Fed: Credit companies admit profiling credit card users
It's rare, but yes: What you buy, where you buy it can be held against you
Hundreds of thousands of credit cardholders' accounts have been zinged in recent years by credit card companies based in part on where consumers shopped, what they bought, who they bought from or who held their mortgages, according to a new federal report issued Friday.
The cardholders were hit with credit limit reductions, interest rate hikes or had their accounts closed by issuers who told federal regulators that decisions to clamp down on credit to these consumers were based on tracking their spending and loan data. Among the consumer shopping practices that triggered negative account changes:
- The location of where transactions were made.
- The identity of the merchant processing the transaction.
- The type of credit card transaction.
- Identity of the mortgage lender.
|WHAT YOU BUY, WHERE YOU BUY SOMETIMES HELD AGAINST YOU|
|Congress ordered the Federal Reserve to study whether banks were mining credit card transaction data -- such as what people bought, where they bought it, who they bought it from -- and using it to take punitive action against those customers. After surveying the 175 institutions that issue 98 percent of the general purpose credit cards in the U.S., the Fed answered: Yes, it's rare, but it happens. Here are the actions the banks took, how many of the 175 banks took them, and how many customers were affected.|
|Action taken ...||... By this many banks ...||... Against this many customers|
|Reduced credit limit
||6 (including 5 of the 25 largest banks)||323,714|
|Raised interest rate||2 (including 2 of the 25 largest banks)
|Closed a credit card account
||4 (including 4 of the 25 largest banks)
|Source: Federal Reserve, "Report to the Congress on Reductions of Consumer Credit Limits Based on Certain Information as to Experience or Transactions of the Consumer," May 2010
The 72-page report, conducted by the Federal Reserve Board, was quick to point out that profiling card users' spending habits was rare among credit card issuers and actually affected a relatively small number of card users. Still, the report gives the first official glimpse at how some in the credit card industry have used a technique called behavioral modeling to mine spending data for clues about whether customers will default on their credit card loans.
The types of behavior that might trigger adverse credit card changes could include changes in spending habits. If you shop at pawn shops, casinos, discount stores and low-end retailers when you haven't frequented these establishments before, it might signal an increased risk of defaulting on credit card loans. Having a troubled mortgage lender in a ZIP code full of foreclosures might mean your net worth is rapidly dwindling. Buying retread tires could also hint at someone just scraping by.
Or, critics say, the changes in behavior could mean nothing but you blew off some steam, shopped with a new friend or found yourself a super bargain.
The Fed report confirmed a practice that was widely reported and criticized by consumer groups: Where you shop and the type of people who shop at that same store could impact your credit limit. The Fed report recounts the response of an unnamed credit card issuer: "In considering whether to reduce the credit line of a given cardholder, this issuer considered the performance and spending patterns of other cardholders with similar credit-related characteristics who shopped at the merchants where the given cardholder had made purchases."
Two credit card issuers reported using merchant category codes and geographic locations of stores along with the customer's payment history to develop risk models. "A third issuer grouped all charges into a small number of very broad categories and considered the performance of its other customers within those categories in deciding on line reductions for a given individual," according to the report.
The report was a mandate of the Credit CARD Act of 2009. It required the Fed to investigate whether credit card companies engaged in these practices at any time during a three-year period before the credit card reform law was enacted.
Is it profiling?
Members of Congress, led by Rep. Maxine Waters of California, were concerned after reports surfaced in 2009 about credit cardholders whose credit limits were reduced and interest rates jacked up because they began to shop at discount stores. Other consumers complained they may have been hit with less favorable credit terms based on where they lived. Waters, who was not available for comment on the new Fed report, suggested last year the practices might be a form of "profiling," targeting certain groups for different treatment.
For those institutions that have used such information, the proportion of cardholders actually affected has been small.
Federal Reserve report
Waters said the actions also appeared to border on credit card redlining, a controversial practice used in the insurance or mortgage industries during the 1960s and 1970s to keep African Americans and other minority groups from purchasing homes in certain neighborhoods or targeting some areas for discriminatory treatment. Today, federal fair housing and equal opportunity lending laws ban such treatment.
The Fed report was based on a survey of 175 financial institutions, including the 100 largest issuers of general-purpose credit cards. Together, the banks, credit unions and lending firms included in the survey issue about 98 percent of the open-end consumer credit cards in the United States. The Fed sent written surveys to the 175 lenders in November 2009. Credit card issuers were asked if they had engaged in any of a list of practices between Nov. 30, 2006 and Nov. 30, 2009. The CARD Act requested information for a three-year period, but because banks are required to keep data for only 25 months, the Fed's results are based on the shorter reporting period of Nov. 1, 2007 to Nov. 30, 2009. Participation in the survey was mandatory. In addition, Fed investigators interviewed other banking regulators and major credit card issuers as part of the study.
