After a record decade of economic growth in the U.S., the current expansion is becoming long in the tooth. S&P Global Ratings economists now place the probability of a U.S. recession within 12 months at between 25% and 30%. Although our baseline economic view is neutral, market participants are attempting to identify the most likely source of a potential economic stall, and some have turned their attention to the set of financial instruments that they believe were the primary cause of the 2007–2009 global financial crisis: U.S. structured finance products.

Broadly speaking, this $11.5 trillion market of securitized products includes bonds backed by cash flows from mortgages, consumer and commercial receivables, leveraged loans, and more. In terms of outstanding debt in the U.S., structured finance is second only to Treasury debt (over $16 trillion). Structured finance has been a widely used funding mechanism for over 30 years and is an established and well-developed part of U.S. fixed income markets that is likely to persist for years to come.

S&P Global Ratings recently undertook a stress test to estimate the impact of a potential market downturn on structured finance ratings. In this hypothetical recession scenario constructed to transpire over a three-year period, U.S. GDP contracts and growth stays negative for six quarters, unemployment exceeds 6% in the second and third years, and both U.S. fixed investment and export growth are assumed to be negative for the first two years. The contemplated recession affects global markets as trade slows and commodities markets weaken. Under this stress scenario, downgrade and default risks for most structured finance sectors and regions are anticipated to be contained to speculative-grade classes (rated 'BB+' or lower), with some risk in low-investment-grade classes, particularly in collateralized loan obligations (CLOs). This topic is explored in our Sept. 4, 2019, publication, "When The Cycle Turns: How Would Global Structured Finance Fare In A Downturn?".

The crisis took a particular toll on the performance and valuation of many securitized products, even some that were not connected with the mortgage market. Apart from Treasury debt, most fixed-income products experienced spread widening, which was problematic for investors with limited liquidity and relatively short-term investment horizons. Because many investors were seeking safety and liquidity at the time, a large segment shied away from structured products in general. As chart 1 shows, issuance of most securitized products declined during and after the crisis--especially non-agency residential mortgage-backed securities (RMBS).

Chart 1

This trend is also reflected by the general post-crisis decline in the degree to which securitization is used as a tool to fund assets--such as cars and leveraged loans--called the utilization rate, shown in chart 2, and defined in terms of outstanding securities divided by total outstanding funding for a given asset (houses, cars, etc.) at a point in time.

Chart 2

Ten Years Onward

During the years leading up to the crisis, U.S. structured finance annual issuance expanded to almost $3.0 trillion from under $1 trillion in 2000, with much of the growth in the mid-2000s coming from non-agency RMBS. Today, overall issuance has returned to levels comparable to those of the early 2000s but remains below the levels immediately pre-crisis (see chart 1). Once the crisis was in full swing in 2008, new issuance of structured finance fell sharply to under $1.5 trillion, with $1.2 trillion coming from agency RMBS and non-agency RMBS contributing little or nothing at all to the total. Only recently has structured finance other than agency RMBS made a comeback.

Many of the major structured finance asset classes still exist and generally resemble their forebears. Some structured products have seen improved collateral, although the changes haven't been uniform across all asset classes. For example, CLOs, which performed well through the crisis with relatively few defaults, have undergone only structural changes. Non-agency RMBS, on the other hand, are no longer backed by the type of subprime and Alt-A collateral that was widespread prior to the crisis (although non-prime and limited documentation mortgages still collateralize a segment of the RMBS market).

The emergence of new regulations has transformed the structured finance market by better aligning the interests of investors and issuers. This has been achieved through the requirement that issuers of many structured finance products in the U.S. must now maintain "skin in the game" through risk retention. A notable exception is that of broadly syndicated loan (BSL) CLO transactions, which are exempt from the risk-retention rules following a court ruling in 2018.

The broadest and most important of the new regulations is the Dodd-Frank Act, which is intended to protect consumers and to regulate banks and lenders in a way that reduces the risk of another major economic downturn. Many of the post-crisis rules, which affect structured finance to varying degrees, are shown in the timeline in chart 3.

Chart 3

In terms of transaction structure, credit enhancement has generally increased, providing increased credit protection for investors in 'AAA' rated classes of many structured products. The post-crisis era boasts more upgrades than any period since 2001, due in part to the healthy economy. To some extent, it also reflects the increased prevalence of sequential structuring. In a sequential structure, senior bonds are paid down before junior bonds receive cash flows--a feature that allows for a buildup of credit enhancement over time.

