Criteria | Structured Finance | ABS: Revised U.S. ABS ...
Post on 25-Apr-2022
10 Views
Preview:
Transcript
Criteria | Structured Finance | ABS:
Revised U.S. ABS RatingMethodology And Assumptions ForCredit Card SecuritizationsGlobal ABS Criteria Officer:Joseph F Sheridan, New York (1) 212-438-2605; joseph_sheridan@standardandpoors.com
Primary Credit Analysts:Ildiko Szilank, New York (1) 212-438-2614; ildiko_szilank@standardandpoors.comFrank J Trick, New York (1) 212-438-1108; frank_trick@standardandpoors.com
Analytical Manager, ABS Term New Issuance:Felix Herrera, CFA, New York (1) 212-438-2485; felix_herrera@standardandpoors.com
Analytical Manager, ABS Term Surveillance:Eric Hedman, CFA, New York (1) 212-438-2482; eric_hedman@standardandpoors.com
Structured Finance Research:Erkan Erturk, Ph.D., New York (1) 212-438-2450; erkan_erturk@standardandpoors.com
Table Of Contents
SCOPE OF THE CRITERIA
SUMMARY OF CRITERIA UPDATE
IMPACT ON OUTSTANDING RATINGS
EFFECTIVE DATE AND TRANSITION
METHODOLOGY
Unemployment Rates And Credit Card Losses
Benchmarking And Peer Comparison
Credit Card ABS Rating Sensitivity To Economic Downturn And Relative
Stability Of Different Rating Categories
APPENDIX
January 28, 2010
www.standardandpoors.com/ratingsdirect 1
771570 | 300001603
Table Of Contents (cont.)
Regression Analysis
RELATED RESEARCH
Standard & Poor’s | RatingsDirect on the Global Credit Portal | January 28, 2010 2
771570 | 300001603
Criteria | Structured Finance | ABS:
Revised U.S. ABS Rating Methodology AndAssumptions For Credit Card Securitizations(Editor's Note: The criteria in this article supplement and partially supersede the criteria described in the "Base Case
And Stress Assumptions" section for "loss rate" of the criteria article "General Methodology And Assumptions For
Rating U.S. Credit Card Securitizations," published Dec. 15, 2008.)
1. Standard & Poor's Ratings Services is refining and adapting its methodology and assumptions for rating U.S. credit
card asset-backed securities (ABS). We are publishing this article to help market participants better understand our
approach to reviewing securities backed by credit card loans. This article is related to our criteria article
"Principles-Based Rating Methodology For Global Structured Finance Securities," which we published on May 29,
2007.
SCOPE OF THE CRITERIA
2. Standard & Poor's revised its criteria for credit card ABS as part of its overarching objective to make ratings
comparable across sectors, geographic regions, and over time. The changes reaffirm the long-standing definitions of
our ratings. We intend for each rating symbol (e.g., 'AAA') to connote a comparable overall view of
creditworthiness wherever and whenever it appears.
3. To help enhance the comparability of our ratings, we recently embraced the use of specific economic stress scenarios
to calibrate our criteria. For our 'AAA' stress scenario, we use the Great Depression, a time of extreme economic
stress. Accordingly, we intend for 'AAA' rated credit card ABS securities under the revised criteria to be able to
withstand that level of economic stress without defaulting (although, they could suffer downgrades under that level
of stress).
4. The principal updates to our criteria are:
• We will supplement our established criteria methodology by comparing all U.S. credit card master trust portfolios
with our U.S. Bankcard Credit Card Quality Index (CCQI). The CCQI will serve as an industry average
benchmark (the benchmark pool), and we will use it to compare and measure all rated pools when we forecast
portfolio-specific stressed charge-off rates and when we evaluate the adequacy of provided credit support;
• We have established a 'AAA' peak charge-off rate of 33% for the benchmark pool that will be operative,
provided the historical relationship between unemployment rates and industry charge-off rates continues to hold,
when expected charge-off rates for the benchmark pool are within a range of about 5% to 10%. By holding the
'AAA' stressed peak charge-off rate constant during normal economic cycles, we effectively increase the loss
coverage multiple when charge-off rates are lower during more benign economic environments;
• We will evaluate the credit risk of an actual pool relative to that of the benchmark pool as well as its peer group.
During our evaluation, we will review originator-specific data on historical performance, pool-specific loan and
borrower characteristics, and our forward-looking base case performance; and
• After we compare an actual pool with the benchmark pool and peer group, we will adjust the actual pool's
specific 'AAA' stressed charge-off rates to the extent there are deviations in originators' historic and expected
www.standardandpoors.com/ratingsdirect 3
771570 | 300001603
performance relative to the benchmark pool and peer group or current loan and obligor characteristics relative to
the peer group.
SUMMARY OF CRITERIA UPDATE
5. The criteria outlined in this article supplement and partially supersede previously published criteria for credit card
ABS described in the "Base Case And Stress Assumptions" section for "loss rate" of the criteria article "General
Methodology And Assumptions For Rating U.S. Credit Card Securitizations," published Dec. 15, 2008.
