Sunday, June 10, 2007

“Uneconomic Transactions,” Or Why I Hate Wall Street

by Tanta on 6/10/2007 09:06:00 AM

So what has given rise to today’s latest installment of WSDS (Wall Street Derangement Syndrome)? In this post of the other day we were attempting to wade through the undistinguished reporting on this current hullabaloo subject of hedge funds accusing investment banks generally and Bear Stearns specifically of “market manipulation.” I observed at the time that, among other things, I couldn’t figure out what sort of mortgage-security-related transactions Bear was alleged to have engaged in: modifications? Repurchases? Purchases of mortgages? Purchase of bonds? Lots of people are focused on what motive Bear might have to do something or other with a subprime RMBS/ABS, the idea being that some action could be taken to prevent write-downs on a security or set of them, such that Bear would avoid making a payout to hedge funds who purchased credit protection from Bear on these “reference” securities. What has been making me crazier than usual is that I can’t get a clear fix on what this “action” is that Bear has been alleged to have taken.

Here’s a new piece of reporting on this, which I have not (yet) found on the web. The source is Dow Jones Newswire, “Artful Hedge Trimming On Wall Street,” by Steven D. Jones (thank you, Brian):

Hedge funds, which insure mortgage-backed securities by purchasing credit default swaps, have complained that some techniques used to restructure loans reduces the value of their swaps. That's because in renegotiating loan terms lenders may pull loans out of the underlying loan pool in a security and reduce the likelihood of default. . . .

But what may appear to be a tempest in a teapot goes to the heart of the swap business. Renegotiating loan terms is one thing, but removing troubled loans from portfolios specifically to avoid paying losses on credit default swaps may erode the value in the swap market that is essential to subprime lending.

"What is objectionable are non-economic transactions in order to avoid making payments to holders of derivative positions," says Harvey Pitt, a former Securities and Exchange Commissioner and advisor to hedge funds in the matter.

When Wall Street firms buy worthless loans out of a portfolio they eliminate the risk of paying off the credit protection the firms sold to hedge funds. Renegotiating loans with homeowners doesn't affect swap investors and benefits lenders and homeowners. But just buying and removing defaulted mortgages from loan pools to avoid paying derivative holders weakens the system, says Pitt.

"My concern is that this type of manipulative conduct will undermine the credibility of the subprime mortgage market," he says.
I confess that there’s something about Harvey Pitt (who last I knew is a former SEC Chairman as well as a former Commisioner, but whatever) making claims about the credibility of the subprime mortgage market that sends my personal irony tolerance well into the danger zone, so let’s not get into consideration of the source here. I first want, however, to highlight the utter absurdity of the claim that, after all the fraud, criminally loose underwriting, and perfectly disgusting capital resource misallocation that has collectively been known as the “subprime mortgage market” in the last few years, it can be asserted with a straight face that there’s any credibility left there to blow.

Ditto goes for the unironic claim that “the swap market . . . is essential to subprime lending.” Well, sure. Flame-retardant suits are essential to surviving the process of driving a motorcycle up a steep ramp at 150 mph and shooting over a series of raging bonfires, but there’s also the idea that you could just not drive a motorcycle up a steep ramp at 150 mph and shoot over a series of raging bonfires and then you wouldn’t really need that fireproof get-up. The “credibility” Pitt is talking about is the “credibility” of stupid subprime lending, a party that’s been and done and over now that the “swap market” has decided it’s no longer a great “risk-free” trade. You want someone to shed tears over that, you’re on the wrong blog.

That nonsense aside, notice once again how difficult it is to really get a handle on what the alleged behavior here is. The implication is that modifying or “restructuring” a troubled loan is OK, but buying a loan out of a pool in order to restructure it is not OK. On the face of it that makes zero sense, and so what is necessary is the further claim that the latter—buying a loan out of a pool in order to modify it—represents a dishonest act by the mortgage servicer because it is a “non-economic” transaction for the servicer. That is, what is assumed is the servicer will lose money doing it that way, and if the servicer is volunteering to lose money, there must be a reason, and the reason must be to avoid paying out on the swaps, which would involve a much bigger loss than the loss on the restructured loans.

Here’s where we need to get all unsloppy with our language and UberNerdly with our concepts, members of the press. I have yet to see anyone in any published source make any reference to an actual shelf or issue in which this practice is alleged to have occurred, so I can’t go to EDGAR, find the deal documents, and tell you exactly what rights and obligations that specific deal gives to a specific servicer in regards to modifications. Until such time as certain folks quit making nonspecific allegations in the press that the rest of us can’t verify because we can’t see what deal you’re talking about—sure, a cynical person might see this as an attempt at market manipulation by the hedgies, designed to drive investors out of those “manipulated” tranches and therefore drive their prices down to increase the swap payout to the hedgies, but of course that would be, like, so totally unfair and mean and unsympathetic besides like totally lacking in actual evidence OMG so I’m not going to do any such thing—we are stuck trying to measure the “credibility” of the manipulation argument by looking at how subprime RMBS/ABS deals usually work. So please take the following with that in mind: there may be deals out there that give a servicer rights or obligations I am not aware of. In what follows I am merely working with what I do know about all the deals I’ve ever seen. I have not seen them all; subprime securitizations have never been a specialty of mine, God help me.