The Asset Types

Chart 4

The RMBS market

Issuance and the shift in utilization 

The utilization rate for housing has always been higher than that of other assets, save leveraged loans, and has drifted up to 70% as of mid-2019 from just under 60% in 2000. Moreover, given the size of the housing market (U.S. mortgage debt is currently about $11 trillion), the outstanding amount of RMBS is substantially greater than all the other structured products in the U.S. combined. While the majority of mortgage funding in the U.S. has come from securitization markets, it is important to distinguish between the two different avenues for mortgage securitization: agency and non-agency. The high housing utilization rate is driven primarily by the agency sector. However, in the run up to the crisis, the utilization rate for non-agency started to approach that of agency. The total RMBS utilization rate, as well as the break-out between agency and non-agency are shown on chart 2.

Two of the major agencies that securitize home loan pools across the country are the government-sponsored enterprises (GSEs) Fannie Mae and Freddie Mac, which are regulated by the Federal Housing Finance Agency (FHFA) and exist to promote homeownership and ensure funding at the best possible rates for borrowers. Mortgages securitized by these GSEs need to conform to specific guidelines that concern credit quality and loan size. One of the hallmarks of RMBS issued by Fannie Mae and Freddie Mac is that the principal and interest payments on the securities are guaranteed by these GSEs. Fannie Mae and Freddie Mac are in turn generally believed to be implicitly backed by the U.S. government. This means that, barring U.S. sovereign risk, the broad market perception is that GSE RMBS present limited credit risk to the investor. Rather, it is interest rate risk (in its various forms) that the investor needs to manage. Credit risk transfer (CRT) deals are an exception. These allow the GSEs to shed credit risk by selling it to the private market. At roughly $20 billion per year, annual issuance in this subsector is relatively small compared to overall agency volume and, because it carries credit risk, it is often classified as non-agency.

Non-agency RMBS do carry credit risk, which is derived from the collateral and redistributed to investors via structural mechanisms similar to those used in other securitized products. While there has been non-agency securitization since the 1980s, there was a marked increase between 2004 and 2006 when approximately half of the securitized home funding in the U.S. came from the non-agency sector (as measured by new issuance RMBS volume). The bulk of this funding came from the subprime and alt-A subsectors, which were characterized by loans to borrowers with weaker credit histories and limited loan documentation.

Just prior to the crisis, the U.S. non-agency market saw over $1 trillion in issuance, roughly on par with agency issuance. Issuance was minimal post-crisis and has returned slowly. Since 2014, the non-agency market has grown, although it has not come close to its 2004-2006 dominance, as the subprime and Alt-A spaces have decreased in prominence. In 2018, there was $95 billion in non-agency issuance, and in 2019, there was $124 billion of issuance volume.

Mortgage collateral quality 

Prior to the slowdown of the U.S. housing market in mid-2006, mortgage lending credit criteria had loosened, and the industry began to experience increasingly widespread underwriting anomalies. Importantly, borrowers relied heavily on the availability of short-term consumer financing and continued home price appreciation. As the variability in underwriting standards increased, the U.S. homeownership rate hit an all-time high of over 69% in 2004 (well above the historical norm of just under 65%), and prices grew at an unsustainable pace. Much of the embedded default risk in non-agency RMBS was masked through home price appreciation and fast borrower equity build-up. When there were defaults, severities were low because the housing market was strong.

When home prices fell sharply across the country in 2007, many overleveraged borrowers started to miss mortgage payments and eventually become delinquent on their loans. Delinquencies and, later, defaults and losses cascaded through many of the securitizations that relied on mortgages--especially subprime lending programs. Over the next several months, what began as a seemingly contained U.S. housing market correction turned into a liquidity crisis that affected broader financial markets and eventually led to a global recession.

In the years after the crisis, the credit quality of mortgages underlying securitizations has broadly improved. Among the dominant non-agency RMBS products are jumbo (which boasts of a particularly strong credit profile), re-performing loan deals, CRT (described above), and non-qualified mortgages (non-QM). The last is the fastest-growing segment, but it remains relatively small, with $25 billion in non-QM RMBS issuance in 2019.

Mortgages without certain prescribed risky features, as defined by the Consumer Financial Protection Bureau (CFPB), are considered QM loans. That is, they conform to the QM rule set forth by the CFPB and are presumed to comply with the ability to repay (ATR) rule, which the CFPB introduced in early 2014 to help lenders assess a home buyer's ability to repay a mortgage loan. While non-QM loans still need to satisfy the ATR rule, they deviate in one or more ways from the CFPB's QM rule. Originators of QM loans benefit from a degree of protection against potential buyback claims and other legal liabilities.

