 top customers which make up approximately 70% of their business and/or face margin compression risk as these banks renegotiate contracts over the next 1-2 years. I believe that FMDs economic book value is about $4.80 and a 3x book multiple is reasonable to aggressive, giving a valuation of $14 - $15 per share to a company that trades for $28 a share. My discounted cash flow analysis which considers competitive pressures and potential losses of customers indicates a range of $12 - $25 per share. If FMDs business model is as untenable as I believe and they lose their larger customers (as I believe they will) and margins deteriorate (which I believe they will), then the stock is worth $12. If FMD maintains their current customers, but margins deteriorate over time to more reasonable levels, the stock is worth $25 per share (and I believe this is the best case scenario and extremely unlikely).

Business Model

FMD receives exorbitant fees in return for helping financial institutions price and facilitate the securitization of private student loans. In 2001, FMD entered into a strategic relationship/transaction with TERI (The Education Resources Institute), the largest guarantor of private student loans, in which FMD acquired TERI’s historical database and loan processing operations. TERI guarantees loans that are securitized by FMD in return for an upfront guarantee fee paid by the borrower and a piece of the residuals of the securitization (economics will be described below). The database acquired has historical data on the FICO scores of borrowers of private student loans as well as the default rates and past performance of these loans. This database is essentially the foundation of FMDs business model and value proposition to customers. FMD uses this database to help financial institutions properly price private student loans upon origination and securitization. However, what FMD actually gets paid for is aggregating private student loans from several different financial institutions into a securitization trust. After facilitating the formation of such a trust, an investment bank actually underwrites the deal, not FMD. This business model is therefore founded on having access to information that some others don’t have, but can create with time. It is also based on having customers that don’t have the scale to pool and securitize loans on their own (which is why the large banks entered into contracts several years ago when the industry was still very small). This model seems to be very untenable as there are no barriers to entry – it is really just a matter of time before margins and returns in this business begin to seriously deteriorate as a result of larger banks performing functions in-house or requiring better economics. In addition to losing customers or facing margin compression from larger banks, FMD will probably experience some direct competition over the next couple of years. There are currently businesses in the industry that are developing a competitive product to FMD and it will take time for these businesses to develop, but they will very likely increase competition and compress excess returns.

70% of FMDs revenues came from three large financial institutions (Bank One/JPM, Bank of America and Charter One) in 2004. I believe that all of these customers will bring this function in house or take some economics away from FMD as it is realizing returns on its equity base of over 30% after adjusting book value and earnings to reflect true economic values. It does not make sense for these banks to be passing on such huge value to FMD for providing a service that banks (especially the three listed above) have the scale to profitably operate in house. These same banks should also be able to create appropriate databases to help them price student loans properly as they are in the business of pricing credit risk (whether it is auto loans, consumer loans, mortgage loans, etc.) I think a good benchmark is the mid-90’s when the credit card industry was growing and the larger financial institutions saw the returns experienced by the mono-lines. It was only a matter of time until the financial institutions developed their own systems and pricing databases in order to be able to compete in this space.

As students of competitive behavior may argue, although FMD does not have any true barriers to entry, it does have somewhat of a first mover advantage in the sense that it will take these bank customers several years to develop their own internal databases. And as a result, FMD will be able to create significant value in the mean time potentially making current stock prices look more reasonable. After making adjustments to book value, FMD is trading at 5.80x book and as indicated above, is currently recognizing 30%+ returns on that book. If this were to continue, one could argue that at 5.80x book this is not a good company to short. As I will illustrate below, I believe that it is 3-4x more profitable for these banks to keep these loans on their balance sheets as they develop their own databases and should be inclined to do so until the databases are fully developed. This alone would destroy FMDs current and expected earnings power. Both Sallie Mae and Student Loan Corp. perform these functions in house and have developed their own databases. People from both companies agreed that it should not be difficult for these large banks to create their own databases and perform FMD functions in house. In addition, if you ask SLM why they don’t get into this business, they will flat out tell you that they have no interest in competing with FMD in this business because they don’t think that it is a sustainable business. SLM is a tough competitor and a category killer that has its own database and would surely enter this business for 30%+ returns on equity if they felt that the business was sustainable. Student Loan Corp would probably enter this business as well if they felt that the business model would continue to provide these returns. Even if several of FMDs bank customers don’t take the function in house, they have serious leverage when renegotiating their contracts and should negotiate down the returns that FMD has been realizing to this point.

