Peeling Back the Onion

A Look at the Inside of a Subprime Security:

WaMu Asset-Backed Certificate WaMu
Series 2007-HE2 Trust Issuing Entity

I have a deal for you. You can buy a WaMu AAA tranche subprime security today for about 68¢ on the dollar, a 32% discount on securities that were priced at 100¢ on the dollar when issued in April, 2007. The BBB tranche is selling for about 19¢ on the dollar. That’s a huge hit in a very short period of time.

What happened? The deal was structured so that they were oversecured, insured, preferred, and highly rated. Yet in a few months they were worth much less.

It is widely believed that most investors never fully understood how they work (I don’t really believe that) so here’s a chance to find out what you don’t know. I downloaded from EDGAR the prospectus for WaMu Asset-Backed Certificate WaMu Series 2007-HE2 Trust Issuing Entity and analyzed it.

The concept isn’t difficult: take 6,723 mortgages written by Washington Mutual, wrap them up into a security called a trust, slice up the trust into about a dozen certificates (“tranches”) of risk, and sell them to the public. How hard is that? It was issued in April, 2007 and contained about $1,600,000,000 in subprime mortgages generated by Washington Mutual.

I’m going to describe how it was structured as simply as possible. I’ve probably glossed over points that the issuers may think are significant. So be it.

The Mortgages

First, WaMu sold 6,723 mortgages to a trust managed, in part, by WaMu. These are all subprime mortgages that it had originated all over the country. They seem to be fairly geographically distributed.

WaMu’s underwriting standards allowed loan-to-value ratios (LTVs) of up to 100% of the appraised value of the mortgaged property. The borrowers didn’t have to prove income or have good credit. Income was whatever the borrower put down on the loan application. That’s why they are subprime.

WaMu had a complex borrower credit rating system which considered credit scores (generally, at least 500), the LTV, whether or not the borrower had any defaults, and debt service-to-income ratios (55%). The better the borrower’s rating, and the lack of secondary financing, the higher they would allow the LTV. The subprime loans were rated in quality by WaMu from A to C.

About 75% of the mortgages had adjustable interest rates (ARMs) and the balance was fixed rate mortgages.

The Certificates

Second, they carved up the Trust into 19 tranches represented by investment certificates. The investors could buy one of 11 A-rated certificates and choose among 3 B-rated certificates. The remaining 5 certificates relate to the credit enhancement features of the Trust. Each tranche had its own mortgages. The mortgages were serviced by WaMu and there was a pooling agreement and trust agreement controlling the operation of the Trust. These provisions and the operations were well thought out and fair to the investors, based on the investment terms.

If the investors had difficulty in understanding the deal it must have been in the allocation of risk. There is a complex system for allocating risk away from the senior tranches and putting it on the junior, subordinate tranches.

About 84% of the mortgages were allocated to the A tranches. The balance went to the B tranches. Which means that specific mortgages were held by the A tranches and specific mortgages were held by the B tranches. The returns of the investors were from the mortgages in which they invested.

The returns were based on a premium over LIBOR. Let’s just say that the higher rated A tranches got lower returns than the lower rated B tranches because of the relative risk of the B to A tranches.

If there were losses in the returns for the A tranches, the income from the B tranches and the other credit enhancement devices would kick in to fund the A tranches.

Credit Enhancement

The interesting part of the Trust is the structure of credit enhancement and the order of investment priorities.

First, all the B tranche certificates and their underlying assets are subordinated to the returns of the A tranche investors. The lower B tranche certificates were subordinated to the higher B certificates.

Second, there are credit enhancement features such as “overcollateralization”, a mortgage insurance policy, and a swap agreement.

When WaMu put this deal together, they had each tranche rated by Moody’s, S&P, and Fitch. The A tranche certificates were all rated AAA. The B tranche certificates were rated AA+ to A- for the first 6 tranches. The last 3 were rated BBB+, BBB, and BBB-. So 11 of the 14 certificates sold to the public were rated in the A’s.

Recall that these are risky loans, unsupported by tax returns or W-2s, have high LTVs, have adjustable interest rates (about 75% of the mortgages), low loan coverage, and borrowers with modest credit scores. Many of the loans were generated by independent mortgage loan brokers.

Because of the structure of the deal, the rating agencies converted risky loans into AAA rated securities. Wait a minute; that doesn’t make sense. But . . . if housing prices continued to rise, who would be hurt?

By the way, most of these securities have been written down by the rating agencies. Each tranche has been written down at least one grade.

Then there was mortgage insurance. About 14% of the A tranche mortgages were insured and about 8% of the B tranche mortgages had insurance. So, the issuers couldn’t conceive that it might be necessary to insure more than that. Hey, we’ll have price appreciation. Don’t worry.