While some of the non-QM collateral is in certain ways weaker than that of prime, the loans appear to be characterized by an absence of risk-layering and by offsetting risk factors. For example, if the loan documentation is weaker than normal, the loan-to-value (LTV) ratio might be commensurately lower. Moreover, the interest rate spread above prime for non-QM loans is substantially greater than what was observed for subprime. Non-QM loans also tend to have better credit profiles as measured by the FICO score and LTV ratio.

Structural changes and risk retention 

Credit enhancement for 'AAA' rated non-QM bonds is in the neighborhood of 30%. This is in contrast to the average credit enhancement of just over 20% for 'AAA' rated subprime bonds prior to the crisis. Non-agency RMBS has generally become more robust, especially for the higher-rated tranches. RMBS deals backed by non-QM collateral typically adopt a mixture of pro rata and sequential payoff structures. At deal inception, the senior bonds are paid pro rata and the subordinate bonds are locked out from receiving any principal payments. If certain triggers fail, however, even the senior tranches adopt a sequential prioritization. Unless the collateral conforms to the QM rule, RMBS issuers generally need to retain an interest in the bonds they produce. This can take the form of either a vertical or a horizontal slice of 5% ownership.

For a detailed comparison between non-QM RMBS and subprime, alt-A, and prime jumbo during two pre-crisis periods, see our Dec. 4, 2017, commentary "How Do Non-QM Loans Stack Up Against Pre-Recession Mortgage Products?" For a general overview of trends in that market, see our Sept. 20, 2019, commentary "Non-QM's Meteoric Rise Is Leading The Private-Label RMBS Comeback."

CDOs versus CLOs

Collateral, not structure, is the key distinction 

While similar in name and even structure to CLOs, collateralized debt obligation (CDOs) should not be confused with CLOs. CDOs were a type of structured finance instrument that proliferated pre-crisis. Because they were primarily supported by U.S. RMBS (including subprime RMBS), they were susceptible to residential housing market risk. CLOs, on the other hand, are securitized pools of senior secured loans to relatively highly leveraged companies. Thus, CDOs and CLOs are securitized by different collateral with different risk characteristics.

CDOs and CLOs do share a number of structural characteristics. The transaction is structured using a special purpose legal entity, which owns the assets that produce cash flows. For both products, the transaction typically has a manager and is financed through the issuance of debt at different tranche levels, as well as a small equity class. Like many structured finance vehicles, both CDOs and CLOs typically had debt issued according to a senior-subordinate structure, with the cash flows from the underlying assets distributed via a prescribed "waterfall"--paying the most senior note classes first, followed by the more junior notes. While all classes of debt issued by a CDO or CLO are supported by the same asset pool, the senior tranches typically have first claim to the cash flows generated by the collateral.

In order to protect the quality of the issued notes and investor capital, both the CDO and CLO structures employ regularly scheduled coverage tests, which include asset coverage (overcollateralization) tests and interest coverage tests at various tranche levels. Typically, if one of the tests fails at given tranche level, interest (and, if needed, principal) payments are diverted from subordinate tranches and used to reduce the balance of senior notes outstanding until the failing test is restored to compliance. Missed interest payments to the subordinate tranches are capitalized and added to the tranche balance, to be paid back with interest-on-interest if the test failure is cured and cash flows are restored to the subordinate tranches under the failing tranche. If par losses are to be suffered, the CDO or CLO equity is first in line, followed by the junior-most subordinate tranches and continuing up the capital stack.

CDO and CLO performance 

The RMBS CDOs issued before the crisis had similar structural features as CLOs, but the losses from the underlying RMBS collateral were severe enough to overwhelm the structural protections in the CDO transactions and produce defaults across the CDO capital stack, including, in many cases, at the 'AAA' CDO tranche level. U.S. RMBS and CDO ratings criteria at the time of the financial crisis contemplated substantial declines in home prices; however, the actual deterioration of the U.S. residential real estate market was more significant than we and others had anticipated and was more severe than we had associated with the contemporaneous macroeconomic stress. In addition, defaults and losses on residential mortgage loans displayed unanticipated sensitivity to declining home prices.

