--By Mark Adelson, chief credit officer, Standard & Poor's
As in past periods of market stress, today's market dislocation is accompanied by sizable
helpings of confusion and misunderstanding. Naturally, part of the confusion relates to why market distress is occurring. More interestingly, though, part also relates to the differing degree to which it has affected different sectors of the fixed-income capital markets.
On the issue of "why," the short answer seems to be that leverage plus speculation fueled by "financial engineering" and derivatives such as collateralized debt obligations (CDOs) were the primary drivers. Financial engineers started with essentially manageable risks from the residential mortgage sector, but then concentrated and amplified those risks. Hindsight reveals that they created securities that were vulnerable to rapid and severe deterioration far beyond what market participants had anticipated.
The general media has blurred this distinction at times by using the term "subprime securities" to refer indiscriminately to both SF CDOs and first-lien subprime mortgage ABS. That confusion may have spread misunderstanding about the true causes and triggers of the current problems in the housing and financial sectors. If policymakers misunderstand the true causes, they cannot frame effective policy responses. If investors misunderstand the true causes, they cannot frame effective defensive strategies. Only with a more incisive understanding of what has happened can both groups craft effective policies and strategies.
Market Distress By Sector
First-lien subprime ABS
Contrary to popular belief, the subprime mortgage sector probably was not the primary cause of the market dislocation (rather, as noted above, the primary cause was the combination of leverage and speculation). However, the subprime mortgage sector clearly was a contributing cause, and it was the key sector that triggered the release of pent-up stresses.
The U.S. housing sector currently is experiencing a more severe contraction than at any time since the Great Depression. That, combined with lax lending standards from 2005 through 2007, is producing high levels of defaults and delinquencies on first-lien subprime mortgages. In turn, we believe that those defaults and delinquencies likely will cause widespread defaults of 2005 through 2007 vintage subprime mortgage ABS tranches initially rated at the 'BBB' and 'A' levels. That result should not be surprising given the high level of stress buffeting the U.S. housing sector. Naturally, securities rated initially at the 'BBB' and 'A' levels embody a lower measure of safety than those rated at higher rating levels.
Indeed, if securities initially rated at the 'BBB' and 'A' levels did not experience defaults resulting from the greatest level of stress in more than half a century, that would argue that they should have been rated higher in the first place. To conclude otherwise begs the question of what level of stress should 'BBB' rated securities withstand: a 100-year, 250-year, or 500-year stress?
From 2005 though 2007, total issuance of first-lien subprime mortgage ABS was around $1.1 trillion. Of that amount, around 3%, 5%, and 10% carried ratings at the 'BBB', 'A', and 'AA' rating levels, respectively. Based on current projections, 2005 through 2007 vintage first-lien subprime mortgage ABS are likely to suffer aggregate losses around $200 billion (see note 2). To be sure, those are large numbers, but they represent only a moderate fraction of the total issuance. As noted above, we expect that tranches that initially carried 'AAA' ratings are likely to suffer few defaults; most had sufficient credit support to withstand losses in excess of 23% and, starting in third-quarter 2006, many had sufficient credit support to withstand losses in excess of 28%. Thus, the overall performance of subprime mortgage ABS is reasonably consistent with the severity of stress that the housing market is experiencing.
Some additional figures help put the size of the "subprime issue" in perspective. At the end of 2005, there were around 109 million occupied housing units in the U.S. Renters occupied 34 million units, while 75 million units were owner-occupied. Of the 75 million owner-occupied units, 25 million were owned free and clear. Fifty million owner-occupied units had mortgage debt. Of those, roughly 44 million had prime mortgages, and around 6 million had subprime mortgages. Of the 6 million households with subprime mortgages, we estimate that the homeowners of between 1.5 million and 3 million are likely to default on their loans and either have loan modifications or lose their homes. Accordingly, by this measure, the "subprime issue" directly affects fewer than 3% of U.S. households.
Why, then, has it become such a headline story? There are two main reasons. First, the subprime situation is connected to the bursting of the U.S. real estate bubble. The inflation of the bubble, during the early 2000s, was one of the key factors that led to the expansion of subprime lending. However, the two are not the same thing (see note 3).