Recall the basic point of securitizing loans: if what you are trying to do is get some risk off your books, you have to get it off your books. This gets you to securitization accounting 101: securitizations can be on-balance sheet debt financings, or off-balance sheet “true sales.” Without wading into the weeds here, the point is that you only get certain liabilities and risks “off your books” if the securitization involves a true sale of the loans to the SPV (the trust entity that actually owns the loans on behalf of the security investors). There are a whole lot of financial transactions, like certain repos, swaps, and collateralized financings, legit and “sham,” that involve party A transferring an asset to party B, but with some understanding that party B can send it right back if it turns out to be “non-economic,” or that party A can demand it back (call it) without market risk. Such transactions are not “true sales”; in the true sale, the seller relinquishes control of the asset, the buyer assumes the risk of the asset, and any return of the asset to the seller would have to be “at market.”

There are, in fact, on-balance sheet mortgage securitizations, but as I believe that is not what we’re talking about in the current situation, I am ignoring those. It would be a lot harder to make claims about “manipulation” if the “reference security” were an on-balance sheet securitization; by definition the sponsor of an on-balance sheet security carries the risk and all “restructurings” would have a clear presumption of being “economic” for the sponsor/servicer.

Theoretically, any “restructuring” of a true off-balance sheet securitization would be economically neutral for the sponsor/servicer. The “economics” now belong to the investors who put capital into the deal. The servicer is paid a servicing fee, and certainly in a badly-written servicing agreement there could be perverse incentives for a servicer to increase its compensation by taking actions that are “non-economic” to the bondholders. We do not particularly seem to be alleging this in present circumstances.

This really is an important point: the sponsor/servicer may not “control” the collateral in an off-balance sheet securitization. The difference between “servicing the loans” and “controlling the collateral” may not be always obvious, but that’s why you hire lawyers to write 80-page pooling and servicing agreements. In all cases, without exception, the servicer must have a fiduciary duty to the trust—which has the fiduciary duty to the bondholders—to service the loans such that gain/loss to the deal is the ultimate legitimate consideration.

The whole hoopla over modifications tends to involve the assumption—fact-based or faith-based—that servicers are modifying loans mostly to benefit borrowers (usually unspoken here is that they do that at regulatory behest; they are not usually assumed to be that altruistic) or to benefit the servicers themselves. They do that either by increasing/maintaining servicing compensation for loans, or by virtue of the fact that the servicer is also the sponsor, the sponsor holds the residual interest in the security (the right to any interest income left over after paying the bondholders, covering losses, and maintaining overcollateralization), and therefore the servicer is identified with the residual, not the funded tranches.

The Fitch report we were looking at the other day, for instance, is in aid of dealing with that latter problem. Fitch declared, basically, that while servicers would clearly be within their rights to modify loans as a loss mitigation practice, those modifications would be taken into account when it came time to “step down” the deal or release the overcollateralization. Short version of what that means: the residual holder would not be able to get its hands on money faster by modifying loans with the sole intent of making the delinquency/default “triggers” look better than they really are.

That might suggest that “the problem”—real or imagined, and I’m not sure how much a lot of this wasn’t imagined—of modifications is on its way to being solved, at least from the perspective of the bondholders. The American Securitization Forum just issued recommendations for standardized deal document language regarding mods that is an attempt to formalize all this on new deals. My own personal view of the whole situation is that there’s nothing earth-shatteringly innovative in the ASF guidelines; it’s not a “new way” of dealing with the ancient problem of modifying loans. It’s a way of sort of dealing with the reality that we have too many mods to do, not that we’re modifying at all.

That is, if the underlying loans weren’t such unspeakable horrors to start with, we wouldn’t be here. But I’m not hearing any market participants—not Fitch, not ASF, not, God help us, a certain former SEC Commissioner—volunteering as how that might be the case. As CR likes to say, “mistakes were made.” The old subprime credit guidelines are “no longer operative.” Let’s all be “constructive” and ignore how we got into this horrible mess in favor of obsessing endlessly about the minor details of how we’re going to try to make lemonade out of it.