By way of contrast, rated CLO tranche defaults during and after the crisis were limited. As of the end of second-quarter 2019, out of more than 11,000 U.S. CLO tranches rated by S&P Global Ratings over the past 25 years, there were only 40 defaulting tranches for U.S. CLO transactions. No cash flow CLO tranche with an S&P Global Ratings original rating of 'AAA' has defaulted. Only one CLO tranche with an S&P Global Ratings original 'AA' rating has defaulted, and the noteholders in this tranche suffered no monetary loss after the default was cured. (See Twenty-Five Years Strong: Update On CLO 1.0 Defaults, Aug. 12, 2019.)

Derivatives: How the risk was spread 

The impact of the housing downturn was exacerbated by derivative instruments. Credit default swaps (CDS) grew the effective subprime investor base to a scale that was much larger than it originally appeared. CDS are essentially insurance contracts on securities that one may or may not own. Prior to the crisis, it was common to use CDS to, in essence, make bets on the outcome of subprime performance. This could be done by simply entering into a CDS contract directly (such as those that were listed in the ABX index) or by buying a tranche of a synthetic CDO.

The U.S. synthetic CDO derived cash flows not from underlying bonds but from CDS on a basket of reference assets, which in many cases were subprime RMBS. The size of the synthetic CDO market grew in the years leading up to the crisis and eventually overtook the transaction volume of traditional cash flow CDOs.

In a synthetic CDO, the issuing special purpose vehicle sells protection on the reference portfolio (via CDS), and the premiums paid by the CDS counterparties (who are buyers of credit protection) provide cash flows to investors in the synthetic CDO. As credit events occur (i.e., defaults or losses on underlying RMBS), contingent payments reduce the cash flow to synthetic CDO investors. If credit events are particularly severe, not only would the more traditional "funded" classes lose their principal, but the investors in so-called "unfunded" classes could also be forced to make contingent payments. The synthetic CDO, which contributed to the contagion of the crisis to the broader financial system, has disappeared from the rated market since the crisis. Chart 1 shows that CDO issuance grew substantially between 2005 and 2007, after which it dropped sharply and ceased to be a material part of the structured finance market.

Market demand for CLOs persists, but certain risks are increasing 

Unlike CDOs, CLOs have no connection to subprime RMBS or any mortgage-related collateral at all. The vast majority of U.S. CLOs today are collateralized entirely by corporate loans, although recently proposed changes to the Volker Rule would allow for a 5% bucket of corporate bonds.

Today, CLOs make up the roughly a quarter of all U.S. outstanding structured finance (excluding agency RMBS)--equivalent to the amount of auto, credit card, equipment, and student loan securities and other ABS combined. While there was an abundance of CLO downgrades (many due to changes in rating agency criteria) during the crisis, there were few CLO tranche defaults. Nearly all of the pre-crisis generation of CLOs (referred to as "CLO 1.0" transactions) have paid down, and in their stead are the even more structurally robust "CLO 2.0s."

For more than 15 years, the leveraged loan market has been largely funded by CLOs and continues to be today, even as the leveraged loan market has more than doubled in size since the end of 2007 (outstanding leverage loans total around $1.2 trillion today, versus less than $600 billion in 2007).

CLO 'AAA' tranches often have 35% or more par subordination today, up from an average of 26% before the crisis. While, as mentioned above, issuers of BSL CLOs (which make up at least 85% of the space) are no longer required to retain an interest in the structures, issuers of middle-market CLOs must still conform to the risk-retention rules (although this segment makes up at most 15% of the overall CLO space). The distinction between these types of CLOs depends on the size of the corporate loan issuers (middle market CLOs generally hold loans from smaller companies with EBITDA less than $100 million) and in some cases how the collateral was obtained: in the open market (BSL) or originated by the manager (some middle-market loans). In the case of the latter, the CLO manager is still subject to the risk-retention rule.

The parallel increase in the size of the CLO and leveraged loan markets raises some questions. As demand increases beyond a threshold in any market, one concern is the loosening of certain credit-related features on the supply side. In the case of corporate-leveraged loans, a trend observed over the past five or six years is the disappearance of particular loan covenants. While so-called "incurrence covenants" have remained in place, "cov-lite" loans without maintenance covenants are the norm in the corporate loan market nowadays. Other indicators of potential distortive effects produced by increased leveraged loan demand might be found in EBITDA add-backs, increased leverage, lack of debt cushions, and loan document concerns. This, along with an increase in the low-rated corporate loan issuers ('B-' and below) and the potential impact on our CLO ratings is a topic that S&P Global Ratings has recently commented on. To learn more, see our Nov. 26, 2019, commentary "To 'B-' Or Not To 'B-'? A CLO Scenario Analysis In Three Acts."

