It will surprise no regular reader of these pages that a major part of the tremendous housing boom the United States has experienced was played by the invention and marketing of mortgage backed securities (MBS). Fans of the old Michael Lewis (who know that the new Michael Lewis seems to have been transformed into a bitter journalist hack, is the subject of an alien abduction and impostor replacement scheme, or perhaps has simply been browbeaten by life into a poorly disguised and very tired political operative) will delight in remembering his descriptions of the founding of Salomon Brothers' mortgage security department (the first on The Street) and the uphill battle Salomon sales folk fought to popularize the instruments. Lewis recounts the "skit" that Salomon sales people acted out for his trainee class in Liar's Poker:
Two played Salomon salesmen; two played managers of a thrift. The plot ran as follows: The Salomon salesmen enter the thrift just as the thrift managers are leaving, tennis racket in one hand and a bag of golf clubs in the other. The thrift men wear absurd combinations of checkered pants and checkered polyester jackets with wide lapels.
[...]
The Salomon salesmen, having schmoozed their client, move in to finish him off. They recommend that the thrift managers buy a billion dollars' worth of interest rate swaps. The thrift managers clearly don't know what an interest rate swap is; they look at each other and shrug. One of the Salomon salesmen tries to explain. The thrift managers don't want to hear; they want to play golf. But the Salomon salesmen have them by the short hairs and won't let go. "Just give us a billion of them interest rate swaps, so we can be off," the thrift managers finally say. End of skit.
That was the sort of person who dealt in home mortgages, a mere sheep rancher next to the hotshot cowboys on Wall Street.
Lewis is being a bit tongue in cheek, but the sophistication spread between the characters he paints is quite understandable. The business of writing home mortgages is, as Lewis aptly describes it, messy. Personal. "Down home." The asset itself, the collateral, is anything but fungible, or liquid. The debtors, too, are anything but fungible and (mostly useless attempts at credit scoring notwithstanding) subject to a great opacity with respect to their ultimate creditworthiness.
And this, I think, is an interesting point. The opacity that accompanies attempts to liquefy the illiquid and its impact on markets, particularly when they have to unwind.
Even messier than their creditworthiness, mortgage debtors are prone, in good times anyhow, to prepay the entire debt or to refinance at a lower rate. There is nothing more annoying to a long-term, fixed income investor than a fixed income instrument that might, randomly, just prepay one day. Just when those interest rates were looking good you get left with a pile of cash and, should you reenter the market, lower returns.
Short of prepayment, there is "curtailment," where debtors make larger payments than scheduled, reducing their debt expense, and, again, complicate the cash flow characteristics of the underlying instrument.
The key thing to remember about both of these effects is that they boost early cash flow (but reduce capital at work) in good times. More importantly, they can have the effect of a net reduction in cash flow in bad times, relative to good times- exactly when you don't want cash flow trending down. Modeling these cash flows is, of course, a non-trivial exercise.
Then there is the fact that people sign mortgages when they are ready to, not on "mortgage signing day," once a year. That means maturities for each mortgage, obviously, vary. Again, this is messy.
All of this conspires to keep them illiquid, and reinforce their tendency to hang around on balance sheets and increase reserve requirements for institutions holding them.
Mortgages also just don't have much scale. To get enough size to be interesting for large scale financial instrument work you have to pool a bunch of them. Once that is done you have yet another layer of opacity. What market actor (except maybe Bill Ackman) is going to poke around and analyze all those individual mortgages?
Of course, pools are appealing in some sense, as slaves to normal distributions will rush to point out, because of the diversification against any single mortgage default they create. So the effort becomes to develop a set of characteristics, standards, that describe a given pool. To "de-unique" them, a prerequisite to introducing some liquidity to the asset class.
How might we classify mortgages? Size of the underlying mortgages, perhaps. Interest rates, of course, are bracketed as descriptors. "Weighted average maturities," which are dollar amount weightings of the time to maturity of the loans in a pool. Prepayment speed. Current loan balance (Low, below $80,000 or $85,000 to Super High, upwards of $150,000 or $175,000). And, of course, in the never ending spirit of hopelessly simplifying the intrinsically complex, ratings- themselves based on a given set of characteristics and standards, "prime" and "subprime" being the main actors in that play.
Supposing a residential mortgage backed security (RMBS) has thousands of collected debtors (a good bet) it represents a somewhat arbitrary classification of an otherwise widely diverse bunch of counterparties. And, technically, investing in an MBS isn't a cash flow investment. It is an investment in the underlying collateral. This is problematic, even if you do have a good model for predicting the ultimate cash flows.
You can quickly see where this leads. Blending different MBS instruments into a portfolio helps if you want exposure to the asset class, but it doesn't change the cash flow issues, and diversification can be time consuming. You either have to monitor a wide variety of MBS instruments (and along with all the portfolio optimization and balancing issues here you can complicate things even more by throwing in models to try to find uncorrelated MBS instrument classes and build a efficient portfolio) all the time, or you have to find a new job. Wouldn't it be nice if some nice manager did all that work for you and gave you predictable cash flows? Something linked to LIBOR for instance? Wouldn't it also be great if you could get the MBS assets off your balance sheet so you could reduce your reserve requirement but still be exposed to (and enjoy the "safe" returns of) the asset class? Enter the CDO.