IMPACT ON OUTSTANDING RATINGS
6. At this time, the changes outlined in this article will not in and of themselves have an impact on our outstanding
U.S. credit card ABS ratings.
7. It is important to note that while this criteria article is focused on credit card losses (also referred to herein as
charge-offs), other variables may have a more significant impact on a transaction's credit profile, ratings, and
performance. For example, a one percentage point decrease in the base-case scenario payment rate could have a
more significant impact on credit enhancement levels and on ratings than a one percentage point increase in the
base-case charge-off rate.
8. This article is intended to provide increased transparency and establish a clear framework for benchmarking pools
of credit card loans and provide insight into the stability of credit card ABS ratings. We believe that all rated
securities meet or exceed the new criteria outlined in this article, and we expect that the consistent application of
these criteria will enhance ratings quality and consistency.
9. As of December 2009, Standard & Poor's maintained surveillance on 770 rated U.S. credit card ABS securities. In
2009 we raised 183 ratings and lowered 96. However, the number of downgrades would have been greater and the
number of upgrades would have been fewer if certain issuers did not provide additional credit enhancement during
the past year. Four-hundred-and-thirteen rated securities benefited from this additional enhancement. Without the
additional support, 236 of 413 ratings (approximately 57%) would have remained in the same rating category. We
would have likely lowered the remaining 177 ratings (approximately 43%) by up to one rating category. For
additional information on ABS transactions that benefited from additional support, see "ABS Credit Ratings Would
Have Remained Relatively Stable Even Without Additional Support," published Jan. 25, 2010, and "U.S. Credit
Card ABS Issuer Report: Additional Credit Support From Originators Results In 154 Upgrades Of Ratings
Previously On Watch Negative," published Oct. 2, 2009.
10. We expect that the criteria changes outlined in this article will enhance the stability of our 'AAA' ratings during
normal economic cycles.
EFFECTIVE DATE AND TRANSITION
11. These criteria are effective immediately for all new and outstanding U.S. credit card ABS transactions.
Standard & Poor’s | RatingsDirect on the Global Credit Portal | January 28, 2010 4
771570 | 300001603
Criteria | Structured Finance | ABS: Revised U.S. ABS Rating Methodology And Assumptions For Credit CardSecuritizations
METHODOLOGY
12. We will explicitly use our U.S. CCQI as an industry benchmark to compare and measure outstanding pools against,
based on collateral performance. The CCQI is a monthly performance index that aggregates performance
information across Standard & Poor's rated bankcard transactions in several key risk areas. We are also
emphasizing the use of peer group comparisons in refining our evaluation of a specific pool relative to other similar
portfolios based on collateral characteristics. The 'AAA' peak charge-off rate for the benchmark pool is 33%. This
benchmark 'AAA' peak charge-off rate reflects our opinion of expected performance under conditions of extreme
economic stress.
13. In addition to stresses on performance variables such as payment rate, purchase rate, yield, and cost of funds, the
credit enhancement supporting a 'AAA' rated security should also be sufficient to withstand a pool-specific 'AAA'
stressed charge-off rate that peaks 12 months following the breach of an early amortization event trigger. We
consider the 'AAA' peak charge-off rate of 33% associated with the benchmark pool as a fixed anchor point.
However, pool-specific variations in historic actual and projected base-case performance relative to the benchmark
pool, as well as loan or obligor characteristics relative to an issuer's peer group, may cause the 'AAA' stressed
charge-off levels for actual pools to vary (higher or lower) from the benchmark pool's 33% peak charge-off level. In
addition, we review pools on an ongoing basis and portfolio-specific base-case and stressed peak loss projections
may change over time. For example, we may adjust our projections if performance deviates from industry trends, if
performance is not in line with what we would expect based on current economic conditions, or if there has been a
material change in underwriting standards, geographic diversification, or product mix.
14. We believe that issuer-specific historical pool performance for a comparable managed portfolio is a useful indicator
of future performance. Therefore, issuer-specific portfolio performance will be a significantly weighted variable in
the comparative analysis. We will consider the managed portfolio's historical performance relative to the benchmark
pool as well as our expectations for future performance in the base-case scenario going forward.
15. When rating or surveilling credit card ABS transactions, we use fundamental credit analysis, which incorporates
various qualitative and quantitative elements. Our ratings arise in large part from certain base-case assumptions for
key variables, including losses, yield, payment rates, purchase rates, and coupon rate. The base-case scenario reflects
our view of how a pool is likely to perform and reflects historical performance during various economic
environments. The base-case scenario is also based on our opinion of the card originator's/servicer's overall
experience, the quality and consistency of its origination activities and underwriting, and its account management,
collection, and servicing practices.
