Christine
Cards: Bruised Under the Surface
http://www.creditcollectionsworld.com/cgi-bin/rs.pl?story=20030218CCWN197.xml
When Capital One announced a higher than expected charge-off rate in its credit card portfolio due to greater exposure to subprime accounts in the third quarter of 2002, it rattled the investing world and credit card lenders. More than that, it's leading to re-examination of automated credit.
Capital One has been known at least a decade for its state-of-the art credit scoring, modeling technology and information-based marketing that has allowed it to leapfrog competitors and become one of the largest and fastest-growing issuers with one of the lowest charge-off and delinquency rates in the industry.
But Capital One was hardly the only issuer to run into trouble with subprime accounts-generally defined as those with a Fair, Isaac credit score of 660 or less. Providian, which specialized in this market, announced charge-offs of more than 17 percent in the second quarter. Metris had similar problems with subprime loans this year, and Internet-only credit card issuer NextCard was forced to close its doors.
After years of fast growth due in large part to modeling and scoring technology enabling instant decisions to give credit to new customers, experts say credit card issuers must emphasize the basics of portfolio risk management when it comes to subprime accounts. Others say they must make better use of existing and new risk management technology.
"All of the advances in models and scoring technology clearly led some folks to be susceptible to this false sense of confidence that the models were always right," says John Grund, partner at First Annapolis Consulting, based in Baltimore. "Modeling began in some issuers to outweigh the need for prudent judgment. We have all been taken back to reality that technology does not put portfolios on autopilot. There's a need for basic, prudent risk management."
Others say the widespread prevalence of subprime lending makes it necessary for banks to rely more heavily on risk management technology to eliminate human interaction. "Most of the credit card issuers could be doing a better job of using the technology and minimizing the human interaction, and that is what distinguishes the successful from the unsuccessful issuers," says Avivah Litan, vp of financial services for Gartner. "They all have access to the same credit bureau information, it all comes down to the proprietary analytical tools that they use to analyze the risk on the data from the bureaus."
This leaves issuers in a bind. With the economic downturn, growth rates have moderated, and competition for accounts that represent the best credit risks is escalating. At the same time, attentions have naturally turned to the less desirable credit risks, or the subprime pool. With fees and annual percentage rates that are often twice those of prime account holders, there are rewards for those who can best mine the subprime markets.
Meanwhile, charge-off and delinquency rates are also much higher-between 15 percent and 20 percent, compared to about 6 percent for prime loans-and so the technology required to manage subprime accounts needs to be more sophisticated. Capital One, citing credit rating data from Equifax Information Services in July 2002, estimates nearly 37 percent of all U.S. credit card loans are for customers with FICO scores of 660 or less.
"That means by definition there are other players in that space, who may not have discrete subprime strategies, but who have loaned to customers who are more risky than the general population," Grund says.
Where risk management vendors say banks have done a particularly good job is implementing technology that allows them to find potential new cardholders and grant them credit almost instantaneously. "Technology is critical in having instant, risk-based decisions on particular prospects, and modeling those customers with high rates of charge-offs," says Gene Devine, svp of Boca Raton, FL-based Data Warehouse Corp., which provides prospect databases for the subprime market, primarily for mortgages. "Risk-based decision engines do that."
Credit issuers have not done such a good job, he adds, in applying management technology to risky portfolio segments on an ongoing basis. "You would be surprised how complacent portfolio lenders get until they see the explosion," Devine adds.
In particular, he says lenders with large subprime portfolios need to do more regression modeling, which statistically determines the probability of how accounts will perform in the future. That information can then be used to determine better risk prospects on the decision-making end, he says. "You plug that into a decision engine and you can tell who will be risk-based and who won't," he adds.
Dennis Capozza, founding principal for risk management firm University Financial Associates, based in Ann Arbor, MI, says another key thing that subprime bankers need is technology that helps them manage different parts of their portfolios based on changing economic conditions. With that in mind, his company has built predictive modeling software that helps lenders predict loan performance based on the underlying economic conditions on both a national and a local level.