After speaking with people in the ABS group of the investment banks that actually underwrite these loans for FMD, you will hear that FMDs value proposition is that they are an aggregator of student loans and are very valuable to smaller financial institutions that don’t have the scale to securitize on their own. For someone like JP Morgan, it may have made sense several years ago when they entered into the contracts since the market was small and they did not have the necessary scale. FMD lacks any real franchise value since they do not originate loans on their own and simply serve as a facilitator. I would liken their job more to the underwriters at the investment banks in the ABS group who receive under 100bps in fees (vs. the 15% or 1,500bps FMD receives) as a direct result of a competitive marketplace. I believe that over time, FMDs offering will become more of a commodity and returns will adjust accordingly. As I mentioned earlier, competition will increasingly appear over the next few years and this should strain margins.


There are three types of loans FMD securitizes from three different channels. Approximately 99% of FMDs loans that it securitizes come from both the school channel and the direct to consumer channel. The school channel is one in which lenders are able to market their loan products via the financial aid office of the school. In this channel, there is intense competition amongst the lenders and schools also have a say in the pricing all of which is reflected in lower loan yields. FMDs school channel must be distinguished from competitors as they cater increasingly to proprietary schools/1-2 year schools in which borrowers are less likely to gain a significant increase in their earnings power. In the direct to consumer channel, lenders are free to operate without any oversight from the schools and due to the fact that it is a relatively nascent market, yields are unsustainably high. If you look at FMDs latest securitization prospectus, you will see that 80% of loans come from the DTC channel and 30% have FICO scores below 700. And as investors who cover this space know, default rates increase drastically as FICO scores go below 700. Many competitors are getting into this market because yields are sufficiently high to compensate for the risk taken. However, I would assume that this opportunity for excess returns will be removed by new entrants. This will cause additional strains on the margins (specifically hurting the residuals which are booked as a result of excess spreads) that FMD has been able to maintain up to this point.

In addition, more and more companies are beginning to offer private loan consolidation options. Collegiate Funding (CFSI) which makes up most of the business that FMD records as Charter One recently renewed their contract with FMD until 2007, but also announced that they would pursue a similar product offering that they would keep on balance sheet. CFSI is also offering private loan consolidation options to borrowers who may be leaving school and beginning to pay off their loans. These same borrowers may have been a huge credit risk previously, but after exiting school and indicating that they will be able to make their payments, CFSI will go to them and offer them lower variable rates to consolidate their loans. This process will essentially cause FMD to experience much higher prepayment speeds than expected (which could destroy their residuals) or it will cause them to reduce their yields, eating into excess spreads. These consolidation options have not yet affected the industry since the market is still young and it takes a few years for borrowers to leave school and exhibit their ability to make payments.

Economics

Let’s walk through a typical securitization that is executed through First Marblehead. If you look at the prospectus for the most recent securitization done through National Collegiate Funding LLC (FMDs securitization trust) dated May 20, 2005, you will see the following:

Proceeds from ABS Issuance ($M)
Initial Trust Student Loans $391
Gain on Sale for Bank Customer 18
Upfront Fee to FMD 33
Reserve Account 116
Pre-Funding Account 84
Deposit to TERI Pledge 2
Cost of Issuance 2
Underwriting Fee 2
Total Proceeds from ABS Issuance $649

Gain on Sale to Bank 4.67%
Upfront Fee to FMD 8.46%

Collateral
Initial Trust Student Loans $391
Reserve Account 116
Pre-Funding Account 84
TERI Pledge Fund 23
Total Collateral $614

Parity Ratio 94.7%

As you can see above, in a typical securitization, the bank recognizes a gain on sale of
4-5%, FMD recognizes an upfront fee that it takes in cash of 6-8% (although in the above securitization, this amount is higher) and the parity ratio is about 95%. In addition to FMDs upfront fee that it receives for structuring/advising on the securitization, it also recognizes an additional structuring/advising fee of about 1% that is paid over the life of the loan (but recognized upfront) and it receives a 75% share of the residual interest in the securitization which is also capitalized and recognized upfront (and TERI receives the other 25%). After adding up all the fees, FMD is recognizing 15% in gains upfront. And this is on top of the 4-5% that the bank customer recognizes upon securitization which gets you closer to 20% in upfront gains (and this still excludes 25% of the residuals which is booked by TERI). Keep in mind that Sallie Mae recognized upfront gains of 9-10% in 2004 on their securitizations and 15% in the first 6 months of 2005. Part of this gap can be explained by the fact that FMD operates more in the DTC channel which experiences higher yields leading to increased excess spread (and increased credit risk, although it seems like risk adjusted returns are still high and have not been competed down yet) and greater weighted average loan length, but according to my math this does not explain the entire gap. The residual interest is a non-cash gain and FMD does not receive cash flows for about 5-6 years or until the excess spread allows the collateral in the trust to build up to the point where there is approximately 103% parity. I believe that FMD is capitalizing these residual interests in the securitizations very aggressively. My calculations indicate that the residual values recorded on FMDs balance sheet at $247 million are actually worth $136 wiping out $111 million in equity or 27% of the equity base. This reduces book value per share to approximately $4.80 from $6.50 and implies that FMD is trading at a price / book multiple of 5.80x. In order to calculate what I believed to be the real value of these residuals based on the cash flows I thought FMD could reasonably receive, I looked at their assumptions in their 10K, securitization trusts and spoke with the CFO. Very simply, they receive 75% of the 245bps of excess spread beginning in year 5-6 after the 103% parity has been reached and they receive this spread until the end of the life of the loan which is on average 10 years. Assuming that defaults are below the 6% collateral that TERI pledges (which given credit quality of the DTC loans is probably too low) and assuming a 12% discount rate (which they use, but given the nature of these residuals one could argue is too low) I calculate the present value of these flows to be about 3.0%. The majority of FMDs securitizations have been executed over the past 3 years so it is easy to go back and adjust the rates at which the residuals are capitalized to get an economic value for the equity in the business.

