Tuesday, March 13, 2007

Tanta Makes the Argument for Conformism and Other News of the Weird

by Calculated Risk on 3/13/2007 03:20:00 PM

CR note for new readers: Tanta posts here occasionally. She is a former bank officer and mortgage lending specialist who is currently on extended medical leave.
Jillysi asked a good question about this odd term “prime conforming” in the comments to an earlier thread, and Dirk van Dijk followed up with a good short explanation. I, who have never been able to leave a good short explanation alone, feel the need to complicate things slightly. Trust me; I have a purpose other than just making things difficult.

In the old days--a few years ago--you would find mortgage people using the term "conforming" in its fullest sense: a loan that conforms in all respects to the requirements of Fannie and Freddie. That includes the maximum loan amount—currently $417,000—but also all the other credit, underwriting, documentation, loan term, etc. requirements of the GSEs. A "nonconforming loan" was one that did not meet GSE requirements in at least one characteristic. It only takes one "fail" out of a long list of requirements to make the loan ineligible for purchase by the GSEs, but of course a loan can fail in more than one way. So the term “nonconforming” in this robust sense does not necessarily indicate the precise degree of deviation, just that there is at least some deviation.

Oldtimers also used the term "conforming balance" to refer to loans that are at or under the maximum loan amount. (This was a useful formulation because the actual conforming loan limit tended to change every year. Wise writers of credit policy did not put the actual current dollar limit in the guidelines or contracts in a hundred different places.) The term for loans over the maximum loan amount is "nonconforming balance" or, more usually, "jumbo." A jumbo loan could have exactly the same credit, product, and documentation characteristics of a conforming balance loan, but be simply too big to go to the GSEs. Or, of course, a jumbo loan could depart from GSE standards in multiple ways. There was a time, certainly, when nearly everyone’s jumbo guidelines were stricter—sometimes much, much stricter—than the GSE’s for things like LTV, FICO, DTI, reserves, and so on. The unsurprising idea here was that if you were lending that much more money, you needed to be that much more cautious. But then a certain housing price bubble came along, and those jumbo loan balances—especially in the conforming to $650,000 range or thereabouts—started to seem more “normal” in certain parts of the country. It therefore began to seem more “normal” that such loans should be offered on the same terms—or even more relaxed terms—than the GSEs use. A very dangerous idea was thus born.

As you can see, once these terms escaped their fairly precise use within the industry, and became more everyday terms in the financial press, they got a little less specific. Dirk is right that in most places the term "conforming" now generally refers specifically to loan balance, not to the long, long list of other requirements a loan must meet to be eligible for purchase by the GSEs. That's actually why you get this odd phrase "prime conforming." (To someone like me, that's an oddly redundant formulation.) "Prime" itself as a credit designation in the mortgage business was traditionally defined as the level of creditworthiness required for loans sold to the GSEs in their standard MBS programs. What happened is that 1) the GSEs started buying some subprime for their portfolios, so you couldn't always assume GSE = prime without carefully specifying that you meant the standard MBS programs, and 2) a huge market opened up for prime-credit loans that don't meet the GSE requirements in some other ways, now generally lumped into the big category called "Alt-A."

The usual story on "Alt-A" by the mortgage industry is that it is prime credit—that's the "A" part of the name—but it is based on some "alternative" way of structuring or qualifying or documenting a loan, the alternative being in reference to GSE standards. The problem this has and continues to create is that there becomes a question of how far off the reservation of GSE standards you can wander before the "Alt" part cancels out the "A" part. This is a question at the loan level, and also at the pool/MBS level.

Probably the big innovation that the GSEs had on the market for mortgage-backed securities, back in the day, was their process of selecting homogenous and uniform loans for their security pools. If you can create a set of rules—underwriting guidelines, product features, appraisal rules, all the way down to very detailed requirements for legal documents, property and title insurance, etc.—and make sure all loans fall within those rules, you can build a security with a fairly high degree of confidence in the assumptions about delinquency, default, prepayment, enforceability of legal rights in the event of foreclosure, and so on. This allowed the agencies to create reliable calculations for the guarantee fee they would need to charge the lenders in order to cover the credit risks of the pools; it also made it possible for various kinds of institutional and depository investors to buy those securities in a market that could establish apples-to-apples prices for them. The regulators of depository institutions had already reviewed the GSE guidelines, as it were, and given their general approval to such standards; the depository didn’t have to wade through a 400-page prospectus for each deal. GSE loans became the vanilla ice cream of the mortgage world.

Of course no one ever claimed that only GSE-style loans were worth making. First, of course, the GSEs by charter were limited to moderately-sized loans, and so there was always a segment of the market—the jumbo market—that was simply outside the GSE sandbox for charter reasons. Also, there have always been all sorts of near misses, one-offs, unique deals, and so on, that are certainly loans worth making, but that just can’t go to the GSEs; banks generally held these in their investment portfolios. Again, although it’s possible that such loans are of weaker credit quality than the GSE standards, it’s also possible that they’re more or less equivalent, if you go to all the trouble of wading through all the variables and risk layers and coming up with a total. The idea of GSE MBS was simply that neither investors nor the GSE’s buy side staff would have to do any “wading,” as an efficiency issue as much as anything else, and that prices for those securities could be set with a lot less fuss. There is of course a huge value to efficiency, if you’re trying to run a trillion-dollar securities market, but it was never intended to be the last word in credit quality.

