When the Federal Deposit Insurance Corp.
(FDIC) closed on its recent RMBS, the headline called it a "pilot securitization of performing single-family mortgages from 16 failed banks."
Translation: expect more to come.
It makes sense to the FDIC to do this, of course, as spreads on the deal indicate a cheap way to finance what is becoming a more and more expensive book to balance. It's a good deal for the investors and a great deal for taxpayers as it stops the need for more funding from Congress.
“The FDIC uses several strategies to sell assets from failed banks," the accompanying press release said. "Securitization is one of the ways in which the FDIC intends to maximize the value of these assets for the benefit of creditors of the failed banks."
And there is certainly a future in the FDIC getting into securitization more and more. Consider loans in non-recourse states. The FDIC may be sitting on these assets for a very long time, especially if the borrower makes it through HAMP trials intact. Perhaps if this pilot model performs for the FDIC as expected, I wouldn't be surprised if we start to see more esoteric deals, such as a HAMP-backed RMBS.
But is that really the motivation behind this deal? Perhaps there is something larger here, such as a motivation to get the secondary markets moving again. In that mode, is the FDIC setting a precedent on what it thinks good securitizations look like? If so, then the FDIC has the market pegged wrong.
The deal is overcollateralized to 15%
which means that if it were rated based on that criteria alone it would almost reach triple-A status, if expected losses are modeled at 3% at the point of credit enhancement.
But the FDIC didn't bother with credit ratings. Didn't have to. The deal bears the full faith of the United States government.
[Requests to the FDIC to discuss FDIC 2010-R1 were unanswered.]
Structurally, the biggest issue is with the re-underwriting of the deal. When Redwood Trust came to market
, its due diligence provider claimed that 100% of the loans needed to be rechecked.
In the case of the FDIC, the original FICO scores and income levels at point of mortgage origination were used – the entire process was pretty much performed electronically. AVMs were the most common form of appraisals, though 10% of the properties were randomly selected for BPOs.
In one pool of loans the average LTV is 111%, in another its 70%. An advisor to the deal told me that it felt more as if the FDIC was purchasing and selling loan pools instead of putting together a securitization.
Here is where the market faces a challenge: could the FDIC structure a RMBS that the industry can follow? Has it been successful without the faith of the US government and instead using the strict Redwood
approach to due diligence? It's not likely considering the borrower is under no obligation to cooperate. Some may even frown not just on giving out updated personal data, but to the concept of MBS itself, so big the name has become. This is especially an acute risk if a failed bank holds the mortgage.
But while it doesn't work for the market, it certainly works for the FDIC.
Jacob Gaffney is the Editor of HousingWire
Write to him