16. We also look at factors such as the detail and depth of data on the collateral, the amount of performance history for
the card originator/servicer through varying economic and business cycles, and examples of operational strategies
the issuer has used to mitigate credit risk. For example, we may use more stressful assumptions for pools with less
than three years of data available, or for pools from originators/servicers with, in our opinion, less experience
managing and servicing credit card asset pools. Our assumptions for such pools generally will rely more heavily on
comparisons with portfolios with similar loan characteristics, vintage data (for example, information on one year of
originations), and worst-case scenario figures.
17. After we determine our base-case assumptions for each variable, we use them as starting points to simulate a
deteriorating credit environment by stressing all variables simultaneously and applying the stresses to the base-case
www.standardandpoors.com/ratingsdirect 5
771570 | 300001603
Criteria | Structured Finance | ABS: Revised U.S. ABS Rating Methodology And Assumptions For Credit CardSecuritizations
assumptions progressively until we reach the maximum stress level (for charge-offs, the peak charge-off). We then
maintain the maximum stress level until the notes' legal final maturity date. The stresses vary for each rating
category.
18. Our base-case and stress assumptions for pool-specific charge-off rates include factors that will affect losses in a
stress scenario. Some of the considerations that affect our base-case loss assumptions include vintage loss
performance, delinquency roll rates, lagged loss rates based on receivables growth, account origination strategies,
account seasoning, FICO credit tier distribution of new accounts and migration of FICO scores for existing
accounts, geographic concentrations, and credit line management strategies.
19. When we compare a pool's historic actual and future expected loss rates with the benchmark pool's historical
performance, we also make adjustments to the actual pool performance based on the quality and consistency of
underwriting over time. For example, if FICO scores are drifting downward, it may be necessary to adjust pool
specific charge-off rates upward. Likewise, for each account and obligor attribute, we will make qualitative
adjustments for variations from the benchmark pool.
20. If a pool's expected charge-offs are higher than what we would expect for the benchmark pool in the base-case
scenario, then the pool-specific 'AAA' stressed peak charge-off will likely be above 33%. 'AAA' stressed peak
charge-off rates that are lower than 33% could also be applied in our cash flow analysis to the extent there are pools
with higher credit quality assessments relative to the benchmark pool.
21. Although the stressed peak loss rate plays a central role in our rating analysis, it's not the only factor we consider.
Our assessment of the interplay of many factors will affect our rating decisions. For example, higher payment rates
by the obligors or higher yield from the underlying assets could offset an increase in losses when we're determining a
new rating or considering a rating action during the surveillance of an existing rating.
22. For greater detail on our established criteria methodology and assumptions for rating U.S. credit card ABS, please
refer to "General Methodology And Assumptions For Rating U.S. Credit Card Securitizations," published Dec. 15,
2008.
Supporting research and rationale for the 'AAA' peak stressed charge-off assumption for the benchmark
pool23. General-purpose bank credit cards were not around in the U.S. during the Great Depression. Bank of America
launched its BankAmericard credit card program—the predecessor of Visa—in 1958. A group of banks joined
together in 1966 and created the InterBank Card Association—now known as MasterCard. In the absence of
historic credit card performance data during periods of extreme economic stresses, we have formed an opinion of
how credit card losses should behave in a 'AAA' stress scenario based on an examination of loss performance during
recent economic downturns and an analysis of the relationship between charge-off rates and key economic
indicators.
24. The rationale supporting our peak charge-off rate assumption for 'AAA' credit card ABS ratings draws from recent
observed data and an analysis of the relationship between credit card charge-off rates and key economic indicators
such as the unemployment rate. During times of recession, some studies have shown that there is a one-to-one
relationship between the unemployment rate and credit card losses. Standard & Poor's published a research study
investigating this relationship as well (see "The Recession—Combined With Severe Home Price Drops—Has
Increased The Unemployment Rate’s Correlation With U.S. Credit Card Losses," published Aug. 31, 2009). Based
Standard & Poor’s | RatingsDirect on the Global Credit Portal | January 28, 2010 6
771570 | 300001603
Criteria | Structured Finance | ABS: Revised U.S. ABS Rating Methodology And Assumptions For Credit CardSecuritizations
on the findings of our regression analysis, we expect that the annual credit card loss rate could reach at least 27% in
an extreme scenario where the unemployment rate reaches 25% (as we saw during the Great Depression).
25. Based on our analysis, we expect that the correlation between the unemployment rate and credit card losses will
remain strong in periods of double-digit unemployment. The regression coefficient may increase beyond the current
level, however, implying that credit card loss rates could rise at a faster pace than unemployment. In other words,
the historical linear regression relationship could become an exponential one in double-digit unemployment rate
environments. Overall, this finding provides the basis for our assumption that, in a 'AAA' scenario, we would see a
rapid increase in charge-off rates that would correspond with rising unemployment rates. This expectation is in line
with our revised 'AAA' stressed scenario industry benchmark of about 33% peak charge-off rates, which reflects our
opinion of expected performance under conditions of extreme economic stress.
For a more detailed review of our regression analysis, please refer to the Appendix at the end of this
article.
