Analytic measures of credit capacity can help bankcard lenders build strategies that go beyond compliance to deliver business advantage

How Much Credit Is Too Much? Analytic measures of credit capacity can help bankcard lenders build strategies that go beyond compliance to deliver business advantage Number 35 April 2010 On a portfolio of 1 million bankcards, integrating FICO Credit Capacity Index into a lender s initial line strategy would boost profits by $1.6 million. The landscape for bankcard lenders has changed significantly. An unstable economy, industryhigh delinquency rates and increased government intervention have had serious impact to both strategies and profitability. In response, many lenders are taking steps to retrench their business models, searching for ways to comply with new regulations, while mapping a return to profitability and reducing risk. A critical focus for both compliance and growth is the ability to accurately ascertain a customer s ability to safely manage additional credit. It impacts: Customer Acquisition. Identifying higher-capacity segments allows lenders to build portfolios responsibly by targeting low-risk customers with higher growth potential. Matching Right Product to Right Customer. Given regulatory constraints on interest changes within the first year of service, lenders are keen to construct the right product offerings from the onset. Credit Lines. Understanding both risk and credit capacity are essential for making the most favorable limit decisions. Cross-Selling. To gain a greater share of wallet, lenders need to understand a consumer s capacity to take on new debt, not just repayment risk on current obligations. Regulatory Compliance. Regulation Z, recently amended to implement provisions of the Credit CARD Act, places a burden on issuers to determine the consumer s ability to pay before extending credit. Today s constrained environment presents a dichotomy for bankcard lenders, who must be riskconscious and tighten credit standards, while still pursuing new opportunities. The good news is that analytic measures of credit capacity can help bankcard lenders walk the fine line of growing the business without over-extending consumers. These analytics measure a consumer s capacity to take on additional credit, and as such, provide powerful new insight into managing consumer debt. www.fico.com Make every decision count TM

Traditional Measures of Ability to Pay Are Limited The CARD Act mandates that card issuers consider either the consumer s income or assets as well as their current obligations, and requires that process to be reasonable. However, traditional options for accomplishing this task have key limitations. Income validation: Income, while somewhat predictive, is often self-reported, difficult to verify and subject to manipulation. Consumers who attempt to report income honestly and not all do often are unclear how to calculate their earnings accurately: Do I report my income as gross or net? Do I include stock income/bonuses? Do I account for spousal support? Even when income is accurate, it does not account for discretionary income relative to cost of living. Consider for example, the same income won t stretch as far if you live in Manhattan vs. Little Rock. Additionally, income validation is costly to gather and maintain, putting further burden on operational resources already under stress with compliance and customer management related to high delinquencies. Income estimators: Typically based on a combination of reported income (where available) plus geographic and demographic inferences, income estimators attempt to roughly approximate a borrower s income. Depending on the quality of underlying data and breadth of a truly representative development sample, income estimators can provide practical guidelines. But they don t provide a completely accurate portrayal of an individual s ability to repay new debt. Validating the precision of these estimations can be costly and burdensome. Measures of debt ratio: These ratios, obtained either from the consumer or the credit bureau, may not be sufficient on their own to use as a measure of ability to pay since they don t consider the risk impact of the additional debt of the new card or line. When used, many lenders rely on such metrics assessed during origination which over time, fail to reflect a customer s current status. Lenders need a more precise measure, frequently updated, that considers how the new line will impact a consumer s ability to pay as agreed. While these measures meet immediate compliance obligations, they provide little or no opportunity to create better decisions and competitive differentiation. Stronger, more empirical and more automated measures of capacity are needed for a more complete picture of consumer credit risk.»» Putting Credit Capacity to Work While income and income estimators are required for compliance, the FICO Score is a strong risk predictor, and thus an essential tool within risk management strategies. But bureau-based risk scores predict credit risk given current obligations not how the risk would change given a change in a consumer s available credit. There are many reasons why different consumers are assigned similar FICO Scores. For example, some with mild delinquency and low utilization may receive the same score as those with no delinquency but high utilization. These various credit profiles within the same score band represent different sensitivities to incremental debt, as shown in Figure 1. The theoretical consumers on this chart have similar risk scores and the same default probability, but different credit capacities. www.fico.com page 2

