I think I understand the Goldman deal

After listening to the conference call and reading the press release and SEC documents, I think I understand the logic of First Marblehead's deal with Goldman Sachs. Recall that when I bought the company, it needed to securitize the loans it helped originate in order to make money. The banks that originally fund the loans only paid the cost of the loan processing in exchange for agreeing to securitize through First Marblehead. As a result, the business model was heavily backloaded and dependent on investors to buy asset-backed securities.

By the end of 2007, the market for all sorts of asset-backed securities had dried up to the point where First Marblehead was not able to sell the loans it helped originate. As a consequence, the banks originating the loans were able to start charging a penalty to Marblehead (a risk I did not fully understand). This left the company in a bind. It could stop originating loans until the market for them cleared up, losing ground to better capitalized companies. Or it could keep making loans hoping they could be sold before running out of capital. Neither choice was very palatable, though curtailing originations was clearly less risky.

Apparently, GS Capital Partners VI Fund, L.P. and First Marblehead "executed a confidentiality agreement, dated as of July 13, 2007", which means they had been considering some sort of deal for half a year. We can't know what the deal would have looked like at that point, but since the confidentiality agreement was signed by a "global, diversified fund dedicated to making privately negotiated equity investments" (according to its press release), we can be pretty sure it would have been a privately negotiated equity investments similar to the one that was finally agreed upon. Of course the terms would have been more favorable in July than they were in December.

It's possible First Marblehead saw the writing on the wall last summer and was ready to sell an equity share in order to get access to a warehouse facility like it got in December. That would open up options in the event secuitization stalled for a quarter or two. Alternatively, it might have facilitated keeping a portion of loans on the books for one reason or another. But there can be no doubt that what First Marblehead wanted, and what the Goldman Sachs fund could provide, was access to capital. So the plan was to begin shifting from the nearly capital-free business model to one that was more capital intensive.

I think the reason the Goldman deal did not go through during the summer is that it wouldn't have looked good to investors or to clients. Low capital requirements made the company a very attractive investment since cash flows could be diverted to dividends and buybacks. It also created an opportunities for First Marblehead's banking clients, who provided capital to originate the loans. So it would have been difficult to sell this deal in July when business looked pretty good.

By December, when it was clear the business was in trouble, the deal could be seen as a baleout, not a sellout. Whether or not it would have been better to just do the deal over the summer is water under the bridge—the real question is was it worth doing in the winter. And given the choice between stepping out of the student loan business or selling an equity share of the company to keep going, there is a strong case to be made for the later, especially if the plan was in the works already.

To make the case, let's look at First Marblehead's growth history:

Fiscal     2007    2006    2005    2004    2003    2002    2001    2000
Revenues  54.61%  34.83% 109.76% 118.11% 121.42% 511.06%  70.08%  72.38%
Earnings  57.37%  47.78% 112.12% 138.96% 157.58% 487.37%  

Now for the sake of argument, assume that without the deal, First Marblehead will lose money for a quarter or two (their Q2 and Q3), and then make enough in the final quarter to be flat on the year. Q1 revenue, which traditionally accounts for a third of the annual total, was up 24% and earnings were up almost 20% from the previous year, so this scenario seems possible. In 2009, we'll assume growth jumps back to 35%, so that earnings are up a total of 35% over two years. I made this scenario as aggressive as seemed reasonable, since it's the one I'm arguing against.

With the deal, we can assume loan volume growth will be close to the 35% we saw in 2006 because there's no evidence student borrowing is slowing down. Q2 will still be a losing quarter and depending on how the accounting works Q3 might be too. But when the market for ABS opens, First Marblehead will have a lot of inventory saved up. We don't know exactly when securitization might start up again, but let's suppose it takes through the end of 2009 to work through the backlog. Two years of 35% growth works out to about 82% total growth compared to 35% without the deal.

Presumably, the securitizations will be less profitable with the deal, because warehousing the loans racks up interest charges. And of course, you have to factor in the 20% share of the company First Marblehead needed to give up. Obviously, there are risks added as well. But its hard to argue against diluting shares 20% over two years to get 47% better growth over the same period.