Sunday, February 24, 2008

Recommendations for fixing mortgage securitization

(Tanta at Calculated Risk) I am generally impressed with the quality of Andrew Davidson & Co.'s analysis of mortgage securitization issues. This particular instance, however, leaves me shaking my head. Certainly I give them credit for trying to find constructive suggestions to make, but I don't think we've drilled down far enough into the issues yet.

This does start out right, I think:

The standard for assessing securitization must be that it benefits borrowers and investors. The other participants in securitization should be compensated for adding value for borrowers and investors. If securitization does not primarily benefit borrowers and investors rather than intermediaries and service providers, then it will ultimately fail.
The trouble is that the standard case for how securitization of mortgages benefits borrowers is simply the observation that it makes capital available for lending at reasonable rates of interest. But that's hardly a benefit if it makes too much capital available for lending at too cheap a cost: supply chases down ever more implausible types of borrowers with ever more implausibly "affordable" mortgages, with ever more aggressive solicitation tactics. We're all learning a lot about the costs to all of us of unlimited mortgage money being provided to the financially weakest of us. Therefore securitization's benefits have to be more than just the provision of capital and the lowest possible interest rates; if securitization cannot function to rationalize lending practices by creating "best practice" lending guidelines and loan product structures outside of which its generous capital is not available, then it always has the potential to blow devastating bubbles.

There's a lot in this essay, but for the moment I just want to focus on the analysis and recommendation for dealing with the front-end part of the problem:
At the loan origination stage of the securitization process, there was a continuous lowering of credit standards, misrepresentations, and outright fraud. Too many mortgage loans, which only benefited the loan brokers, were securitized. This flawed origination process was ignored by the security underwriters, regulators, and ultimate investors. . . .

First, originators should be held responsible for the quality of the origination process. Investors in mortgage‐backed securities rely on the originators of loans to create loans that meet underwriting guidelines and are free of fraud. Borrowers rely on originators to provide them with truthful disclosures and fair prices.
Notice how, in the first paragraph, we slip in the first two sentences from "the quality of the origination process" (something quite obviously under the control of the originator) to "underwriting guidelines," which in any securitization practice I know of are either outright stipulated by the issuer in all respects (Ginnie Maes, standard-contract Fannies and Freddies, a lot of private pools) or are at best negotiated between lender and issuer (most private pools, some GSE business). Once the guidelines are either published by the issuer or agreed to in negotiation between issuer and originator, then it is indeed the originator's job to meet them.

But a whole lot of these loans that are failing right now were originated as 100% CLTV stated-income loans, because the guidelines agreed to by the issuer allowed that. I am scratching my head over the logic here: I spent most of the early years of this decade, just as a for instance, blowing my blood pressure to danger levels every time I looked at the underwriting guidelines published by ALS, the correspondent lending division of Lehman. ALS was a leader in the 100% stated income Alt-A junk. And I kept having to look at them because my own Account Executives keep shoving them under my nose and demanding to know how come we can't do that if ALS does it. I'd try something like "because we're not that stupid," and what I'd get is this: "But if ALS can sell those loans, so can we. All we gotta do is rep and warrant that they meet guidelines that Wall Street is dumb enough to publish." Every lender in the boom who sold to the street wrote loans it knew were absurd, but in fact they had been given absurd guidelines to write to. What on earth good did it do to have those originators represent and warrant that they followed underwriting guidelines to the letter, when those guidelines allowed stated income 100% financing on a toxic ARM with a prepayment penalty?
Currently, investor requirements are supported by representations and warranties that provide for the originator to repurchase loans if these requirements are not met. However, when there is a chain of sales from one purchaser to another before a loan ends up in a securitization, investors may find it hard to enforce these obligations. Similarly, borrowers who have been victimized by an originator may have nowhere to go to seek redress if the company that originated their loan goes out of business. Some in Congress are proposing “assignee liability” as a solution to this problem.
Investors don't "find it hard to enforce these obligations" unless they did zero financial due diligence on the last party in the chain. The way it works, your contract is with the last party to own the loan. If you bought loans from Megabank who bought them from Regional Bank who bought them from First Podunk who bought them from Loans R Us who funded the application for a broker, your contract is with Megabank and if the reps are false, Megabank supplies the warranty (the repurchase or indemnification). Megabank can go collect from Regional, who can go collect from Podunk, as far back as it takes to find the original misrep. It's always possible, of course, that the broker made true reps to Loans R Us, who made true reps to Podunk, but it was Podunk who misrepped to Regional (because Podunk bought under a set of guidelines that might have worked with some other investor, but then decided to slip these loans into a deal with Regional, even though Regional published different rules). The assumption that it is in fact always the first originator (the one who closes the loan) who fails to follow guidelines is a huge logical flaw. The contract theory underlying this--that A in contract with B can enforce terms against D who was in contract with C who was in contract with B--startles me as well.

This necessity of "chained pushbacks" certainly does cause grief: eventually, bad loans get back to the originator, but not nearly soon enough. It could take two years for the thing to work through, and delaying negative consequences is never a good thing in terms of keeping incentives aligned properly. But it doesn't hurt investors: they get paid back at par right away from the first party in the chain. In fact, that may explain why, as a rule, they've never particularly cared about the whole problem of "aggregating," or having whole loans work their way from small local originators into the hands of large financial institution counterparties before they are securitized. The investors think, more or less correctly, that their risk is covered because while the loans might have been originated by Loans R Us, net worth $37,000, they were sold to the investor by Megabank, net worth o' billions, and that takes care of the counterparty risk. Which is to say, it puts the counterparty risk on Megabank, who puts in on Regional, etc.