»» Figure 1: Same Risk, Different Capacity (Theoretical Example) Larry, Mary and Harry Similar FICO Score, Different Capacity ESTIMATED DEFAULT RATE 50% 40% 30% 20% 10% Legend Capacity Level Larry Low Mary Medium Harry High 0 0 250 500 750 1000 1250 1500 1750 2000 BALANCE CHANGE What s missing in today s bankcard lending strategies is the ability to determine, For consumers who look equally risky, which can more safely manage additional credit? in other words, more precise measures of credit capacity. Capacity measures would allow bankcard lenders to reallocate loss exposure and reserves toward consumers best able to repay debt, not to mention protect its customers long-term credit health and loyalty. And proactive management of consumer debt levels can go a long way in addressing today s legislative and consumer advocate pressures for greater consumer protections within lending practices (see sidebar on next page). FICO Credit Capacity Index (FICO CCI) generates a complementary measure that predicts which consumers can more safely manage new or increased credit. Combined with FICO Scores, FICO CCI helps bankcard lenders better target and set initial credit amounts and product terms, and refine account management actions such as credit line assignment and authorizations, within the guidelines of new regulations. Research shows that FICO CCI can effectively capture and leverage capacity. Validation results for both new and existing revolving accounts show that FICO CCI effectively rank-orders consumers most likely affected by incremental debt within each risk score range those whom, without a change in debt, would have the same expected risk of default. FICO CCI is extremely effective within existing credit strategies. Across income groups and risk levels, there are segments with relatively high, moderate and low capacities to manage increased debt, as shown for the mid-fico Score range of 665 699 in Figure 2. This knowledge can be used to target or refine offers of new credit for each group. www.fico.com page 3

»» Figure 2: Opportunities to Target Action FICO 665 699 by Income and FICO Credit Capacity Index TM Results on pooled bankcard sample POPULATION % 60% 50% 40% 30% 20% Legend Capacity Level Low Medium High 10% 0% UNKNOWN < $30K $30K 54K $55K 79K $80K 114K $115K+ FICO CREDIT CAPACITY INDEX WITHIN INCOME GROUPS Beyond regulatory compliance the value of capacity measures Using a model that empirically estimates a consumer s capacity to assume additional debt can help card issuers more accurately measure ability to pay. This, in turn, helps them: Increase control over loss exposure and reserves. Lenders can refine line assignments toward consumers best able to repay debt, and limit exposure for those posing the highest default risk. This would help reduce loss reserves and reallocate working capital to» more profitable areas of business to alleviate over-indebtedness. Grow portfolio profits conscientiously. Line assignments can more closely correspond with what a consumer can safely handle, minimizing losses. Demonstrate responsible lending practices. Proactive management of consumer debt loads would help address consumer advocacy and legislative pressures to alleviate over-indebtedness.»» Improve customer satisfaction, retention and corporate image. Public promotion of responsible lending practices could help boost customer ties and corporate image, and attract and retain more good customers. www.fico.com page 4