CDOs reduce the number of underlying instruments in the pool to 200 or less, permit the issuer to use leverage to boost return, earn a management fee and to retain any excess return after the CDO notes, issued against the RMBSs or other securities, are paid. These notes should, therefore, enhance liquidity since they are "un-unique," and (hurray!) have reduced balance sheet impact.
CDOs also introduce another opaque layer. What does a buyer of instruments from a CDO know about the underlying asset pools? Let's just see:
It is a condition to the issuance of the Notes that the Class A Notes be rated "AAA" by Standard & Poor's Ratings Services, a division of The McGraw-Hill Companies, Inc. ("Standard & Poor's") and "Aaa" by Moody's Investors Service, Inc. ("Moody's," together with Standard & Poor's, the "Rating Agencies"), that the Class A2 Notes be rated "AAA" by Standard & Poor's and "Aaa" by Moody's, that the Class A3 Notes be rated at least "AA " by Standard & Poor's and at least "Aa2" by Moody's, that the Class A4 Notes be rated at least "A" by Standard & Poor's and at least "A2" by Moody's, that the Class B Notes be rated at least "BBB+" by Standard & Poor's and at least Baa1 by Moody's, that the Class B2 Notes be rated at least "BBB" by Standard & Poor's and at least "Baa2," by Moody's, that the Class B3 Notes be rated at least "BBB-" by Standard & Poor's and at least "Baa3" by Moody's, and that the Class C Subordinated Notes be rated at least "BBB-" by Standard & Poor's. The Class D Subordinated Notes will not be rated by any Rating Agency.
Not much. Time to play golf yet?
As you can see from the Placement Memorandum excerpt above, issuers of CDOs quickly figured out that they could tranche out the returns of their pool in different instruments. I've touched on the motivations for that sort of split before, citing a paper referenced some time ago by Alea. In short:
For those issues where our model predicts a higher optimal number of tranches, we find that additional uniquely-rated tranches are associated with higher prices for the issue as a whole. This suggests that structuring is allowing issuers to exploit market factors – such as greater investor sophistication and heterogeneous screening skills related to asymmetric information – to their advantage via tranching.
All this lends the appearance of liquidity. We have "fungibled" the unfungible and liquefied formerly illiquid assets. But, just a second there. There is a big difference between "potentially liquid" and "liquid." True, they have all the ingredients of liquidity now, but that does not a market make. I touched on this issue as well with reference to corporate and municipal issues, pointing out via the always interesting "Accrued Interest" that:
According to TRACE, which is a system for tracking corporate bond trading, about 4,300 TRACE eligible corporate issues trade each day. That's less than 3% of the total corporate universe. I note that trading activity in non-TRACE eligible bonds is not publicly reported. According to the MSRB, about 14,900 municipal bonds trade each day, or just over 1% of the universe.
This starts to make matters interesting when you delve into what I will crudely call "second order liquidity." What happens when you have to unwind a CDO? Whatever the liquidity of its instruments, what is the liquidity of the underlying assets, the collateral? What is that composed of, and how liquid is it?
And this is where I return to the concept of "opacity."
So, traveling up the securitization chain, which opacity membranes do we pass through?
First, between the debtor and the lender: The character of the ultimate debtor. Reduced to credit score (a somewhat mysterious, black art sort of ranking). Between the lender and the collateral, the quality of the collateral itself.
Second, between the lender and the pool sponsor, the quality of collateral and debtor, further complicated by the nuances of individual mortgage originator policies on credit and credit verification- for which no easy classifications exist. There might actually be an additional level of opacity here, if the lender used an "outsourced" origination mechanism (bet on it). As to the quality of the collateral... as we move through opaque membranes, this is reduced almost entirely to purchase price (mortgage size, Jumbos, etc. etc.). If we tried, could we possibly think of a more dangerous metric for collateral quality than acquisition price? One that would have an even more exaggerated value inflation effect during a bubble?
Third, between pool sponsor and CDO issuer, just about anything about the basic collateral, aside from generalized descriptive characteristics of the RMBS instruments. A reduction in creditworthiness score to "prime" or "subprime," is now evident.
Fourth, between CDO issuer and everyone else, anything specific at all about underlying collateral, including what the portfolio actually holds look today, to the extent the CDO manager is even remotely active. These are reduced to ratings agency assessments of the instruments issued by the CDO (which, we assume, involve some scrutiny of the underlying assets and liabilities, albeit still obscured below the structure of RBMS instruments). The underlying liquidity, two levels lower, is also effectively masked, unless the holder of CDO issued notes has a in-depth understanding of the CDO pool, what instruments compose it, the markets they trade in and their liquidity properties (read: unlikely).
Astute Going Private readers will enjoy these sorts of structural questions, understand their intrinsic connection with issues of structure and leverage, the impact of illiquidity on markets, and a careful analysis of incentives- looking each step of the way for information asymmetries (and what do you think "opacity" is exactly?) as they formulate the investment thesis for their next shorting opportunity in multi-tiered, opacity laden, "liquefied" "securitizations."