New types of data to determine prospective buyers’ creditworthiness may increase homeownership opportunities for Black and Hispanic Americans. However, standards need to be developed for the use of such data—which come from sources outside traditional credit bureaus—to ensure equitable and responsible lending practices, experts said Thursday during a National Association of REALTORS® webinar on alternative credit scoring.
Ann Schnare and Vanessa Perry, authors of the white paper “Tipping the SCALE: How Alternative Data in Credit Scoring Promote or Impede Fair Lending Goals,” said 21% of Black households and 19% of Hispanic households are considered “unscorable” because they don’t have access to traditional credit. Schnare and Perry suggested that three alternative data sets could help lenders determine “credit-invisible” consumers’ eligibility for loan products, including:
- Credit proxies: Payment histories for bills such as rent, utilities, and other financial obligations, which can suggest a person’s ability to pay and likelihood to default.
- Banking data: Account balances, check-writing history, and savings—information that can tell lenders whether a person is financially stable.
- Nonfinancial personal data: Data harvested from a person’s digital footprint, such as spending patterns and social media activity suggestive of their purchasing power or financial status.
“Minorities are far more likely to be ‘unscorable’ or have relatively weak credit scores using traditional credit bureau data,” said Schnare, president of AB Schnare Associates, a consulting firm specializing in housing and mortgage finance. “Incorporating additional data into the credit evaluation process can open doors for many deserving borrowers and boost minority homeownership rates.” She added that some countries where credit agencies don’t exist are already using alternative data to shape lending practices.
But such data could be misused without legal safeguards, said Perry, professor of marketing, strategic management, and public policy at The George Washington University’s School of Business. “When we start looking at how and where people spend their money, those kinds of indicators can be proxies for race, gender, or neighborhood,” Perry said. “Some of the same issues can apply in the case of social media data. People are assigned a trustworthiness score based on who they associate with online, which is obviously problematic.”
Schnare said banking data likely is the most promising source of alternative data for credit scoring because it’s a more objective way to evaluate someone’s finances and requires the consumer’s approval, which offers a layer of privacy protection. No matter which source of data is used, though, any lending practices around alternative credit scoring should meet the standards of what Schnare and Perry call SCALE.
- Societal values: Does the practice respect social and ethical norms, such as the right to privacy?
- Contextual integrity: Regardless of predictive value, is it relevant to mortgages?
- Accuracy: Does the data accurately reflect the household’s financial situation?
- Legality: Would the use of the data have a disparate impact on protected classes?
- Expanded opportunity: Would the use of the data increase the number of qualified borrowers?
“The rise of big data greatly expands the options for credit scoring,” Perry noted. “However, predictability is not enough to justify the use of certain kinds of data. Their use must also be consistent with broader social and ethical values.”
NAR President Charlie Oppler said during the webinar that the association plans to use Schnare and Perry’s research to help shape its policy positions and inform its future advocacy efforts on credit scoring.
“A borrower’s credit report and credit score are the gateway to a mortgage,” Oppler said. “But for too long, inaccurate credit reporting methods have raised the cost to borrow while limiting access to mortgage credit for prospective borrowers, particularly those from minority populations and rural communities.”
No comments:
Post a Comment