Second, key changes in the mortgage business occurred while the housing bubble was inflating. Structured finance assets--particularly subprime mortgage ABS--became the main asset class backing CDOs. Demand from the CDO sector pushed spreads tighter on 'BBB' rated subprime mortgage ABS. By the end of 2005, CDO demand for the securities was so strong that it drove other investors and bond insurers out of the sector. Now, in hindsight, it appears that without the moderating influences of traditional investors and the bond insurers, subprime mortgage originators became increasingly lax in their underwriting practices and started to make riskier loans.
Thus, in the subprime mortgage sector, the severe stresses from the bursting of the housing bubble were aggravated by the effects of lax underwriting practices starting in 2005.
Still, as noted above, current projections suggest to us that the vast majority of first-lien subprime mortgage ABS--those initially rated at the 'AAA' level--should not suffer high levels of defaults. And, finally, for those that do default, the severity of loss is expected to be quite small.
Second-lien ABS
The performance of second-lien ABS is notably poorer. Many of the tranches initially rated 'AAA' from deals issued from 2005 through 2007 are at risk of defaulting. Fortunately, the dollar volume of such deals--roughly $150 billion--is much smaller than the volume of first-lien subprime mortgage ABS during the period.
Although the housing and residential mortgage sectors are experiencing significant stress, we do not believe that today's conditions are sufficiently extreme, by themselves, to explain widespread defaults of securities initially rated at the 'AAA' level. So, what happened in the second-lien area? In our opinion, second-lien mortgage lending changed during the mid-2000s--and past experience gave a false signal about future performance under stressed conditions. One key change was an increase in so-called "piggyback seconds," which are purchase-money second-lien loans. Another key change, related to the first, was significantly higher combined loan-to-value ratios (CLTVs) in second-lien loans.
A third key change was in consumer behavior. Around the time home prices started to decline (late 2006), consumers became increasingly willing to default on their second-lien loans. They realized that second-lien lenders were unlikely to initiate foreclosures. Defaulting consumers concluded that damage to their credit reports was not a consequence that was bad enough to deter them from defaulting. The old-fashioned notion that defaulting would bring disgrace, shame, dishonor, and stigma appears to have been supplanted by the idea that the decision to default is purely an economic one.
Although the first two changes were visible at the loans' origination, the third was not apparent until performance of the loans could be observed. The combined effect of all three was an unprecedented increase in defaults on second-lien mortgage loans. The early 2000s' "conventional wisdom" about second-lien mortgage loans--which virtually all market participants embraced, including investment banks, issuers, and investors--viewed the product as quite risky and requiring higher credit support levels than any other residential mortgage product. However, the historical experience upon which the "conventional wisdom" had rested had no precedent for what subsequently happened. The combination of more piggyback loans, higher CLTVs, and changing consumer behavior produced never-before-seen levels of risk in second-lien loans. The result was that virtually everyone in the sector was blindsided.
Fortunately for the capital markets, the second-lien mortgage ABS sector is much smaller than the first-lien subprime sector. As noted above, from 2005 through 2007, the total issuance of second-lien mortgage ABS was only around $150 billion, compared with more than $1 trillion for the first-lien subprime mortgage sector. Thus, although the intensity of credit problems is very high--reaching the senior, 'AAA' level of many deals--the total impact is not significantly greater than that within the first-lien subprime sector.
SF CDOs
The SF CDO sector has suffered the worst of all. Starting around 2004, SF CDO issuance volume started to increase rapidly. Professionals in the sector fueled the expansion by creating SF CDOs backed primarily by subordinate tranches of subprime mortgage ABS deals and second-lien mortgage ABS deals. In fact, the demand from SF CDOs for subprime mortgage ABS was so strong that it outpaced the available supply. To satisfy the higher demand, investment banks started creating "synthetic" ABS using derivatives known as credit default swaps (CDSs). Ultimately, the volume of synthetic ABS in the SF CDOs exceeded the volume of actual securities by a factor of roughly 4 to 1.