OK, so how come this hasn’t, actually, “settled” this modification problem? It’s because Fitch and ASF and those nosy-parker regulators are thinking about the interests of bondholders and sponsor/servicers—with the occasional nod to those borrowers. It appears that we have not yet worried sufficiently about the swappers. That is to say, in all of this we have been considering modifications as “economic” transactions for somebody. They are supposed to be “economic” for the bondholders, to whom the primary, ruling, overarching fiduciary duty here is owed. “Economic” does not mean “at a profit”; the whole great long Fitch document we just looked at is dealing with the problem that “loss mitigation” is about taking the smallest loss possible, not preventing loss entirely and certainly not making a buck. Servicers are on notice that they cannot modify loans unless it is the “best execution” in reference to the loss taken by the deal as a whole, not any class of securities within it, up to and importantly including the residual. If acting in the best interest of the deal is “non-economic” for the servicer, well, bummer. That’s what a fiduciary duty can do for you. Servicing mortgage loans also has risks. This is not a particularly controversial idea to any adult mortgage servicer I know.

What you haven’t seen in the Fitch or the ASF document, however, is any direct discussion of this question of taking a loan out of a security in order to modify it, as opposed to modifying a loan that remains in the security. And that seems to be the big to-do today with the hedgies. Why is that? Well, it is never the point to buy a loan out of a pool in order to do anything with it whatsoever, including but not limited to modifying it. These are not “managed” pools; that’s the whole “off-balance sheet true sale” thingy. The deal documents (at least the REMIC ones) that I am aware of do not give the trust the right to sell loans out of the pool for gain for any reason; the underlying loans are “held to maturity” in the security, not “held for sale.”

All the trust can do is force repurchase of certain loans in a very specific circumstance: if the loan is defective, if it violates the reps and warranties of the loan purchase agreement, or it experiences an EPD and hence invokes the EPD covenant, the trust forces the sponsor to either buy it out of the security at par, or, in most deals, to substitute another, substantially similar, non-defective loan in its place. The latter is an option for the sponsor, and it is quite often not a practical choice for a number of reasons.

It is important to understand the “economics” of this particular issue: the most likely result is that nobody wins. The security gets paid par, which limits its loss, but the security doesn’t make money by returning principal early. That’s supposed to teach you not to buy junky loans from people that will have to be put back; there must be some risk to the security in put-backs or else the crap these investors buy would be even worse than what it is today. (There’s a moral there: put-backs are a “crisis” these days because nobody ever thought they’d have to deal with them; the belief they’d never have to deal with them allowed this ridiculous loosening of guidelines.)

The sponsor buys the loan back at par—it does not owe a penalty premium to the security to “make up the loss”—but it now owns a wretched junk loan that is almost guaranteed to be, if not actually “worthless,” then worth a whole lot less than par. That’s supposed to teach you not to sell junk loans or make misrepresentations when you sell loans.

To sum up, this particular means of “selling” a loan out of a security is “non-economic” to everyone involved, but it is supposed to be non-economic to everyone involved. I repeat: these are not “at market” sales where there would be a gain to the deal. They are “money back warranties,” if you will. If you’ve ever gotten your “money back” on a defective toaster, only to discover that it’s approximately 50% of the price of a new toaster, you understand how this game works.

So I for one do not think we are probably talking about true buybacks here, although anything is possible. There is one other way a loan can be “bought out of” a security, and that is a clause that is typically (but certainly not universally) present in a security servicing agreement that gives the servicer the right, but not the obligation, to buy a seriously delinquent loan (at least 90 days past due, and often at least 90 but no more than 120 days past due) out of the security. Let me say first that we are talking about private-issue non-guaranteed RMBS, not Fannie or Freddie or Ginnie MBS. In those latter, guaranteed securities, either the servicer or the guarantor has the obligation to buy a delinquent loan out of the pool, because in a guaranteed security the bondholders do not take principal losses. For our purposes, we’re talking about securities where bondholders have to take write-downs if they occur. This is why the defaulted-loan removal clause is a right, but not an obligation, of the servicer; if it were an obligation it would be a “guarantee,” and that whole “off-balance sheet” risk-layoff thingy would be a major problem.

You therefore ask: so why the hell would a servicer ever want to do such a stupid thing as buy a seriously delinquent loan out of the pool, when it has the right to leave it there and make the security eat the loss? There are certainly situations in which everyone involved gains by this practice. Imagine that the loan in question is an ARM, the pool is an ARM-only pool, and the only way to fix up the loan and avoid foreclosure is to modify it to a fixed rate. The deal does not want a fixed rate loan in it; that messes up the net yield and cash-flow and may even violate the prospectus (if the prospectus says there are only ARMs in the pool).

So the servicer buys the loan out at par, which means no principal loss to the deal, modifies the loan to a fixed rate, and either holds it to maturity or sells it as a scratch & dent whole loan or resecuritizes it some day in a “reperforming” or “seasoned” deal. Some folks got a bit concerned in the comments the other day about this “resecuritization” of these loans. You really don’t have to worry, usually, about anybody making any pots o’ money off of this; the whole thing is rarely more than a break even, if that much. And the new security is a junk security, not a “new production” security, so it’s not like anyone is fobbing a seasoned modified loan off on someone else as a new warranted loan.