The obvious thing for security issuers to do, if they want direct liability of the loan originator, is to buy the damned loan from the loan originator. Or, maybe:
Assignee liability, however, may create risks for investors and intermediaries that they are unable to assess. As an alternative to assignee liability, an updated form of representations and warranties – an origination certificate – would be a better solution. An origination certificate would be a guaranty or surety bond issued by the originating lender and broker. The certificate would verify that the loan was originated in accordance with law, that the underwriting data was accurate, and that the loan met all required underwriting requirements. This certificate would be backed by a guarantee from the originating firm or other financially responsible company.

The origination certificate would travel with the loan, over the life of the loan. By clearly tying the loan to its originators, the market would gain a better pathway to measure the performance
of originators and a better means of enforcing violations. Borrowers would also have a clear understanding of whom to approach for redress of misrepresentations and fraud.

While risk arising from economic uncertainty can be managed and hedged over the life of the loan, the risks associated with poor underwriting and fraud can only be addressed at the initiation of the loan. Such risks should not be transferred to subsequent investors, but should be borne by those who are responsible for the origination process.
Wall Street security issuers don't want to buy loans directly from any given originator, because that would require them to have loan purchase and sale agreements with a bazillion little counterparties. They like the idea that Megabank has agreements with one hundred counterparties, who each has agreements with one hundred counterparties, etc. The cost of all this managing of relationships and moving loans up into the biggest buckets--the "aggregator" bucket--never goes away, it just isn't carried operationally by Wall Street.

So the idea here is to keep the middle-men and intervening aggregation of whole loans, but to make sure the original party who closed the loan never gets off the hook by having that party issue some kind of surety bond that would be guaranteed by somebody. So Loans R Us, assuming it has enough capital to back a guarantee, issues this certificate stating that it followed ALS underwriting guidelines to the letter. The loan blows up because ALS underwriting guidelines are stupid. Now what?

And of course this is simply meaningless in the context of originators who go out of business. Go ask the monolines how much a credit guarantee is worth if the counterparty doesn't have any money. As it currently is, regulated depository loan originators are required to reserve for contingent liabilities on loan sales: if you have the risk you might have to repurchase a loan, you have to reserve something for that. State-regulated non-public non-depositories may not have such strict requirements, or may not have them enforced. So if you want to assure that originators appear to have the financial strength to make reps and warranties, then you need to look into financial and accounting requirements for originators. Forcing them to come up with a surety bond--putting investors ahead of the originator's other creditors in a potential bankruptcy because said investors don't want counterparty risk--is a bit much.

That's one of those few cases were you can, in fact, pretty much guarantee that credit costs to consumers will rise unnecessarily. If you just want originators to keep skin in the game, there's this old-fashioned way of securitizing loans called "participations" that you might look into. In that case, the investor only buys a fraction of the loan asset (not a fraction of a security, a fraction of a loan), with the originator retaining some percentage interest. That shares the risk between two parties, but also the reward: if you retain a 10% participation interest in a loan you sell, you get 10% of the monthly payment. If you buy 100% of a loan and collect 100% of the payment, and yet you force the seller of the loan to warrant its risk forever, that seller will find some way to be compensated for that--meaning more points and fees to borrowers.

The idea of assignee liability is that you did, in fact, agree to a set of underwriting guidelines when you bought loans or invested in a pool of loans. If you agreed to guidelines which are harmful to borrowers, then this capital you are pouring into the mortgage market is not helping borrowers. The essential confusion here is between failure to follow responsible guidelines and faithful following of irresponsible guidelines. My sad news for the investment community: a whole lot of what you are suffering from is the latter, not the former.

How can anyone possibly require more proof of that? Starting in 2007, investors rapidly pulled out of the 2/28 ARM subprime product. They just announced they wouldn't buy it any longer. And it went away. You do not have a bunch of mortgage brokers still selling 2/28s to borrowers, or correspondent lenders still throwing 2/28s into new securitizations. As you might have noticed, you don't have new securitizations. You always had the power to click your heels together three times and return to the land of just not buying the paper, but I guess you didn't know that until the pink witch showed up.

And you will note that what immediately happened after you all stopped agreeing to those goofball underwriting guidelines was that a bunch of marginal originators immediately went belly-up. That's all the business they could get: writing junk paper for foolish investors. You put them out of business. You should not be sorry about that, except for the part about how you did business with them for so long that now you might have a bunch of worthless contractual warranties. This is called learning by doing. The solution to it is not to go back to buying any old dumb loan that you can get someone to offer a warranty on.

The solution is for investors to refuse to get within 20 feet of a mortgage-backed security that is backed by dumb loans. If you do not know what a dumb loan is, you might want to consider investing in a different kind of instrument. If the guidelines are not dumb, then by all means hold those originators to every last dotted i and crossed t in their contracts, because it is true that the quality of the process is what the originators control. But if you tell them it's OK for them to make loans without seeing docs, without requiring down payments, without worrying about ability to repay, then that is what you get and what we all get.

No comments: