This will follow up on Barbara Mishkin’s January 14, 2016 blog “FTC warns use of big data may violate federal consumer protection laws.” At the American Bar Association’s Consumer Financial Services Committee meeting last month in Park City, Utah, Bryce Stephens and Jeffrey Langer from the CFPB provided insight as to how the CFPB treats the use of big data by creditors. The main point made by Mr. Stephens, an economist with the CFPB, is that the CFPB is unlikely to object to the use of big data by creditors to reconsider credit applications that would otherwise be denied. The CFPB will more closely scrutinize the use of big data when it is used to deny credit in the first instance.
Mr. Stephens indicated that the CFPB is continuing to develop an approach to analyzing the use of big data. The concern with using big data is that it may present fair lending issues if its use causes a disparate impact. Mr. Stephens advised that when a creditor determines that a big data factor may be leading to a discriminatory impact, the creditor should determine whether (1) the factor is highly correlated to the discriminatory impact, (2) there is a good basis for continuing to use that factor, and (3) there is a better variable that could be used for the same purpose that does not lead to a discriminatory impact. With respect to the FTC report and guidance regarding big data discussed in Barbara’s blog, Mr. Stephens stated that there is an “expectation that lenders comply.”
Jeffrey Langer, the Assistant Director of Installment and Liquidity Lending Markets in the Bureau’s Research, Markets, and Regulations Division, stated that to the extent that creditors are using third parties for collection or use of big data variables, creditors must follow the CFPB’s third party vendor management policy. Mr. Langer warned that creditors should be careful when using data that is predictive but could also have a disparate impact.
For more information about big data, attend our webinar, Big Problems with Big Data? FTC Report Warns Against Using Big Data, which we will host on February 17, 2016. A registration link appears here.