Unemployment Rates And Credit Card Losses
26. Generally speaking, the unemployment rate mirrors the economic environment. During the Great Depression in the
1930s, the unemployment rate reached 25%. In December 1982, the unemployment rate was 10.8%, the highest
we've observed since the monthly unemployment data became available in 1948. Since then, the unemployment rate
has hovered around 6%, and it remained under 8% through 2008 (see chart 1). It then passed 8% in February
2009. By June 2009, the unemployment rate had climbed to 9.5% and then surpassed 10% in October 2009.
Unemployment was at 10% at the end of December 2009.
Chart 1
www.standardandpoors.com/ratingsdirect 7
771570 | 300001603
Criteria | Structured Finance | ABS: Revised U.S. ABS Rating Methodology And Assumptions For Credit CardSecuritizations
27. The reason why there is a significant relationship between the unemployment rate and credit card losses is simple:
Unemployment tends to increase during economic downturns and the loss of income leads to borrowers' inability to
repay credit card loans (see chart 2). Unemployed consumers are also more likely to rely on their credit cards, and
they may use their credit cards to compensate for their lost income, but only for a while. Many credit card
borrowers who fall behind on their monthly bills and are unable to find jobs will eventually face credit card
delinquencies and charge-offs.
28. In addition, other borrowers who are employed and have relatively strong credit profiles but are worried about job
security during periods of economic stress may reduce their credit card purchases to increase their savings. This can
lower the revolving debt and consequently increase charge-off rates. This is why market researchers use the
unemployment rate as a key determinant of delinquencies and bankruptcies during economic crises. As observed in
the current credit environment, this relationship becomes stronger in a stressed economy because traditional
non-employment-related sources of credit card loan repayments decrease. That is, many U.S. households may
experience a significant decline in their worth held in the stock market and a more limited ability to extract depleted
equity from their homes. As a result, those households may struggle to meet their credit card obligations, which, in
turn will generally lead to higher credit card losses.
Chart 2
Standard & Poor’s | RatingsDirect on the Global Credit Portal | January 28, 2010 8
771570 | 300001603
Criteria | Structured Finance | ABS: Revised U.S. ABS Rating Methodology And Assumptions For Credit CardSecuritizations
Benchmarking And Peer Comparison
29. To help maintain ratings comparability across the U.S. credit card ABS sector, we will compare the credit risk of a
master trust portfolio with that of the industry average CCQI benchmark pool as well as other master trust
portfolios that we consider to be in the same peer group. The factors that we will review include originator-specific
data on risk management and operational strategies, historical performance, pool-specific loan and borrower
characteristics, and our forward-looking base case expected performance. The peer group analysis affords analysts
the opportunity to make comparisons across issuers and can be useful in identifying trends and market
developments that may be less apparent when looking exclusively at a single portfolio or originator.
30. We use our CCQI as a proxy for the average U.S. credit card portfolio. As of October 2009, the U.S. CCQI tracked
approximately $401.6 billion bankcard receivables backing Standard & Poor's rated credit ABS. In aggregate, the
U.S. bankcard CCQI tracks the performance of receivables which represent approximately 45% of the total U.S.
revolving consumer debt balance. The benchmark pool is an aggregate of all bankcard pools tracked by the CCQI.
31. The CCQI aggregates performance information across Standard & Poor's rated bankcard transactions in the
following key risk areas: receivables outstanding, yield, payment rate, charge-off rate, delinquencies, base rate, and
excess spread rate. Each master trust's weighting is determined by the size of its outstanding eligible principal
receivables divided by the total outstanding eligible principal receivables for all trusts included in the index. The
resultant CCQI weighted average performance variables are then determined.
32. We also publish comparative portfolio-specific data in our U.S. Credit Card ABS Issuer Report (Issuer Report) to
complement our CCQI. The Issuer Report highlights issuer-specific performance and provides the collateral data for
all credit card receivables pools backing outstanding public/144A credit card ABS transactions that we rate. The
report includes comparative data on pool performance as well as collateral characteristics such as average credit
limit, average account balance, credit line utilization rate, geographic concentrations, and age of accounts. Similar
statistics are used to compare pools within an issuer's peer group when we review a master trust as part of the rating
process.
Credit Card ABS Rating Sensitivity To Economic Downturn And Relative StabilityOf Different Rating Categories
33. Standard & Poor's ratings generally reflect our opinion of the level of stress that a bond with a particular rating
should be able to withstand without defaulting. The higher a bond's rating, the stronger, in our opinion, is the
bond's capacity to meet its financial obligations during adverse economic scenarios.
