Criteria | Structured Finance | ABS: Revised U.S. ABS

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; [email protected]

Primary Credit Analysts:Ildiko Szilank, New York (1) 212-438-2614; [email protected] J Trick, New York (1) 212-438-1108; [email protected]

Analytical Manager, ABS Term New Issuance:Felix Herrera, CFA, New York (1) 212-438-2485; [email protected]

Analytical Manager, ABS Term Surveillance:Eric Hedman, CFA, New York (1) 212-438-2482; [email protected]

Structured Finance Research:Erkan Erturk, Ph.D., New York (1) 212-438-2450; [email protected]

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


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.


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


771570 | 300001603


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


771570 | 300001603


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


771570 | 300001603


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


771570 | 300001603


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.


771570 | 300001603


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'.


771570 | 300001603


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


771570 | 300001603


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


771570 | 300001603


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


771570 | 300001603


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



Intercept (0.11) 0.04 (2.63) 0.95


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



Intercept 0.04 0.01 3.25 0.14


Regression equation 4 - reduced data


Intercept (0.03) 0.01 (2.57) 1.13


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


771570 | 300001603


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.


771570 | 300001603


• "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.


771570 | 300001603


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.


771570 | 300001603

Documents

Criteria | Structured Finance | ABS: Revised U.S. ABS