Improving Assessment of Ability to Pay How does FICO CCI compare with commercially available income estimators or self-reported income as a predictive measure of a consumer s ability to pay? FICO conducted research on a large pool of bankcard accounts to determine the accuracy of FICO CCI in predicting future risk of payment default compared to self-reported income and a monthly debt service ratio. Figure 3 charts future bad rates (90+ days delinquent) for the FICO 8 Score, FICO CCI, selfreported income and the monthly debt service ratio (MDS%). 1 This graph provides a comparison of the ability of each measure to independently rank-order credit risk. Each measure is ranked from the worst (expected highest risk) decile to the best (expected lowest risk) decile as follows: FICO Score lowest FICO Score range in decile 1 up to highest FICO Score range in decile 10. FICO CCI lowest CCI in decile 1; highest CCI in decile 10. Income lowest income in decile 1; highest income in decile 10. MDS% highest MDS% in decile 1; lowest MDS% in decile 10. Figure 3: FICO 8 and FICO CCI More Accurately Predict Ability to Pay Decile Ranking New and Existing Accounts General Performance (90+) 80% BAD RATE (90+ DAYS DELINQUENT) 70% 60% 50% 40% 30% 20% 10% Legend FICO 8 FICO CCI Self Reported Income Monthly Debt Ratio (MDS%) 0% 1 2 3 4 5 6 7 8 9 10 DECILE The FICO 8 Score, which is designed to rank-order credit risk, was indeed the most predictive of future delinquencies. We see that those in the highest-risk deciles resulted in higher bad rates, while those in the lowest-risk deciles resulted in lower bad rates. This reinforces the fact that FICO Scores continue to be essential in risk management strategies. 1 MDS% = Ratio of the monthly debt service and monthly income. Monthly debt service is a measure of the expected minimum payments consumers would have, based on their outstanding balances on their credit reports. A high MDS% indicates the consumer has a large portion of their income already consumed by minimum payments, while a low value indicates very little debt relative to the income. www.fico.com page 5

The chart also shows that FICO CCI is a much stronger predictor of ability to pay compared to self-reported income and the monthly debt service ratio, which separated future risk less effectively. Therefore, using FICO CCI to further segment accounts within the FICO 8 Score band is the most accurate and prudent way to ascertain both risk and ability to pay. Since most lenders will use the income measures in combination with other criteria, the study further investigated the ability of FICO CCI and income to differentiate risk for given FICO 8 Score ranges. Figure 4: FICO CCI Is More Predictive Within FICO 8 Score Bands FICO 8 by Estimated Income & CCI New and Existing Accounts General Performance (90+) 60% BAD RATE (90+) 50% 40% 30% Legend Low CCI Low Income Med CCI Med Income High CCI High Income 20% 10% 0% 300 649 650 699 700 759 760 799 800 850 FICO 8 SCORE In Figure 4, FICO CCI is shown to more effectively isolate the least risky (solid blue line) and riskiest (solid red line) subpopulations compared to the income estimator (dotted lines), which shows little variance in bad rate by FICO Score bands. For example, consumers in the 650 699 FICO 8 Score range with a high FICO CCI are roughly 14% lower risk than those with the highest estimated income grouping (19% vs. 22% for 90+ days delinquent). The bottom line is that while these measures can be useful within strategies and for compliance, they do not provide nearly the predictive value as FICO CCI in determining which consumers can more safely take on incremental debt.»» Managing Capacity A Practical Approach So, how would a bankcard lender use FICO CCI in practice? Let s explore a sample strategy for new bankcard accounts. For prescreen acquisitions and originations underwriting, strategies often include FICO Scores, income and other measures. www.fico.com page 6