34. The more junior the notes are in the capital structure, the more likely it is that their stress scenario is closer to that
of the existing base-case scenario. In our benchmark pool, we define the 'B' stress scenario as one where the 'B' peak
charge-off rate is also the expected loss given the existing base case economic scenario. The 'B' stress scenario is
consistent with the base-case scenario and will change to reflect changing economic conditions or outlook. As such,
the junior notes' credit enhancement and ratings will be more sensitive to changes in our base-case expectations than
the more senior notes in the capital structure. For example, 'B' rated securities during a booming economy featuring
rapid employment and income growth would be at greater risk of downgrade or default than higher rated securities
as the economy weakened going forward.
www.standardandpoors.com/ratingsdirect 9
771570 | 300001603
Criteria | Structured Finance | ABS: Revised U.S. ABS Rating Methodology And Assumptions For Credit CardSecuritizations
35. Since the stress scenarios for the more senior tranches are further removed from the existing base-case scenario
(more stressful), we would expect that 'AAA' rated senior tranches would have relatively stable enhancement levels,
and ratings, until the downturn hits a point that is beyond a level considered consistent with a normal economic
cycle.
36. Although we assign our initial ratings based on our outlook at that time, reality may play out differently. As our loss
expectation for pools backing an outstanding credit card ABS changes because of changes in the market and
economy, we review the ratings.
37. We would consider the relationship between projected losses and currently available credit enhancement before
taking any rating action. Other factors that would likely influence a potential rating action include the seasoning of
the portfolio, the structural features of the transaction, and the expected number of months before maturity. We
typically would also consider the performance of the other primary performance variables including portfolio yield,
payment rate, and purchase rate. Ratings can be adversely affected by an adverse change in any of theses variables,
and the effects of deterioration in one variable could be offset by stability or improvement in another. Moreover, a
rating may be more sensitive to a change in a performance projection that is not related to charge-offs. For example,
a one-percentage point decrease in the base-case payment rate could have a more significant impact on credit
enhancement levels and on the ratings than a one percentage point increase in the base-case charge-off rate.
38. Provided the historical relationship between the unemployment rate and industry charge-off rates continues to hold,
the 'AAA' stressed peak charge-off rate for the benchmark pool of 33% should remain operative in periods of mild
to moderate economic stress when unemployment rates are below 10%. When charge-offs are stressed to 33% and
the expected industry average charge-off rate ranges from 5% to 10%, this corresponds with a loss coverage range
of about 6.6x-3.3x the base case. In the past, the stress multiples we observed were typically in the 3x-5x range for
'AAA' rated credit card ABS. The increase at the upper end of the loss coverage band reflects the "floor" concept
that we are instituting by holding the 'AAA' stressed peak charge-off rate at 33% even if expected charge-off rates of
5% are being realized in a mild economic environment. Although we expect that our 'AAA' stressed peak charge-off
rate assumption for the benchmark pool to remain constant at 33% during normal economic cycles, we recognize
that if economic and market conditions migrate significantly beyond the normal ranges for cyclical fluctuations,
even the 'AAA' stressed peak charge-off rate would have to be higher. We expect that the peak charge-off rate
assumption for the benchmark pool would range from about 3x to 3.5x the base-case scenario in an economic
environment where the unemployment and base-case charge-off rates are between 10% and 20%. For the
benchmark pool, the 'AA', 'A','BBB', and 'BB' peak loss scenarios are increasingly less stressful relative to 'AAA'
stress case and are expected to be in the range of about 2.5x-5.0x, 2.0x-3.75x, 1.5x-2.5x, and 1.25x-1.5x the
base-case when the base-case charge-off rates are in the 5%-20% range. The base-case and stressed peak charge-off
rates will vary (higher or lower) for actual securitization trusts based on a portfolio- and servicer-specific analysis
and stressed charge-off rates as a multiple of the base case will not be the same for all pools. For example, subprime
pools with charge-offs that are materially higher than the industry average are likely to have higher stressed peak
charge-off rates that are a lower range of multiples of the base case for that pool. In the 'B' rating scenario, peak
losses are unstressed or equal to 1x the base case. Rating stresses and peak charge-off assumptions for actual pools
will vary based on differences in pool characteristics and historic performance relative to the benchmark pool.
39. As our outlook for the economy, the credit card industry, and specific pools change, we reassess the potential impact
of these changes on our outstanding ratings. This impact will be greatest for credit card ABS rated 'B' and 'BB' and
least for securities rated 'AA' and 'AAA'.
