Analysis of Federal Reserve Research on Behavioral Scoring
Published: August 11, 2010
CRL Comments On "Report to the Congress on Reductions of Consumer Credit Limits Based on Certain Information as to Experience or Transactions of the Consumer"
The Federal Reserve recently issued a report based on a survey of credit card issuers that asked whether they had taken adverse action on consumers' accounts based on cardholder patronage of merchants by geographic location, merchant type or transaction type. This practice is commonly known as "behavioral scoring." The media have highlighted examples of card issuers judging consumers to be a high risk after observing that consumers engaged in marriage counseling sessions, frequented bars, or even shopped at Walmart.
The Fed report provides valuable information on a topic with little public information and confirms that some of the largest issuers in the country continue to use behavioral scoring tactics. Data in the report suggests that more than half of American households may have a credit card account with an issuer influenced by where the cardholder shops or what they buy.
CRL's comments on the Federal Reserve report fall into three main points:
1) General underreporting of the incidence of behavioral scoring among issuers.
The report's methodology understates the prevalence of behavioral scoring. It uses a narrow definition of when a consumer is affected, focuses on actions in a single month rather than a longer period, has the potential for bias, and may not fully account for the use of third-party vendors that provide behavioral-scoring models.
2) Certain adverse events experienced by consumers go unexamined.
The only outcomes considered relevant were increases in the cost of credit, reductions in the amount of credit available (line reductions), and account closures. Moreover, these three events were only counted if a letter – known as an "adverse action notice" – was sent to the consumer. Further, to be captured in the report, that adverse action notice must specifically indicate that behavioral scoring was the reason for the rate hike, reduced credit, or closed account. In reality, however, virtually every discretionary choice by an issuer regarding how to handle a credit card account will involve the issuer's risk assessment, often resulting in adverse outcomes for the consumer. Examples of discretionary choices that go unreported in the study include critical decisions about whether to lower rates or increase credit lines.
3) Impact on minority and low-income consumers is left unexamined.
The report does not thoroughly examine the affect of these practices on minority and low-income consumers and is unable to reach any conclusions regarding whether there is a disparate impact. Data on this may be limited, but the report could have explored this issue in more depth. CRL suggests ways to get closer to an estimate of the impact on these groups.
 Report to the Congress on Reductions of Consumer Credit Limits Based on Certain Information as to Experience or Transactions of the Consumer, Board of Governors of the Federal Reserve System, May 2010 (available at http://www.federalreserve.gov/BoardDocs/RptCongress/creditcard/2009/consumercreditreductions.pdf )