3 Credit Bureau

Larger Issues however, Spirer concedes issuers that take this action create larger issues. "The less data issuers provide, the less valuable it is," he explains. "This begins to undermine the third-party objective data resource of the bureaus themselves."

Recently, San Rafael, Calif.-based Fair, Isaac and Company Inc., the 3 Major Credit Bureau leading credit-scoring company, conducted research to determine what impact the elimination of credit-limit data would have on its scoring models if 75% of the issuers were not reporting. Sally Taylor-Shoff, director of credit-bureau products, says that with Fair, Isaac models, if high-balance or credit-limit information is missing, then the entire account would be ignored in determining potential risk.

Generally, she says, many scoring models use credit limits to evaluate credit utilization. The result would be an account getting a lower score indicating higher risk if it had a high credit limit with an issuer that is no longer reporting that data, but low balances. On the other hand, Taylor-Shoff explains, a consumer who is up to the limit on the trade line that is no longer reporting credit limits may appear to be a better risk than he would have because that trade line is being ignored.

"Any reduction of valuable data at the credit bureau will have an impact on the value of that data, and therefore the value of any pools such as scoring models that use that data," she says.

What the Fair, Isaac research determined, was that on an aggregate rather than individual level, there wasn't much of an impact. Overall, the research showed lower scores, but only by a few points, Taylor-Shoff says. Approximately the same amount of accounts moved up as those that moved down in score, thus, canceling each other out.

"The predictiveness of the models is still very high," she says. "What that tells us is if we are eliminating some of the limit data, then there is still so much data at the bureaus that at least with the Fair, Isaac risk scores, there wasn't that much impact in the predictive power."

However, Taylor-Shoff doesn't want to imply that credit-limit information is not important.

"The bureaus provide value by providing the data," she says. "The data has value in it. Any threat to the amount of data that is at the bureau is a problem for the bureaus as well as the entire credit industry."

Certain models will be affected more than others depending on how the scoring model treats credit-limit information and what it is trying to predict, she says. Issuers will have to do their own research on how their models treat missing data.