CFPB’s ‘BISG’ Algorithm Not Designed to Determine Race, Creator Says

The controversial algorithm behind the Consumer Financial Protection Bureau’s formula to determine disparate impact was originally created for healthcare research, according to an in-depth article published this week in the Los Angeles Times.

The Bayesian Improved Surname Geocoding algorithm, called BISG for short, is a formula that utilizes data from the Census Bureau, and it turns out, was created by Marc Elliot — a statistician at nonprofit research outfit Rand Corp.

Those in the auto finance industry have become familiar with the bureau’s use of BISG in its methodology for determining disparate impact. BISG estimates race and ethnicity based on an applicant’s name and census data.

Yet Elliott contends that BISG was designed to look at large groups of people, not to guess the race of individuals, according to the report. “If you want to know the difference in the percent of people with diabetes among people who are black and people who are white, you can answer that question much more accurately than asking, ‘Is this particular person black or white?’” he said. “That’s an inherently harder question.”

The CFPB’s use of BISG has been publicly criticized in the industry as far back as November 2014 when the American Financial Services Association commissioned and published a study entitled Fair Lending: Implications for the Indirect Auto Finance Market, which determined that  that the proxy correctly identified the race of African American less than 25%.

Regardless of the industry’s protest of BISG, the algorithm has been used to determine alleged discrimination at auto finance companies for the past three years, starting with Ally Financial Inc. in 2014.

Read a full history of the BISG algorithm here.

For more auto finance insights like this, register for the upcoming Auto Finance Summit, Oct. 5-7 at Bellagio Las Vegas. Visit www.AutoFinanceSummit.com for more information.