After all that, the best I can speculate is that Bear is being accused of exercising its right, as servicer, to buy seriously delinquent loans out of these pools, in a manner that is alleged to be “non-economic” for Bear as these are loans a rational servicer would leave in the pool for the bondholders' butts to get bitten by. The allegation is that Bear is willing to do this counter-intuitive thing because 1) doing so means there are fewer seriously delinquent loans in the pool and 2) that means that the step-down triggers do not fail which means 3) the subordinate tranches or OC accounts take fewer write-downs or 4) the subs/OC receive stepped-down payments that increase their value or 5) the tranches in question avoid a rating downgrade and 6) all or any of that means that Bear avoids an even more “non-economic” problem involving the credit default swap payouts. The accusation is that Bear loses a little on buying out these yucky loans, as opposed to losing a lot more by settling with those who bet on the yucky performance of the security.

This, friends, is naked “class warfare.” Who wins? The bondholders get principal back on a loan that would undoubtedly have otherwise generated a loss (if not, it would not be “non-economic” for Bear to buy it out). The bondholders also get a “cleaned up” security that is worth a better market price and that steps down faster to increase returns to the holders of the riskiest tranches. The borrowers get a modification that at least in theory helps them hang onto their homes. Possibly “spillover” rushes for the exits are averted, as the value of these securities is stabilized, if not necessarily improved. The hedgies are screwed, though.

I am sorry this is so long and everything, but we have to look at this in detail, because after all of this I for one want to know just exactly where and when all this so-called “insurance” did anything for the subprime mortgage market that could be called helpful, let alone “essential,” to keeping it going. Let us ask: if we didn’t have all this “insurance,” what would we be doing? We’d be modifying those loans right and left in a desperate attempt to staunch the wounds for the bondholders. Funny, that’s what we’re doing with all this “insurance.”

Would there even be that many wounds if a bunch of bright lights on Wall Street had not had this “insurance” available to it? Take the CDS racket out of the equation, and bagholders would be bagholders, they’d know they were bagholders, and the sense God gave an artichoke might have encouraged them to display markedly less enthusiasm for this crap unless those underwriting guidelines were less ridiculous and the due diligence bore some relationship to “due” and “diligent.” As it stands, I don’t see much here except a huge moral hazard, on the one hand, and simple opportunistic punting, on the other. Is Bear “manipulating” a market? In a market this distorted, how, exactly, are you going to define “manipulation”?

This is the point where the Total Pains in the Ass (TPITA, a sophisticated market term) step in and bloviate about how if we didn’t have all those brave hedgies out there speculating with OPM on defaults, the bond investors wouldn’t have touched this stuff with a ten-foot pole, and then the lenders wouldn’t have made the loans because they wouldn’t have been able to hold that risk themselves, and then those borrowers wouldn’t have been given the “opportunity” to become debt-slaves on some overpriced real estate with granite countertops, if they’re lucky, or in line in bankruptcy court or evicted by the sheriff if they’re really lucky. Maybe some of you all aren’t old enough to remember how we used to have to destroy the village in order to save it, but I for one have no nostalgia for those days.

So yes. If, in fact, Bear is buying delinquent loans in order to modify them when it has a legitimate excuse to modify them but leave them in the deal, I would agree that it certainly has the appearance of “manipulation” and Bear has some ‘splainin’ to do. How these hedge funds can have the unmitigated gall to call the kettle black is beyond my talents to ‘splain. There is an old saw about how the child who murdered his parents gets no sympathy for being an orphan. But it isn’t—it shouldn’t be—about “sympathy.” It wasn’t about “sympathy” when we were reading endlessly about poor deluded borrowers or poor deluded lenders or poor deluded bondholders. It’s about how we’re going to return some rationality, transparency, and accountability to the residential mortgage market before every participant plus all the innocent bystanders go down in flames.

That is not something we will accomplish by continuing to think in terms of eliminating risk, or the possibility of “risk-free trades.” I’m the last person to want to eliminate securitization and go back to the bad old days of excessive risk concentrations in depositories. I have no problems with risk being moved to some holder who can withstand it, or risk being shared, dispersed, or “de-linked” such that any given holder of part of it can withstand that part. This Kool Aid about how you can financial-wizard-engineer some magical step in the chain where risk just disappears and everybody is perfectly hedged is insanity.

Moral hazards have to stop. Everyone has to take a chair. And sit on it. If that puts Bear and Paulson both in jail or in damages, fine by me. Just get them out of the driver’s seat of the residential mortgage market and the price of having a roof over your head before we cripple the real economy past the point of redemption. Please stop helping us.