Use of Facebook ‘Likes' in Lending Decisions Raises Concerns

All Banking Law, All in One Place. Bloomberg Law: Banking is the comprehensive research solution that powers your practice with access to integrated banking-related legal news, analysis,...

By Kery Murakami

Oct. 30 — The increased use of Yelp reviews, Facebook “likes” and other social media interactions to determine a borrower's creditworthiness has caught the attention of regulators in Washington, amid concerns that the data is too subjective and could result in discriminatory lending patterns if not used properly.

Supporters of the practice, which include the credit reporting giant Experian and online marketplace lenders such as Kabbage, say positive social media interactions can help fill in the gaps on credit histories and help otherwise credit-challenged businesses obtain loans. But consumer advocates are concerned that some borrowers will be left behind if there's an overreliance on social media information and other alternative data, and that borrowers will have a more difficult time challenging their credit scores or credit reporting errors.

“The increasing amount of nontraditional data used in credit underwriting decisions may raise consumer protection questions,” Antonio Weiss, an adviser to Treasury Secretary Jacob Lew, said last week. “Just because a credit decision is made by an algorithm, does not mean it’s fair,” he said in Oct. 29 remarks at the Information Management Network Conference on Marketplace Lenders.

Weiss’s comments reflect some of the feedback that Treasury has received from a recent request for information on the growing marketplace lending industry. Consumer advocates are wary about the use of social media data and other types of information in lending decisions and credit scores, because it's not clear how the data is being applied.

“We have serious concerns about the discriminatory impact of using big data to determine a consumer’s creditworthiness,” the National Consumer Law Center (NCLC) said in its Sept. 30 comment letter. “Because big data scores use undisclosed algorithms, it is impossible to analyze the algorithm for potential racial discriminatory impact.”

Without more information, the concerns are still somewhat theoretical, NCLC Associate Director Lauren Saunders told Bloomberg BNA Oct. 19.

Social media poses “special risks,” the group said in its comment letter. Black people, for instance, tend have to have lower incomes and lower credit scores, the NCLC said. So if a potential borrower is judged on the credit scores of their Facebook friends, “that data can lead to clear discrimination against minorities compared to white borrowers,” the letter said.

The U.S. Public Interest Research Group and the Center for Digital Democracy raised similar concerns in separate letters on the request for information. The consumer advocacy groups said Treasury and other financial regulators should ask lenders to disclose what alternative forms of information they’re examining and how it’s being used.

Accurate Data?

Of particular concern, Weiss said in his remarks, is the ability of consumers to protest factual errors that influence their credit scores under laws such as the Fair Credit Reporting Act. “This is an area of real concern for consumers even when they have more robust protections,” he said. “The existing model is far from perfect, but applicants at least have the right to check — and correct — their personal data that gets used in credit decisions.”

The NCLC cited a study it conducted in 2014, in which 15 volunteers requested or purchased the information five data brokers had about them. The reports were “riddled with inaccuracies,” the study said. “The reports contained errors in estimated income, nearly doubling the salary of one participant and halving the salary of another, and eleven of the fifteen reports incorrectly stated the volunteer’s education level.”

Although the Treasury Department has no regulatory jurisdiction over marketplace lenders, it is working with state and federal regulators to strike a balance between allowing innovation and protecting consumers, Weiss said.

Help for Small Business

For many small businesses, the use of social media comments and other ways to determine creditworthiness brings promise.

While largely used by online marketplace lenders thus far, a process being developed by Experian could widen the use of nontraditional data to other lenders, Eric Haller, executive vice president of Experian DataLabs, the company’s research and development arm, said in an Oct. 23 interview with Bloomberg BNA.

The program, which may factor in data from Yelp, Facebook, Twitter and FourSquare to compare a business with others of the same type, will be offered to all of Experian's customers, including traditional banks, Haller said.

Experian will only use social media with business loans. Small businesses — particularly new startups that do not have established credit histories — are struggling to find credit, Haller said. Factoring in high Yelp reviews, or positive Twitter comments from customers, will only help their efforts to obtain financing, he said.

For lenders, using social media to evaluate potential borrowers improved the ability to predict defaults of commercial loans by 40 percent, he said, citing DataLabs' research.

If there is a bias, Haller said, it's that restaurants, beauty shops or other businesses that have social media interactions with customers have an advantage over other types of businesses.

Online Lenders Respond

In response to Treasury's request for information, a number of marketplace lending companies like Kabbage said that by operating online, they are able to incorporate a number of different factors into their evaluation of loan applicants. That allows the lenders to made credit decisions quickly.

Kabbage said its use of real-time information from bank accounts, social networks and web analytical services — accessed with the consent of applicants — gives the online lender “a more comprehensive view of the customer.”

“Because the data relates to all aspects of a customers' business (in the case of SMB loans), it is possible to see information related to volume (how robust the business is), character (if the business is able to repay, will they actually repay) and stability (likelihood of business persistence),” Kabbage said in its Sept. 30 letter.

In an Oct. 26 e-mail to Bloomberg BNA, a spokeswoman for Kabbage said the online lender does not use social media for the initial decision on whether to approve a business for a loan. But a business could qualify for a higher line of credit, the company said, if their social media interactions show they respond quickly to customers' questions, or if customers have nice things to say about them.

“These factors are in no way tied to socioeconomic variables for the business owner and focus on quality of service provided to customers,” the spokeswoman said.

To contact the reporter on this story: Kery Murakami in Washington at kmurakami@bna.com

To contact the editor responsible for this story: Mike Ferullo at mferullo@bna.com

The prepared remarks from Weiss are available at http://www.treasury.gov/press-center/press-releases/Pages/jl0238.aspx.

The comment letter from the National Consumer Law Center is at http://src.bna.com/Sq.

The NCLC's 2014 study can be found at http://www.nclc.org/issues/big-data.html.

The letter from online lender Kabbage is at http://src.bna.com/SG.