Then there was a swap agreement. The swap was set at a LIBOR-based rate which was about 4.76%. So, if the payments on the mortgages were less than the LIBOR-based rate, the swap counterparty (ABN AMRO) had to make the payment. Well, what if there are large amounts of defaults? The swap agreement says that if the securities are downgraded, the swap agreement could be terminated.

The securities were “overcollateralized.” About $60,000,000 of mortgages were set aside to cover losses (the C tranche). Not a bad concept. But that’s only about 3.75% of the total mortgages.

All in all, these weren’t badly designed structures. Aside from the Midas part of turning junk into gilt-edge, they were reasonably, lawyerly, MBA-ishly, well thought out and structured to prevent fraud, theft, embezzlement, and the usual safeguards that give the appearance of solid, Big Firm securities.

But they did turn junk into gold. And that was a fatal flaw in these deals when the dam broke and housing prices collapsed. Were they really aware of that risk?

Investment Risk Factors

If you’re a lawyer, the risk portion of the prospectus is one of the areas where the wolves can get you. You’ve got to mention all the risks reasonably expected to arise from an investment in the certificates. Let’s say you forgot one risk and the investment craters. You can safely assume the investors will sue. They’ll sue the lawyers and the issuers, and anyone else standing in the vicinity. Obvious stuff.

Look at the Risks section of the WaMu prospectus.

The first risk they point out is that subprime mortgages have higher delinquencies than other mortgages. Okay.

The second risk points out that the housing market is getting bad. The title of this risk is: “Recent Developments in the Residential Mortgage Market May Adversely Affect the Return on Your Certificate.”

It says, in part:

Recently the residential mortgage market in the United States has experienced a variety of difficulties and changed economic conditions that may adversely affect the yield on your certificates. Delinquencies and losses with respect to residential mortgage loans generally have increased in recent months, and may continue to increase, particularly in the subprime sector. In addition, in recent months, housing prices in many states have declined or stopped appreciating, after extended periods of significant appreciation, a continued decline or an extended flattening or those values may result in additional increases in delinquencies and losses on residential mortgage loans . . .

Etcetera, etcetera, and you could lose all your investment as a result. Slightly understated?

This gets back to my original dismay over the timing of this deal. This was sold in April, 2007! New home sales were collapsing, inventory was rising, house appreciation flattened or was declining, and default rates, especially with subprime mortgages had been increasing since about August, 2005 and took a big jump in June, 2006.

Why, therefore, would anyone buy into one of these deals? The risks were obvious and were printed on the first page of the prospectus’s risks section. But no one saw the bust coming.

An interesting paper on the subprime crisis by Yuliya Demyanyk and Otto Van Hemert dated December 10, 2007 and published on the St. Louis Fed’s web site (“Understanding the Subprime Mortgage Crisis”) says that issuers were aware of the increased risks in the subprime lending market, and yet the spread between the subprime rates and the prime rate continued to narrow (one would expect subprime mortgage rates to be substantially higher in light of the perceived greater risk). The risk wasn’t masked by housing price appreciation because at this point prices had flattened.

I find the implications disturbing: a) risk was not adequately priced into the subprime rate, yet b) they knew it, c) they kept selling these securities, and d) investors were still gobbling these securities up. I understand the amount of money seeking returns at the time may have been suppressing the perceived risk in terms of spread. But what were they all thinking?

Demyanyk has published additional studies of lending cycles and says this cycle is no different than the others. I think that says a lot about human nature. When “everyone” is doing it, the sheep get in line. They ignore the warnings and the fees flow. When the cycle busts, they are as surprised as anyone that it happened to them.

I would suggest that the reasons lie in greed and, well, greed. The investment bankers who invented these securities were pretty slick. They came up with the term “subprime.” When you think about the word “subprime” it’s actually a misleading term. These aren’t just slightly less than “prime” mortgages as the name would suggest. These are risky mortgages. In fact they should have been names “junk” mortgages. If we had a “Junk Loan Crisis” perhaps it would be clearer to the public. If investors were buying junk mortgages maybe they would have asked more questions. But the clever investment banks spun the “junk” part into “subprime.” Does subprime sound that risky?

By spinning junk into gold everyone made a lot of money. The fees from this WaMu deal were in the millions if you consider the underwriting fee, service fees, legal fees, banking fees, rating fees, and the like. I don’t begrudge anyone a fee.

What is wrong is that their assumptions about the nature of this investment were wrong. No one had the balls to suggest that when the merry-go-round of housing appreciation stopped, as all cycles like this eventually do, that these securities could be worth less, a lot less. Where were the rating agencies? Where were the silk stocking investment bankers when investors needed protection? Where were the hedge fund advisors? Too busy collecting fees?