The Fed's report does not identify any of the credit card issuers involved in the study. Of the major credit cards issuers, only American Express acknowledged in 2009 that it had used information about where consumers shop to lower credit limits. But after negative news coverage and public outcry over the practice, the company said it would discontinue the practice. The Fed said that several of the card issuers in the survey said they would discontinue monitoring practices because of negative publicity about their methods.
Of the 175 institutions, six -- including five identified as among the largest commercial banks -- indicated they had engaged in at least one of the questionable practices that had contributed to the decision to lower the cardholders' credit limits. Of the six issuers, two also said they had raised interest rates and four said they closed accounts based on cardholders' behavior. No credit unions in the survey reported using any of the practices.
Relatively rare practice
The proportion of cardholders affected was small compared to the total number of U.S. credit cards and the cases where profiling behavior or shopping habits were used were "relatively rare" incidents. "For those institutions that have used such information, the proportion of cardholders actually affected has been small," according to the report. It cited the example of two credit card issuers that together have 53 million active accounts. Only about 340,000 of those accounts had credit limit reductions -- for any reason -- in November 2009. Of those, only about 1,900 accounts had credit limits reduced because of any of the questionable monitoring practices during that month.
All credit card issuers said the decisions to reduce credit or increase rates were not based solely on the spending data. The issuers weigh a number of factors when reviewing accounts, including past payment history. According to the Fed: "Discussions with card issuers and their federal supervisors found that other metrics used to manage credit risk, such as measures of late payment behavior or measures of cash advance activity, are much more important factors in reducing lines or setting interest rates than the practices" reviewed in the study. "Moreover, significant changes in broad economic conditions can be important factors leading to changes in account terms or account closures."
If the delinquency arose from 'excessive' gambling, they would cut the available credit line or close the account.
|-- Federal Reserve report|
The report revealed that some credit card issuers are also monitoring other types of transactions not specifically requested for review in the CARD Act. Some issuers said they consider how much consumers spend and how frequently they use their credit cards as indications that they may be at greater risk of defaulting on their credit card loans. Four banks said gambling activity of cardholders who were behind on their payments was a factor in reducing credit limits. "If the delinquency arose from 'excessive' gambling, they would cut the available credit line or close the account," according to the report.
Cash advances as red flag
In addition, "One card issuer indicated that, for some of its new cardholders, it monitors cash advance activity in conjunction with card purchases at certain types of merchants (for example, jewelry and electronic goods stores). According to this issuer, it has found such behaviors among new cardholders to be an early warning of elevated credit risk."
In another instance detailed in the report, a credit card issuer admitted it had lowered credit limits on cardholders with low credit scores if they did not shop at the retailer affiliated with the card issuer. "This issuer reported that it found that cardholders with lower scores who did not shop at the affiliated retailer had elevated delinquency rates," according to the Fed.
For those uneasy about the amount of personal data that is tracked, the Fed offers this explanation of the tracking methods used by industry: "Card issuers that make use of transaction-related items in adjusting account terms tend to rely on fairly general metrics -- for example, the total amount spent on a purchase rather than the amount spent on a specific item; broad merchant category codes, such as hardware stores, entertainment establishments, or grocery stores, rather than a specific merchant identity; or general geographic location of the merchant rather than the street address, census block group or census-tract designation, or local five- or nine-digit ZIP code of the business."
Racial discrimination unclear
The report concluded it was "impossible to know" whether there was a link between negative credit card changes and demographics, such as the race or income level of the cardholder. The Equal Credit Opportunity Act (also called Regulation B) prevents lenders from collecting information about the race or ethnicity, among other things, of borrowers. However, a 2008 Boston Fed study by economist Ethan Cohen found a potential link when he studied U.S. Census tract data based on the ZIP codes that are provided on credit reports. People living in white neighborhoods were more likely to be approved for new credit cards than those living in black neighborhoods.
Fraud detection tool
The Fed made no specific recommendations, but urged Congress to be cautious in writing any law to ban or limit monitoring of cardholder spending. The same tools used to monitor spending habits for credit reduction purposes are also vital in detecting fraudulent use of credit cards.
"Transaction-specific information, such as the identity of the merchant or the amount of a charge, is an essential element of fraud detection and prevention systems," the Fed report notes. "In this context, credit card issuers routinely use transaction-related information to temporarily block access to accounts or to close accounts until the cardholder can be contacted to verify the authenticity of the transaction."
"The identity of the merchant, the merchant location, and the size and timing of charges are all central elements of fraud detection and prevention systems," according to the report.
Published: May 28, 2010
- Poll: Fraud alert false alarms common – As banks get more aggressive in combating fraud, about 7 in 10 Americans contacted about possible fraud received at least one alert in error ...
- PayPal lending arm fined for deception, abuses – PayPal Credit, formerly known as Bill Me Later, misled customers and ratcheted up their costs with unfair practices, consumer bureau charges ...
- US proposes anti-steering rules on campus cards – A proposed federal rule restricts marketing and fees on campus-affiliated cards that are lucrative for the schools but not the best choice for students ...