FICO CCI would be added to existing strategies, providing new information on credit capacity not otherwise captured. This would help determine target offers and initial line assignments, especially near existing cutoff zones, as illustrated in Figure 5 where FICO CCI is used in addition to a FICO Score and income measure. Figure 5: Using Capacity in Practice Simple New Account Acquisition Very High Accept Maximum Line High FICO Score + Income Medium Decrease Line from Standard Standard Line Assignment Increase Line from Standard Low Very Low Decline / Accept Minimum Line Low Medium High FICO CREDIT CAPACITY INDEX TM While this sample strategy is somewhat basic 2, it demonstrates the overall approach for using FICO CCI: Consumers with very low FICO Scores and low income, as well as those with high FICO Scores and high income, would be treated the same as before, with few exceptions. Consumers with FICO Scores in the lender s operating range, missing income or other borderline areas, would be assigned either higher or lower lines, depending on capacity. For example, in the mid-fico Score range of 700 720, there would be opportunities to increase lines for highcapacity consumers, but reduce total lines by the same amount when capacity is low. This would minimize future losses and retain consistent exposure levels. A similar approach to Figure 5 could also be used in account management. FICO CCI would be included on top of existing measures, such as internal analytics, FICO Score, current delinquency and relationship status. FICO CCI is a powerful tool for use within existing lending strategies not a replacement for other risk measures. It s still essential, for instance, to consider the consumer s projected risk to repay the debt. After all, you may not want to extend credit to high-capacity consumers who are also high-risk. That s why FICO CCI and FICO Scores work hand-in-hand. 2 For more complex strategies, consider using FICO CCI in conjunction with additional measures such as internal scores, application information and down payment for more accurate risk assessment. www.fico.com page 7

Just as the FICO Score s rank-ordering of risk is used in strategies to determine whether to extend credit to consumers and doesn t provide a yes/no answer on its own FICO CCI operates much in the same way. It serves as a key part of the strategies that answer the question How much new debt is too much? Adding techniques like strategy optimization can help determine the optimal combination of risk measures to make fully informed credit decisions. A customized alignment process could also be incorporated into account management procedures to assess how customers in various segments behave relative to a range of credit lines, in order to determine specific line assignments. Top US Lender Results Tell the Story In recent validations on top US lenders bankcard portfolios, FICO examined the effectiveness of FICO CCI to identify consumers who could safely assume more credit. The results reflect that consumers at the same FICO Score range can have a wide variance in their ability to manage additional credit. The findings indicate that by knowing a consumer s ability to manage new credit, lenders can take actions to reduce delinquency rates and identify areas of profitability for real financial impact. FICO examined the bad rates (90+ days delinquent) for these lenders accounts by FICO 8 Bankcard Score range, shown in Figure 6. This was further segmented by the low, medium and high capacity as designated by FICO CCI. (Note: In this chart, ranges below 630 were removed to better demonstrate FICO CCI separation in realistic lending ranges.) The results clearly demonstrate that within each score band, FICO CCI accurately identified the highest bad rates attributed to the lowest CCI values and the lowest bad rates attributed to the highest CCI values. Figure 6: Determining Credit Capacity Within Segments of Similar Credit Risk FICO CCI Bad Rate by FICO 8 Score Range 16% BAD RATE (90+) 14% 12% 10% 8% 6% Legend Capacity Level Low Medium High 4% 2% 0% 630 669 670 694 695 714 715 739 740 759 760 779 780 799 800+ FICO 8 SCORE RANGES www.fico.com page 8

Figure 7 demonstrates that FICO CCI captures a different dimension of credit than a FICO Score and is not merely a finer ranking of risk within a FICO Score range. The average FICO Score within these ranges and within the FICO CCI assignments is very similar, and in many cases equal. When compared with the variance of FICO CCI in Figure 6, however, the difference in measurement factors is clear. Figure 7: FICO CCI Captures a Different Dimension than a FICO Score Average FICO 8 Score by FICO CCI Level 900 800 Capacity Level Legend Low Medium High 700 AVERAGE FICO 8 SCORE 600 500 400 300 200 100 0% Below 630 669 670 694 695 714 715 739 740 759 760 779 780 799 800+ FICO 8 SCORE The benefits of FICO CCI become further apparent in Figure 8. CCI levels are plotted against bad rates and incremental increases in balances (expressed as percentage increases over prior balances). When examining these different levels of incremental revolving balances against bad rates for the same risk population, the variance of bad rates is higher for the low CCI population. www.fico.com page 9