"Executives get blindsided by changing economic conditions," he says. "You have to have a forward-looking perspective, and try to get that perspective into the lending institutions so they understand how current economic conditions are affecting their expected loan performance."
Some subprime bankers agree. Irving J. Levin, who founded subprime credit card company First Consumers National Bank, and is now CEO of Genesis Financial Solutions, a Portland, OR-based company that buys and collects non-performing loans, says the recent troubles of subprime credit card issuers are largely due to the economy taking portfolio managers by surprise. "It was a case of the managers not being prudent and the economy caving," he says.
Levin says the nature of lending to the subprime credit market is more risky, and when the economy is distressed, "people on the margins are the ones who most acutely feel the pain, because they lose their jobs and homes the quickest." It doesn't help that card issuers are among the most aggressive marketeers out there, some sprinkling debt like pixie dust to people in long-term binds that make it easy to say yes and hard to pay later. He added that "it would be bad enough if there were not a credit expansion bubble over the last 10 years, but this was like two trains running into each other."
Hollis Fishelson-Holstine, vp of analytics for Fair, Isaac, says the FICO scores her company produces are sophisticated tools that enable bankers to glimpse at consumer borrowing behavior for a period of two years out. Bankers, she says, typically add their own predictive modeling to this score, taking into account a mind-boggling number of variables, such as the geography where they hope to build business, state of the economy and pricing models. "A lender has to constantly have a look at their origination policies and ongoing management, keeping in mind a complex mix of objectives, such as maximizing profits and managing their contingent liabilities," she says.
Levin says increased regulatory scrutiny of subprime lenders will also increase the need for portfolio management technology. "The regulatory community is really cracking down and intensely scrutinizing those with large subprime exposure," he says. "In a funny way, technology will be a necessary element of companies who want to be in this market to meet regulatory requirements with better data tracking systems and better controls of the data in terms of managing lending lines."
2003-02-18
http://www.creditcollectionsworld.com/cgi-bin/rs.pl?story=20030218CCWN197.xml
When Capital One announced a higher than expected charge-off rate in its credit card portfolio due to greater exposure to subprime accounts in the third quarter of 2002, it rattled the investing world and credit card lenders. More than that, it's leading to re-examination of automated credit.
Capital One has been known at least a decade for its state-of-the art credit scoring, modeling technology and information-based marketing that has allowed it to leapfrog competitors and become one of the largest and fastest-growing issuers with one of the lowest charge-off and delinquency rates in the industry.
But Capital One was hardly the only issuer to run into trouble with subprime accounts-generally defined as those with a Fair, Isaac credit score of 660 or less. Providian, which specialized in this market, announced charge-offs of more than 17 percent in the second quarter. Metris had similar problems with subprime loans this year, and Internet-only credit card issuer NextCard was forced to close its doors.
After years of fast growth due in large part to modeling and scoring technology enabling instant decisions to give credit to new customers, experts say credit card issuers must emphasize the basics of portfolio risk management when it comes to subprime accounts. Others say they must make better use of existing and new risk management technology.
"All of the advances in models and scoring technology clearly led some folks to be susceptible to this false sense of confidence that the models were always right," says John Grund, partner at First Annapolis Consulting, based in Baltimore. "Modeling began in some issuers to outweigh the need for prudent judgment. We have all been taken back to reality that technology does not put portfolios on autopilot. There's a need for basic, prudent risk management."
Others say the widespread prevalence of subprime lending makes it necessary for banks to rely more heavily on risk management technology to eliminate human interaction. "Most of the credit card issuers could be doing a better job of using the technology and minimizing the human interaction, and that is what distinguishes the successful from the unsuccessful issuers," says Avivah Litan, vp of financial services for Gartner. "They all have access to the same credit bureau information, it all comes down to the proprietary analytical tools that they use to analyze the risk on the data from the bureaus."