Based on our analysis, we believe that corporate loan recovery rates could weaken in the next downturn due to increased corporate leverage along with the proliferation of covenant-lite loans and weaker loan documentation. Historically, recovery rates of first lien leveraged loans have been in the neighborhood of 75%-80%. Our analysis suggests that this might decrease to the mid-60s during the next economic downturn.

We believe these risks are captured by our CLO analysis through related analysis we carry out on the corporate loans and loan issuers found within CLO transactions. S&P Global Ratings rates about 95% of companies that make up the underlying broadly syndicated loan CLOs, including both the issuer credit and recovery rating (note that this excludes middle-market CLOs, which make up less than 15% of annual CLO issuance in the U.S., where the underlying loans are unrated and S&P Global Ratings uses credit estimates to help assess the credit risk of these underlying assets). Our CLO criteria use our credit ratings, credit estimates, and recovery ratings on the underlying assets to inform our default and recovery assumptions employed in our CLO ratings analysis. This is described in our June 26, 2019, commentary, "Credit FAQ: Understanding S&P Global Ratings' Updated CLO And Corporate CDO Criteria".

The ABS markets

The ABS market is made up of numerous asset classes; however, the three largest are auto loan/lease, student loans, and credit cards. All three markets are well-developed, with utilization rates as described in chart 2 above. As with many structured products, the 5% risk-retention rule typically applies to auto loan/lease, student loan, and credit card ABS.

Auto ABS 

Auto ABS issuance peaked in 2005 at $106 billion and dropped precipitously during the credit crisis, falling to $45 billion in 2009. Issuance remained depressed until 2014 when it returned to volumes near $100 billion. In 2019, the market saw $111 billion in auto ABS issuance. As with the housing market, there is a subprime segment, which has been present for years. Today, consistent auto ABS issuance has brought the auto share of securitized loans into the range of 12% to 14%.

Subprime auto ABS losses rose considerably beginning with the 2015 vintage, despite a favorable economy. This was largely due to increased competition in the industry, weaker credit standards, and deep subprime auto loans representing a greater share of securitization volume. However, most issuers are reporting stable to improved loss performance as a result of the tighter credit standards that they implemented in mid-2016 and 2017. Non-investment-grade classes of auto ABS now exceed pre-crisis levels, with demand partly driven by the search for yield. At the same time, these deals are not structured as tightly (in terms of performance triggers) as when there were bond insurers, which structured and insured the senior tranches of these transactions (from the mid-1990s to 2007) and acted as the control parties when servicer transfers were required. As a result, there is now potentially more risk to holders of the non-investment-grade classes compared to pre-crisis.

Subprime auto ABS, which has been increasing and now makes up about a third of total auto loan ABS issuance, presents a degree of risk to speculative grade investors. However, these subordinated tranches make up only a small portion of overall subprime auto loan ABS issuance. In our opinion, the risk of contagion beyond the subprime auto ABS market in the case of an economic market downturn is limited because of the relatively low (compared to housing and leveraged loans) utilization rate of subprime auto loan ABS, the lower size of the total outstanding auto debt versus that of the housing market (at just over $1 trillion, auto debt outstanding is roughly 10% that of housing debt), and the lack of related derivative products.

For more on the risks to speculative-grade noteholders, see the Feb. 20, 2018, commentary "Fewer Structural Protections In Today's Speculative-Grade Subprime Auto Loan ABS."

Credit cards 

The credit card ABS market has shrunk in recent years. Credit card ABS has been affected by the banks' ability to obtain alternative sources of funding with lower cost than securitizations. A combination of a GAAP accounting change, cheap deposit funding, and a contracted credit card lender market, in which mostly big banks securitize, has led to a rapid decline in utilization to the 10%-15% range, where it sits today.

As with the auto market, the total outstanding credit card debt in the U.S. is slightly over $1 trillion. The main changes in credit card ABS are improvements to the collateral quality (e.g., average pool FICO scores are higher, and greater borrower seasoning is typical) and the fact that, relative to pre-crisis levels, there is higher credit enhancement in the bank card deals.