Many of the SF CDOs focused primarily on subprime mortgage ABS rated initially at the 'BBB' level. Those SF CDOs were dubbed "mezzanine" deals. Others focused primarily on securities rated at the 'A' and, to a lesser degree, 'AA' levels. Those were called "high-grade" SF CDOs. Current projections suggest that the vast majority of mezzanine SF CDOs will suffer defaults into their tranches that initially carried ratings at the 'AAA' level. Although the matter is somewhat less clear, we believe that a high proportion of high-grade SF CDOs is likely to suffer the same fate. We expect that the dollar volume of defaulting securities initially rated at the 'AAA' level, as well as the magnitude of losses on those securities, is likely to exceed the volumes of defaults and losses on 'AAA' securities from the first-lien subprime mortgage and second-lien mortgage ABS sectors.
When the dust finally settles from the current market dislocation, we believe that losses on SF CDOs likely will amount to between $200 billion and $300 billion, including a notable proportion of securities initially rated at the 'AAA' level. Those will represent the major share of losses booked by financial institutions and other market participants.
By comparison, we believe that losses on second-lien mortgage ABS (including those backing SF CDOs) likely will amount to roughly $50 billion, also including a high proportion of securities initially rated at the 'AAA' level. As noted above, losses on first-lien mortgage ABS (including those backing SF CDOs) likely will amount to around $200 billion, very little of which will have an impact on securities initially rated at the 'AAA' level. A large share of those losses will not be felt directly by investors, but only indirectly because the securities were repackaged into SF CDOs.
Ramifications
Although many securities have already suffered severe deterioration of credit quality, most of them have not yet suffered payment defaults. Nonetheless, a number of institutions have booked staggering writedowns attributable to SF CDOs (see note 4). For example, from third-quarter 2007 through second-quarter 2008, Citigroup announced total writedowns of roughly $25 billion attributable to SF CDOs (see note 5).
Likewise, Merrill Lynch reported net writedowns attributable to SF CDOs of $23.2 billion for 2007. Later, Merrill Lynch reported SF CDO-related writedowns of more than $3.0 billion for first-quarter 2008 and of around $5.0 billion for second-quarter 2008 (see note 6). At the end of July, Merrill Lynch sold roughly $30 billion (face amount) of its remaining SF CDO positions at a price of just 22 cents on the dollar, and, in doing so, financed 75% of the purchase price, receiving only around 5 cents on the dollar in cash.
UBS, a major Swiss financial institution, had a similar experience. It booked SF CDO-related writedowns of $12.8 billion for 2007 (see note 7). Then it booked further SF CDO-related writedowns of $1.5 billion in first-quarter 2008 and $2.2 billion in second-quarter 2008.
Apart from their impact on banks and securities firms, SF CDOs have been a major factor in the deterioration of the bond insurance sector. Most of the major players in that sector have suffered downgrades in the past year. Only a few have managed to retain 'AAA' ratings.
The overall result has been a significant dislocation of financial markets, including a contraction of liquidity for all types of complex and esoteric assets. Liquidity has contracted for simpler instruments as well, though to a somewhat lesser degree. Bid-ask spreads for everything except instruments of the highest perceived quality have become very wide. This, in turn, has caused a noticeable tightening of credit in both the consumer and commercial spheres.
Silver Linings
On the other hand, the news is not all bad. First, many SF CDOs included CDSs on subprime mortgage ABS rather than the actual securities. Because a CDS is a two-party contract, every dollar lost by one party is a gain for the other (see note 8). Thus, for every dollar of SF CDO-related losses attributable to CDS, there ought to be an offsetting dollar of gain elsewhere in the global financial system. Unfortunately, losses, but not the offsetting gains, seem to have been concentrated in many institutions that are key players in maintaining the smooth operation of the global financial system.
Second, the 2008 market dislocation has revealed some of the systemic risks posed by high leverage and speculation fueled by financial engineering and derivatives. Indeed, some commentators have recently argued that the over-the-counter (OTC) market model for fixed-income derivatives and other complex instruments is largely to blame for the current problems (see note 9). Identifying problems is the first step to solving them. Recognition of the role that complex derivatives and financial engineering have played may give rise to wise policy responses that protect the capital markets from similar disruptions in the future.
Finally, from the narrow standpoint of the U.S., sales of SF CDOs to investors outside the U.S. may have exported some of the losses related to the bursting of the U.S. housing bubble. This has the potential effect of softening the blow to the U.S. economy and the U.S. financial system.