Figure 8: Refining Treatments for Low-Risk Segments FICO 8 Score 715 759 Segment BAD RATE (90+) 7% 6% 5% 4% 3% Legend Capacity Level Low Medium High Pop% 24% 52% 23% 2% 1% 0% NO INCREASE (56%) 1% 20% (9%) 21% 100% (14%) 101% 300% (9%) 300%+ (12%) PERCENT REVOLVING BALANCE CHANGE (PERCENT POPULATION) While this 715 759 score range typically represents a low-risk population, these lenders can now segment by CCI values and treat the segmented populations differently. For consumers falling in the low CCI population, where bad rates are higher for all incremental levels, the lenders may decide to decline credit or reduce initial credit limits. For those with high CCI values, they may decide to offer increases in credit lines, cross-sell promotions for other products and grant preferential customer service treatments.»» Calculating Financial Advantage FICO performed a validation to assess the financial value that FICO CCI could bring to a bankcard lender s origination activities. The study assumed different credit limit strategies were applied to consumers falling in targeted categories, resulting in significant acquisition of new revenue and loss avoidance for the lender. In the Figure 9 example, a population of 1 million new accounts booked was segregated by FICO Score range and by FICO CCI value (low, medium and high). The research found populations for action and applied following assumptions: Initial lines range: $6,000 $12,000. Increase initial line: $2,000 for high CCI. Lower initial line: $2,000 for low CCI. Revenue per good vary by FICO Score and FICO CCI: range $410 $510. Loss per bad vary by FICO Score and FICO CCI: range $4,900 $7,300. www.fico.com page 10

The resulting net profit to the lender was $1.6 million. Figure 9: Using FICO CCI Results in a Net Profit of $1.6 Million Scenario: Change from Champion Line Decision Adjust Initial Credit Card Lines from Current Champion +$2,000 for High CCI; -$2,000 for Low CCI Within FICO Score Ranges Lines Increased Lines Decreased CCI Challenger Impact % of Accounts with Different Initial Lines 14.5% 10.2% 4.3% Total Exposure Change $290,000,000 ($203,750,000) $86,250,000 Goods 140,854 96,651 Bads 4,146 5,224 Bad Rate 2.9% 5.4% Incremental Revenue $4,330,302 -$3,759,457 $570,845.12 Incremental Loss $3,345,994 -$4,326,486 -$980,492.35 Incremental Profit $984,308 $567,030 $1,551,337 Incremental Profit per Account $0.98 $0.57 $1.60»» Modeling for the Future Like the FICO Score, FICO CCI is built on credit bureau data and is designed to rank-order consumers according to risk for use within lending strategies. But FICO Scores reflect consumer risk based on today s credit mix. By contrast, FICO CCI measures consumer risk if he/she takes on future incremental debt. This new debt can be in the form of new credit accounts or increases in existing accounts. FICO CCI, now available at FICO, is based on patent-pending technology called Future Action Impact Modeling. Unlike traditional bureau-based risk modeling, Future Action Impact Modeling can isolate consumer sensitivity to new behaviors not currently present on the credit report and infer tolerance for incremental future debt. www.fico.com page 11

Conclusion Today s lending challenges call for new approaches to carefully target customer acquisition, manage collections activities and comply with government mandated regulations. As we emerge from the economic crisis, tools such as FICO CCI offer lenders a significant competitive advantage by adding a new dimension for a more complete picture of consumer risk. Learn more: Get additional information about FICO CCI Read the Insights paper, Credit CARD Act: Move Ahead of the Curve Contact us at 1-888-342-6336 or info@fico.com The Insights white paper series provides briefings on best practices, research findings and product innovations from FICO. To subscribe, go to www.fico.com/. For more information US toll-free International email web +1 888 342 6336 +44 (0) 207 940 8718 info@fico.com www.fico.com FICO, Credit Capacity Index and Make every decision count are trademarks or registered trademarks of Fair Isaac Corporation in the United States and in other countries. Other product and company names herein may be trademarks of their respective owners. 2010 Fair Isaac Corporation. All rights reserved. 2662WP 04/10 PDF