This leaves issuers in a bind. With the economic downturn, growth rates have moderated, and competition for accounts that represent the best credit risks is escalating. At the same time, attentions have naturally turned to the less desirable credit risks, or the subprime pool. With fees and annual percentage rates that are often twice those of prime account holders, there are rewards for those who can best mine the subprime markets.
Meanwhile, charge-off and delinquency rates are also much higher-between 15 percent and 20 percent, compared to about 6 percent for prime loans-and so the technology required to manage subprime accounts needs to be more sophisticated. Capital One, citing credit rating data from Equifax Information Services in July 2002, estimates nearly 37 percent of all U.S. credit card loans are for customers with FICO scores of 660 or less.
"That means by definition there are other players in that space, who may not have discrete subprime strategies, but who have loaned to customers who are more risky than the general population," Grund says.
Where risk management vendors say banks have done a particularly good job is implementing technology that allows them to find potential new cardholders and grant them credit almost instantaneously. "Technology is critical in having instant, risk-based decisions on particular prospects, and modeling those customers with high rates of charge-offs," says Gene Devine, svp of Boca Raton, FL-based Data Warehouse Corp., which provides prospect databases for the subprime market, primarily for mortgages. "Risk-based decision engines do that."
Credit issuers have not done such a good job, he adds, in applying management technology to risky portfolio segments on an ongoing basis. "You would be surprised how complacent portfolio lenders get until they see the explosion," Devine adds.
In particular, he says lenders with large subprime portfolios need to do more regression modeling, which statistically determines the probability of how accounts will perform in the future. That information can then be used to determine better risk prospects on the decision-making end, he says. "You plug that into a decision engine and you can tell who will be risk-based and who won't," he adds.
Dennis Capozza, founding principal for risk management firm University Financial Associates, based in Ann Arbor, MI, says another key thing that subprime bankers need is technology that helps them manage different parts of their portfolios based on changing economic conditions. With that in mind, his company has built predictive modeling software that helps lenders predict loan performance based on the underlying economic conditions on both a national and a local level.
"Executives get blindsided by changing economic conditions," he says. "You have to have a forward-looking perspective, and try to get that perspective into the lending institutions so they understand how current economic conditions are affecting their expected loan performance."
Some subprime bankers agree. Irving J. Levin, who founded subprime credit card company First Consumers National Bank, and is now CEO of Genesis Financial Solutions, a Portland, OR-based company that buys and collects non-performing loans, says the recent troubles of subprime credit card issuers are largely due to the economy taking portfolio managers by surprise. "It was a case of the managers not being prudent and the economy caving," he says.
Levin says the nature of lending to the subprime credit market is more risky, and when the economy is distressed, "people on the margins are the ones who most acutely feel the pain, because they lose their jobs and homes the quickest." It doesn't help that card issuers are among the most aggressive marketeers out there, some sprinkling debt like pixie dust to people in long-term binds that make it easy to say yes and hard to pay later. He added that "it would be bad enough if there were not a credit expansion bubble over the last 10 years, but this was like two trains running into each other."
Hollis Fishelson-Holstine, vp of analytics for Fair, Isaac, says the FICO scores her company produces are sophisticated tools that enable bankers to glimpse at consumer borrowing behavior for a period of two years out. Bankers, she says, typically add their own predictive modeling to this score, taking into account a mind-boggling number of variables, such as the geography where they hope to build business, state of the economy and pricing models. "A lender has to constantly have a look at their origination policies and ongoing management, keeping in mind a complex mix of objectives, such as maximizing profits and managing their contingent liabilities," she says.
Levin says increased regulatory scrutiny of subprime lenders will also increase the need for portfolio management technology. "The regulatory community is really cracking down and intensely scrutinizing those with large subprime exposure," he says. "In a funny way, technology will be a necessary element of companies who want to be in this market to meet regulatory requirements with better data tracking systems and better controls of the data in terms of managing lending lines."
2003-02-18