Standard & Poor’s | RatingsDirect on the Global Credit Portal | January 28, 2010 10
771570 | 300001603
Criteria | Structured Finance | ABS: Revised U.S. ABS Rating Methodology And Assumptions For Credit CardSecuritizations
40. Chart 3 illustrates the expected stressed charge-off rates at various rating categories based on the base-case
charge-off rates (5% to 15%) in various economic scenarios for the benchmark pool. In the 'B' rating scenario, the
peak charge-off rate would equal the base-case forecast. For the 'AAA' rating scenario, provided the relationship
between charge-off rates and economic variables such as unemployment remains in line with historical patterns, the
'AAA' peak would be expected to remain constant at 33% for the benchmark pool until base-case charge-off
percentage forecast moves into the double digits. 'AAA', 'A', and 'BBB' are the three most common ratings in the
credit card ABS market. Chart 3 illustrates that the peak loss rate assumed when assigning a rating in the 'B'
category moves with the base case forecast under all economic conditions and as ratings move up the scale, the
stressed peak loss rate assumed is less vulnerable to change as the economy moves away from benign conditions to
more stressed conditions. As a result, 'AAA' ratings are generally expected to be relatively stable during normal
economic cycles and lower rated securities are generally expected to be more pro-cyclical as stressed charge-off rates
are expected to change at earlier stages of economic declines. Further, the point in a normal economic down cycle
with rising unemployment rates where a rating is vulnerable to change is expected to occur sooner as the ratings
move down the scale. In other words, as the economy moves from benign conditions with low unemployment rates
to more stressful conditions, the 'BBB' stressed loss assumptions are likely to move higher before the 'A' stress case
assumptions, and they are both expected to change before the 'AAA' stress case would need to be adjusted.
Although base-case and stressed charge-off rates will vary for actual securitization trusts based on a portfolio- and
servicer-specific analysis and are subject to change as discussed herein, this same basic rating scenario framework
will apply in evaluating the adequacy of provided credit enhancement.
41. We believe that benchmarking and the use of specific economic scenarios to calibrate our ratings criteria will
enhance ratings comparability. However, the changes to our criteria outlined herein do not in any way diminish the
www.standardandpoors.com/ratingsdirect 11
771570 | 300001603
Criteria | Structured Finance | ABS: Revised U.S. ABS Rating Methodology And Assumptions For Credit CardSecuritizations
importance of issuer- and portfolio-specific analysis. Credit card ABS master trust pools are actively managed and
some card issuers will employ account origination and management strategies that may be more successful at
mitigating risk in challenging economic environments than others. In addition, other portfolio-specific credit
characteristics, including performance variables such as portfolio yield, payment rate, and purchase rate, are
important considerations in our analysis.
APPENDIX
Regression Analysis
42. For our analysis, we used information from our monthly CCQI (1992 to 2009) and other economic data, such as
the unemployment rate, the S&P Case-Shiller Home Price Index, the S&P 500 equity index, personal bankruptcy
filings, personal income, the Federal Reserve's outstanding consumer credit amount, and GDP growth rate from
Bloomberg.
43. Based on the pair-wise correlations, which are consistent with our expectations, we observed that the changes in
credit card losses are positively correlated with changes in the unemployment rate (0.60) and in personal bankruptcy
filings (0.31), while credit card losses are negatively correlated with changes in home prices (-0.53), personal income
(-0.44), the GDP growth rate (-0.29), and the S&P 500 (-0.23) (see appendix table 1). The unemployment rate is the
most significant economic variable that affects credit card loss rates, followed by the home price index. Similarly, the
changes in unemployment rates are negatively correlated with the changes in personal income, the equity market,
home prices, and the GDP growth rate. When two or more independent variables are correlated, they basically
measure the same relationship. Given the high correlation among independent variables, we believe it's reasonable to
assume that unemployment conveys essentially the same information for personal income, the equity market, GDP,
and to a certain extent, home price behaviors in our regression analysis. As a result, we could consider eliminating
these variables from our key factors list and regression equations when necessary.
Appendix Table 1
Correlation Coefficients
12-monthchange in
loss rate
12-month changein unemployment
rate
12-monthchange in
personalbankruptcy
filings
12-monthchange in
personalincome
12-monthchange inconsumer
creditoutstanding
12-monthchange in
S&P 500
12-monthchange in
S&PCase-Shiller
GDPgrowth
rate
12-Month changein loss rate
1.00 0.60 0.31 (0.44) (0.01) (0.23) (0.53) (0.29)
12-month changein unemploymentrate
0.60 1.00 0.12 (0.80) (0.30) (0.72) (0.43) (0.57)
12-month changein personalbankruptcy filings
0.31 0.12 1.00 (0.11) 0.03 (0.05) (0.25) (0.13)
12-month changein personalincome
(0.44) (0.80) (0.11) 1.00 0.31 0.68 0.38 0.44
12-month changein consumercreditoutstanding
(0.01) (0.30) 0.03 0.31 1.00 0.22 0.12 0.22
Standard & Poor’s | RatingsDirect on the Global Credit Portal | January 28, 2010 12
771570 | 300001603
Criteria | Structured Finance | ABS: Revised U.S. ABS Rating Methodology And Assumptions For Credit CardSecuritizations
Appendix Table 1
Correlation Coefficients (cont.)
12-month changein S&P 500
(0.23) (0.72) (0.05) 0.68 0.22 1.00 0.23 0.64
12-month changein S&PCase-Shiller
(0.53) (0.43) (0.25) 0.38 0.12 0.23 1.00 0.36
GDP growth rate (0.29) (0.57) (0.13) 0.44 0.22 0.64 0.36 1.00
44. While the actual levels of credit card losses and unemployment rates are somewhat correlated, the relation is more
meaningful and stronger when we use the rolling 12-month percent changes to estimate the functional relationship
between them (see appendix chart 1).