FMDs high book multiple can only be supported by significant growth moving forward. I believe that FMD should not and will not experience significant growth in book value as a result of a couple of potential alternatives that could materialize. These alternatives include bank customers retaining loans on their balance sheets, bank customers continuing to securitize, but only if FMD receives less appealing economics per transaction, excess spread being bid down in the DTC channel due to increased competition or excess spread being maintained but only as a result of taking on increased credit risk. In order to better understand why I believe this is inevitable, I will provide the economics to banks when they retain loans on their balance sheet and when they securitize the loans.

The way I analyzed this trade-off for banks is to assume that since these private loans are considered 100% risk weighted assets, banks would need to keep capital of approximately 8% against those assets until the loan was repaid in full. In addition, the excess spread on these loans would be greater because banks would keep the credit risk for these loans and the fee for taking on this risk would be capitalized in the principal amount of the loan, effectively increasing the yield the bank receives on the amount it gave to the borrower (you could also assume that banks don’t take on the credit risk and thus pay a guarantor, thus reducing the excess spread and also reducing the credit risk). I assume that default rates are about 5% which is in line with what several lenders in this space have expressed and I also assume recoveries of about 60%. Based on these assumptions, you can see below that value created is much greater for the bank when the loan is retained.

Initial Capital Infusion 8.00%
Excess Spread (1) 3.50%
Guarantee Fee 7.50%
Default Rate 5.00%
Recovery Rate 60.00%
Economic Value Created when Held on BS 14.78%
Economic Value Created when Securitized 4.50%

(1) Underestimates excess spread because funding costs should be lower than ABS investors require.

As a result of the above calculations, it would seem to me that banks will begin to keep loans on balance sheet or if they do want to securitize, they will not settle for the current gain that they are receiving and will thus negotiate to take some value from FMD.


Customers

JPMorgan Chase – Jamie Dimon has a large say in the decision regarding whether JPM will bring the services FMD performs in-house. The company is not willing to tell you much, except that they are focused on maximizing risk-adjusted returns and will keep loans on balance sheet if that is more profitable than selling them. They recently terminated their relationship with SLM in which they sold all of their FFELP originated loans to SLM. They are now originating a percentage of these federally funded loans on their own and keeping them on balance sheet since the economics are better than selling them for a gain on sale to SLM. I think that the decision to do this will be very similar to the decision they will be faced with when the contract with FMD comes up for renewal. In the very least, I think they will keep more loans on balance sheet due to superior economics (decreasing the volume that goes through FMD) and there is a good chance that they will not renew at all and perform all the pricing/securitization functions in house as they do for all their other loans (auto, consumer, mortgage, etc.).

Bank of America – I spoke with them and after a few questions, I found myself being yelled at by the IR guy. He yelled that it did not matter what the economics looked like on balance sheet as opposed to off-balance sheet. Regardless of economics, he continued, BAC would maintain its relationship with FMD. He either has no understanding of this decision making process, understands the decision and was just annoyed because of a plethora of calls or BAC does not care about giving economics away to its suppliers/outsourcers. Basically, there are many different ways to interpret the call.

CFSI – Already bringing functions in house. They recognize that it is significantly more profitable to hold loans on balance sheet and will begin to do so. They will maintain their relationship with FMD and securitize loans from certain channels (mostly DTC) through FMD until 2007 while competing with FMD in other channels. Given that they also plan to securitize private loans on their own (in addition to maintaining a percentage of their originations on balance sheet) I believe that when the FMD contract comes up for renewal in 2007, CFSI will not renew or renew with significantly less volume going through the FMD channel. CFSI believes that they can develop their own pricing databases over time and in the mean time they feel that they will be able to securitize without FMDs databases.

Catalyst

Catalysts

-One or more of the larger bank customers bringing function in house
-Larger customers increasingly keeping a greater percentage of their private loan originations on balance sheet
-Other competitors entering the space and competing away the excessive margins
-Investors beginning to properly discount the potential threats to FMDs business model even though many of the larger contracts are not up for renewal until 2007