So the big problem with “Alt-A,” from the beginning to right at the moment, was and is that you have to wade into the weeds. A lot. Much of the cries of outraged innocence we’re hearing lately from some investors stems from the fact that they didn’t wade very deeply into the loans underlying the securities they bought. There are lots of reasons for this; to my mind, the biggest one is that a bunch of people seemed to think that one could invest in Alt-A pools at the same operational cost as GSE MBS. But, of course, that’s missing the whole point. In any case, investors became more and more reliant on the rating agencies, whose job was largely to build models that would take these risk-layered loans and end up slapping some overall credit designation on the pool or the security or a tranche thereof. Many folks, right about now, are taking the rating agencies to the woodshed over some of the assumptions in those models and the rating agencies’ apparent inability to react (downgrade) fast enough when new data on performance catches up with some of these loan products underlying these pools. I can’t argue with the critics of the rating agencies in the specifics. On the other hand, I find myself chuckling a little—in a cynical way—over the underlying problem here. For the rating agencies to become fully effective modelers and predictors of credit risk for Alt-A is, at some level, for them to reinvent a certain wheel that the GSEs invented a generation or more ago on the “conforming” side. In other words, we’re expecting the rating agencies to homogenize the heterogeneous so that bond buyers can trade RMBS without doing all that wading and paying all those analysts and complicating all those VaR models. We want mocha java praline mango on a sugar cone at vanilla levels of efficiency.

Theoretically, that can be done—one can build guidelines and other requirements that standardize the Alt-A world fairly reliably, or at least break it down further into somewhat more homogeneous sub-classes—but I confess to being surprised that anyone thinks the rating agencies are the parties to do it. The GSEs were not simply standardizing other people’s loans: they were the buyer, the one on the hook for the credit guarantees, the party of the big risk kahuna. (So were/are the mortgage insurers, which is why you see a whole lot of overlap between MI and GSE credit guidelines.) I’m not sure what other people expect by asking fee-based rating agencies who are not owners of the risk to get in there and standardize the Alt-A market, but what I expect is not very much. My own view is that being on the hook in the event of default tends to concentrate the mind in ways that making fees off of the volume of securities issued does not. Call me cynical if you want, but I think I’m just making a pretty obvious point about how risk works.

So why, you ask, are not the real risk-owners—the buy-side Wall Street firms and big originators who invest in or securitize Alt-A—making it an efficient and standardized market? Well, because it’s a competitive market. Those who like to complain about the GSEs being “monopolies” like to imply that the more competition in the mortgage marketplace, the better things will be. I have my doubts about some of that for lots of reasons, but I will note that the other side of “monopoly” is homogeneity, efficiency, and transparency to the investor (at least as far as the underlying loans are concerned; I’m not talking about GSE accounting practices here. In fact I have no objection to FASB holding a “monopoly” on accounting standards, to make a related point.) Once “competing” products or loans or security issues enter the marketplace, by definition the things can’t just be more of the same. In a marketplace unprepared to do the analysis, due diligence, and complex price-modeling on these “competing products,” it became, after the dust settled, a fairly predictable question of the highest apparent yield getting the most bids. Anyway, one reliable sign that there’s a bubble going on is not so much the repeated statement that “it’s different this time,” but the repeated habit of using “the same” analytical or due diligence or pricing tools on what is clearly not “the same” precisely because it was marketed to you in the first place as “not the same old same old.” I cannot help offering a general “duh” here to certain institutional investors who have just discovered something that was hidden in plain sight.

This is rather a long way of saying that there are reasons—big, important reasons—why this terminology you’re all being confronted with is so confusing. I suspect most non-experts think that the industry has some secret decoder ring that it uses to make precise statements about what all these terms really refer to, and are endlessly frustrated when people like me keep saying otherwise. But it is, in fact, otherwise. When you look at statistics that show, for instance, that certain kinds of loans accounted for less than five percent of the mortgage market five years ago, but last year may have accounted for a third or a half or even three-quarters, what you are seeing is the development of a “nonstandard” market, one that is becoming big enough to require its own “standard” language, after the fact, as it were. It is not that a bunch of industry insiders sat down at the planning desk back in 2001-2002 and said, “let’s call this stuff “prime conforming” or “Alt-A” or “exotic” or what have you. Someone asked in the comments yesterday what “non-IO” meant. It means a loan that amortizes, as opposed to an IO loan, which does not. But why would anyone use such an odd formulation as “non-IO”? Because before the mania for IO loans, no one even used a label for “amortizing” loans. Outside of a tiny slice of a highly specialized market, they all amortized. Given a limited data set, it was not a useful characteristic to remark on. Nowadays, one of the first questions you have to ask about a loan is whether and how (positively or negatively) it amortizes. If you have never heard the phrase “positive amortization,” now you know why. It had to be made up as the opposite of “negative amortization,” even though it is also a bizarrely redundant term, like “prime conforming.”

At the end of the day, one of the reasons that the regulators are struggling to rein in the industry is, exactly, that they first have to define terms and classes of loans that do not have “traditional” meanings. They are doing what the GSEs had to do back in the 70s and 80s. And the whole process is muddled and manipulated by certain folks who have an interest in things remaining “competitive,” which you can bet is quite often code for “unregulated.” My own view is that the proponents of “efficient markets” and the magical properties of “competition” as such have a fair amount of explaining to do after the last several years. What we are seeing now in the mortgage market can, if you want, be seen as a real-life experiment in what happens when “conformity” becomes a bad word and “alternatives” are always assumed to be good things. Those of us who want to see the “grownups back in charge” aren’t really asking to take the economy back to 1983 or to underwrite every loan to GSE standards; we simply want the risk management and regulatory decisions being made by someone who has enough big-picture grasp on market processes to be able to see through marketing slogans and the psychological enticements of terms like “alternative,” and preferably has either lived through or studied periods of credit crises enough to be hard to fool with fast-talk. I wish I could tell you that I am holding my breath.