Many bank card issuers have refrained from adding receivables and accounts to their securitization trusts in recent years, as most trusts have significant excess collateral and receivables to back the new issuances. Therefore, account seasoning in the trusts has been improving. Moreover, transaction documents typically constrain the levels of new receivables that can be added to existing master trusts. Added receivables and accounts cannot typically exceed 15% of the total portfolio in any three-month period and 20% in any 12-month period. So while banks' balance sheets may show some volatility in their credit card receivables quality, the collateral quality of their card securitization trusts is generally stable.

Student loans 

Today student loan ABS issuance volume is considerably lower than when Federal Family Education Loan Program (FFELP) student loans were originated prior to 2010. Although FFELP originations ended in 2010, the credit quality for the FFELP-backed ABS remains strong due to the guarantee (of 97% of the principal) on defaulted loans from the U.S. government.

Over the past few years, the ABS new issuance of refinance and in-school private student loans has roughly offset the decline in FFELP loans, which has helped to stabilize the level of student loan ABS issuance. There have been substantive changes to private student loan ABS structures since the crisis. For one thing, the credit enhancement is higher than it was pre-crisis. Also, while sequential structuring was not present in many transactions prior to the crisis, it has been broadly adopted post-crisis and is present in all but a small number of student loan deals.

Student loan debt, now mostly on the government's balance sheet, is not funded via ABS to the extent it was pre-crisis, and its utilization rate has fallen to just above 10% today from near 40% in 2007.

The CMBS market

The commercial mortgage-backed securities (CMBS) market has undergone various changes. There was $96 billion in CMBS issuance in 2019 (excluding agency-issued multifamily transactions), and outstanding CMBS currently sits near $750 billion (including the agency market), roughly on par with what it was pre-crisis. The difference is that there has been a shift away from the private label market to the agency space. Also, balance sheet lenders have taken market share away from CMBS over the past decade. Pre-crisis, conduits (i.e., multi-borrower deals) dominated the market, accounting for over 90% of issuance. In recent vintages, however, single-borrower deals (which have a better performance history than conduits) make up about 40%-50% of issuance.

CMBS issuers need to either retain equity themselves or assign it to a third party. Credit enhancement levels are substantially higher now relative to pre-crisis conduits. For example, relative to pre-crisis CMBS, credit enhancement is roughly twice as high for 'AAA' tranches and two to three times higher on the 'BBB' rated bonds. Underwriting standards have also improved. For example, pro forma underwriting, in which rental rate increases were assumed from the start of the loan, are not present in this cycle. Underwriting is now based on in-place cash flows and leases. Lastly, changes have been made that we believe improve the level of protection for senior noteholders relative to earlier vintage CMBS structures. For example, "appraisal reductions" now exist, which limit the ability of the servicer to advance amounts that may not be ultimately recovered in a property liquidation.

While the overall losses on CMBS were minor during and after the crisis, this is not to say there weren't collateral delinquencies, bond downgrades, and mark-to-market losses incurred by investors who sold into an illiquid market. Overleveraged holders of this debt were exposed to serious financial risk. Also, investors that needed to maintain a portfolio of 'AAA (sf)' rated assets may have been forced to liquidate assets at prices well below the purchase point. However, CMBS prices generally recovered despite substantial spread widening.

Post-crisis, a competitive lending environment spurred by low interest rates contributed to the reduction in securitization as a funding source for commercial real estate. However, chart 2 shows that the utilization rate has since stabilized in the mid-to-high teens percentage area.

S&P Global Ratings Has Strived To Improve Transparency And The Quality Of Its Ratings

Since the crisis, S&P Global Ratings has strived to improve the quality and transparency of its ratings. Rating criteria have been updated for most types of securities affected by the crisis, including RMBS, CMBS, and CLOs. Structured finance criteria changes have generally led to the establishment of higher credit enhancement levels to support higher ratings.

In our June 3, 2009, commentary "Understanding S&P Global Ratings' Rating Definitions," we outlined a set of hypothetical stress scenarios as benchmarks for calibrating our criteria across different sectors in order to maintain comparability of our ratings. Our May 3, 2010, "Credit Stability Criteria" commentary explicitly recognized credit stability as an important factor in our ratings. In that paper we outlined our expectations for the maximum potential deterioration we associate with each rating category under moderate stress conditions.

We believe that the publication of our category-related stress scenarios and credit stability criteria adds helpful transparency to our ratings. Details about updates to our methodology and criteria specific to individual asset classes are outlined in various publications, which are publicly available.