Takeaways
The subprime problem, the housing bubble, and the market dislocation are related, but they are distinct events. Understanding their interrelationship is essential for identifying the weak points of the financial system and for framing defensive strategies (for investors) and corrective policies (for policymakers).
The financial markets have the opportunity to learn from recent experience and to become stronger. It now remains to be seen whether investors and policymakers will heed the lessons and seize the opportunities. At a minimum, all types of market participants should now be on notice that leverage and speculation born of financial engineering and complex derivative activities can produce unanticipated results.
Notes
(1) The fate of the tranches initially rated at the 'AA' level is the most uncertain. Those are the tranches on the "cusp" of current projections. In our view, there is a reasonable chance that a significant proportion of those tranches could default. It is debatable whether the stress in the housing and residential mortgage sectors is severe enough, by itself, to explain defaults of such securities.
(2) Losses on the underlying subprime mortgage loans likely will be somewhat higher. A portion of the losses on the loans are absorbed by "excess spread" within each deal. The result is that ultimate losses on the securities are lower than the losses on the related loans.
(3) The housing bubble itself merits headline attention because it affects all of the 75 million households that own their own homes. Beyond the mortgage sphere, declining home prices can have important effects because they reduce both household wealth and consumer spending.
(4) To date, most of the writedowns reflect market value losses rather than actual credit losses. In the short run, market value losses can exceed projected credit losses because of supply-demand imbalance in the secondary market.
(5) Citigroup earnings announcements report the quarterly SF CDO writedowns as follows: third-quarter 2007: $1.8 billion; fourth-quarter 2007: $14.3 billion; first-quarter 2008: $5.7 billion; and second-quarter 2008: $3.2 billion.
(6) For first-quarter 2008, Merrill Lynch had net writedowns of $1.5 billion on SF CDOs and $3.0 billion of credit valuation adjustments related to hedges with financial guarantors, most of which relate to U.S. super-senior SF CDOs. For second-quarter 2008, Merrill Lynch had net writedowns of $3.5 billion on SF CDOs and around $1.5 billion of credit valuation adjustments related to hedges with financial guarantors, most of which relate to U.S. super-senior SF CDOs.
(7) According to a special report to shareholders, dated April 18, 2008, about two-thirds of total UBS losses in 2007 were attributable to the bank's CDO desk. About half of the bank's total losses for the year came from super-senior CDO positions.
(8) The equivalence of gains and losses under a CDS breaks down if the party obligated to pay defaults on its obligation. Collateral-posting requirements mitigate that risk in many CDSs, but they do not fully eliminate it.
(9) Whalen, C., "Yield to Commission: Is an OTC Market Model to Blame for Growing Systemic Risk?" J. of Structured Fin., vol. 14, no. 2 (Summer 2008).
Analytic services provided by Standard & Poor's Ratings Services (Ratings Services) are the result of separate activities designed to preserve the independence and objectivity of ratings opinions. The credit ratings and observations contained herein are solely statements of opinion and not statements of fact or recommendations to purchase, hold, or sell any securities or make any other investment decisions. Accordingly, any user of the information contained herein should not rely on any credit rating or other opinion contained herein in making any investment decision. Ratings are based on information received by Ratings Services. Other divisions of Standard & Poor's may have information that is not available to Ratings Services. Standard & Poor's has established policies and procedures to maintain the confidentiality of non-public information received during the ratings process.
Ratings Services receives compensation for its ratings. Such compensation is normally paid either by the issuers of such securities or third parties participating in marketing the securities. While Standard & Poor's reserves the right to disseminate the rating, it receives no payment for doing so, except for subscriptions to its publications. Additional information about our ratings fees is available at www.standardandpoors.com/usratingsfees.
Any Passwords/user IDs issued by S&P to users are single user-dedicated and may ONLY be used by the individual to whom they have been assigned. No sharing of passwords/user IDs and no simultaneous access via the same password/user ID is permitted. To reprint, translate, or use the data or information other than as provided herein, contact Client Services, 55 Water Street, New York, NY 10041; (1)212.438.9823 or by e-mail to: research_request@standardandpoors.com.