45. Beginning with a regression model, we considered most relevant variables regardless of any multi-collinearity
problem in the regression model. Multi-collinearity results from high correlations among independent variables,
such as unemployment and personal income or unemployment and home price changes, which makes it difficult to
determine reliable estimates or regression coefficients.
46. The regression equation 1 in appendix table 2 shows that percent changes in unemployment, bankruptcy filings,
outstanding consumer credit, and home prices are the key loss determinants with acceptable t-statistics. If we
eliminate the home price variable—a potential source of multi-collinearity—from the equation (see regression
equation 2 in appendix table 2), the model's ability (r-square) to explain the behavior of loss rates is around 38%,
down from 62% in regression equation 1. This highlights how important the percent changes in home prices are in
explaining the volatility of credit card losses. We then eliminated all independent variables and kept only the percent
www.standardandpoors.com/ratingsdirect 13
771570 | 300001603
Criteria | Structured Finance | ABS: Revised U.S. ABS Rating Methodology And Assumptions For Credit CardSecuritizations
changes in unemployment rates to explain the credit card loss behavior (see appendix chart 1). In appendix table 2,
regression equation 3, with an r-square of about 36% and significant t-statistics for the unemployment rate, suggests
that the contribution of other variables (other than home prices) to explain credit card loss behavior is very limited
once the unemployment rate is present in the model. That's why researchers typically use unemployment as the sole
variable to predict credit card losses. While the multi-collinearity between unemployment and home prices may, to a
certain extent, be an issue in the regression analysis, the correlation and regression analysis highlight how important
home prices and the unemployment rate are in explaining credit card loss behavior.
Appendix Table 2
Regression Results
Regression equation 1
R square: 61.7% Coefficients Standard error t-stat P-value (%)
Intercept (0.06) 0.03 (1.74) 8.45
12-month change in unemployment rate 0.33 0.08 4.13 0.01
12-month change in personal bankruptcy filings 0.06 0.02 2.45 1.56
12-month change in consumer credit outstanding 3.03 0.43 7.02 0.00
12-month change in S&P Case-Shiller (1.36) 0.14 (9.83) 0.00
Regression equation 2
R square: 37.7% Coefficients Standard error t-stat P-value (%)
Intercept (0.11) 0.04 (2.63) 0.95
12-month change in unemployment rate 0.70 0.09 7.76 0.00
12-month change in personal bankruptcy filings 0.12 0.03 3.89 0.02
12-month change in consumer credit outstanding 2.22 0.54 4.11 0.01
Regression equation 3
R square: 35.6% Coefficients Standard error t-stat P-value (%)
Intercept 0.04 0.01 3.25 0.14
12-month change in unemployment rate 0.78 0.08 10.41 0.00
Regression equation 4 - reduced data
R square: 70.00% Coefficients Standard error t-stat P-value (%)
Intercept (0.03) 0.01 (2.57) 1.13
12-month change in unemployment rate 0.99 0.05 18.26 0.00
47. Finally, after attempting to control the effects of the expansion period in the early 1990s (due to rapid rise of
consumer credit) and the bankruptcy period in 2005 (following a change in bankruptcy law) in the regression
models, we removed these two time periods from the analysis to understand the undistorted functional relationship
between the unemployment rate and credit card losses (see regression equation 4 in appendix table 2 and appendix
chart 2).
48. During the expansion and bankruptcy periods, unemployment wasn't the obvious driver of credit card losses. In
fact, the rolling correlation coefficients between the unemployment rate and credit card losses were negative during
those times due to bankruptcy code changes and the expansion of the credit card industry, which eventually led to
higher losses in 1995-1997. We eliminated about 27% of the monthly data points where the monthly
unemployment rate and credit card loss relationship was negative for regression equation 4 (see appendix table 2
Standard & Poor’s | RatingsDirect on the Global Credit Portal | January 28, 2010 14
771570 | 300001603
Criteria | Structured Finance | ABS: Revised U.S. ABS Rating Methodology And Assumptions For Credit CardSecuritizations
and appendix chart 2). As we expected, r-square reached the 70% level with significant t-statistics for the
unemployment rate variable: The coefficient for the unemployment rate is almost one (it is 0.99), which means that
a 100% increase in the unemployment rate roughly results in a 100% increase in credit card losses during times of
high correlation and economic downturns. We also used the regression equation 4 in appendix table 2 to calculate
the tightly correlated relationship between actual and predicted credit card losses.
RELATED RESEARCH
• "Criteria | Structured Finance | ABS: General Methodology And Assumptions For Rating U.S. Credit Card
Securitizations," published Dec. 15, 2008.
• "The Recession—Combined With Severe Home Price Drops—Has Increased The Unemployment Rate’s
Correlation With U.S. Credit Card Losses," published Aug. 31, 2009.
• "U.S. and Canada Credit Card Quality Index Report: Losses Climb 60 Basis Points After Two Months Of
Improvement," published Jan. 14, 2010.
• "U.S. Credit Card ABS Issuer Report: Additional Credit Support From Originators Results In 154 Upgrades Of
Ratings Previously On Watch Negative," published Oct. 2, 2009.
• "ABS Credit Ratings Would Have Remained Relatively Stable Even Without Additional Support," published Jan.
25, 2010.
• "Criteria | Structured Finance | ABS: How Credit Card Bank Downgrades Could Affect Ratings On U.S. Credit
Card ABS," published Oct. 7, 2008.
• "Criteria | Structured Finance | ABS: Methodology And Assumptions For Rating Short-Term U.S. Money
Market-Eligible Notes Backed By Credit Card Receivables," published July 2, 2009.
www.standardandpoors.com/ratingsdirect 15
771570 | 300001603
Criteria | Structured Finance | ABS: Revised U.S. ABS Rating Methodology And Assumptions For Credit CardSecuritizations
• "U.S. Economic Forecast: To A Prosperous New Year," published Jan. 11, 2010.
• "2010 Forecast: The Recovery Is Slowly Picking Up Steam, But Some Obstacles Could Still Derail It," published
Dec. 23, 2009.
These criteria represent the specific application of fundamental principles that define credit risk and ratings
opinions. Their use is determined by issuer- or issue-specific attributes as well as Standard & Poor's Ratings
Services' assessment of the credit and, if applicable, structural risks for a given issuer or issue rating. Methodology
and assumptions may change from time to time as a result of market and economic conditions, issuer- or
issue-specific factors, or new empirical evidence that would affect our credit judgment.
Standard & Poor’s | RatingsDirect on the Global Credit Portal | January 28, 2010 16
771570 | 300001603
Criteria | Structured Finance | ABS: Revised U.S. ABS Rating Methodology And Assumptions For Credit CardSecuritizations
S&P may receive compensation for its ratings and certain credit-related analyses, normally from issuers or underwriters of securities or from obligors. S&P reserves the rightto disseminate its opinions and analyses. S&P's public ratings and analyses are made available on its Web sites, www.standardandpoors.com (free of charge), andwww.ratingsdirect.com and www.globalcreditportal.com (subscription), and may be distributed through other means, including via S&P publications and third-partyredistributors. Additional information about our ratings fees is available at www.standardandpoors.com/usratingsfees.
S&P keeps certain activities of its business units separate from each other in order to preserve the independence and objectivity of their respective activities. As a result,certain business units of S&P may have information that is not available to other S&P business units. S&P has established policies and procedures to maintain theconfidentiality of certain non-public information received in connection with each analytical process.
Credit-related analyses, including ratings, and statements in the Content are statements of opinion as of the date they are expressed and not statements of fact orrecommendations to purchase, hold, or sell any securities or to make any investment decisions. S&P assumes no obligation to update the Content following publication in anyform or format. The Content should not be relied on and is not a substitute for the skill, judgment and experience of the user, its management, employees, advisors and/orclients when making investment and other business decisions. S&P's opinions and analyses do not address the suitability of any security. S&P does not act as a fiduciary oran investment advisor. While S&P has obtained information from sources it believes to be reliable, S&P does not perform an audit and undertakes no duty of due diligence orindependent verification of any information it receives.
No content (including ratings, credit-related analyses and data, model, software or other application or output therefrom) or any part thereof (Content) may be modified,reverse engineered, reproduced or distributed in any form by any means, or stored in a database or retrieval system, without the prior written permission of S&P. The Contentshall not be used for any unlawful or unauthorized purposes. S&P, its affiliates, and any third-party providers, as well as their directors, officers, shareholders, employees oragents (collectively S&P Parties) do not guarantee the accuracy, completeness, timeliness or availability of the Content. S&P Parties are not responsible for any errors oromissions, regardless of the cause, for the results obtained from the use of the Content, or for the security or maintenance of any data input by the user. The Content isprovided on an "as is" basis. S&P PARTIES DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, ANY WARRANTIES OFMERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE OR USE, FREEDOM FROM BUGS, SOFTWARE ERRORS OR DEFECTS, THAT THE CONTENT'S FUNCTIONINGWILL BE UNINTERRUPTED OR THAT THE CONTENT WILL OPERATE WITH ANY SOFTWARE OR HARDWARE CONFIGURATION. In no event shall S&P Parties be liable to anyparty for any direct, indirect, incidental, exemplary, compensatory, punitive, special or consequential damages, costs, expenses, legal fees, or losses (including, withoutlimitation, lost income or lost profits and opportunity costs) in connection with any use of the Content even if advised of the possibility of such damages.
Copyright ( c ) 2010 by Standard & Poor’s Financial Services LLC (S&P), a subsidiary of The McGraw-Hill Companies, Inc. All rights reserved.
www.standardandpoors.com/ratingsdirect 17
771570 | 300001603
top related