From the very start of the Paycheck Protection Program last year, it was clear that minority entrepreneurs, especially Black business owners, struggled more than White borrowers to find a willing lender. A new research project indicates that the problem was particularly pronounced at smaller banks — and human bias appears to be the main reason.
The majority of Black borrowers who received aid from the $800 billion relief program got their loan from a financial technology company, not a bank, according to an economic working paper released Monday. The skew toward those so-called fintechs was far sharper among Black borrowers than any other racial group.
“I was taken aback by the striking disparity — it was a surprising and unexpected fact, and we wanted to figure out why,” said Sabrina T. Howell, an assistant professor of finance at New York University’s Stern School of Business and the lead author of the paper.
It turned out that the automated loan vetting and processing systems used by the fintechs, as well as some of the nation’s biggest banks, significantly improved approval rates for Black borrowers, the researchers found. They didn’t find such stark gaps for any other racial group they examined, including Asian and Latino applicants.
The findings come amid growing scrutiny of how algorithmic systems can inadvertently perpetuate biases. Regulators like the Consumer Financial Protection Bureau are examining whether lenders using such systems run afoul even inadvertently of fair-lending laws.
But Howell said her new research helped illustrate how technology could also help level the playing field.
“The human brain is a much scarier black box than any machine- learning algorithm,” she said. “You can constrain an algorithm to meet fair-lending standards and you can ensure the data it trains on isn’t biased. That may be hard to do, but it’s a clear and objective possibility. Whereas when you have a human loan officer who is in front of someone and making a decision, you can never do that.”
A trade group for small banks, the Independent Community Bankers of America, defended its members, saying that community lenders had “outperformed the rest of the banking industry in serving minority-owned, women-owned and veteran- owned businesses.”
In particular, the group criticized the steps the researchers had to take to determine the race of applicants. Collecting data on borrowers’ ethnicity was optional for lenders, so Howell and her colleagues used Census Bureau data on business owners’ locations and surnames to project what race they were likely to be. The banking group said those methods turned the research into “an unreliable guessing game.”
But Sergey Chernenko, an associate professor of finance at Purdue University’s Krannert School of Management who was not involved in Howell’s research, said the new paper aligns with his own findings on race-based gaps in Paycheck Protection Program lending. At an economic conference next month, he is presenting a paper that concluded that Black-owned businesses were disproportionately left out of the relief program.
“This fits very well with and complements our finding that minority-owned businesses were less likely to get loans because of racial bias, and to the extent that they do get them, they’re more likely to get them from fintechs than banks,” Chernenko said.
The government designed the Paycheck Protection Program to be virtually risk-free for lenders: They would advance small companies up to $10 million the size of the loan was based on the company’s head count and payroll and the government would then pay off the loans in full for business owners that followed the rules. If the borrower defaulted, the government would still repay the lender. In theory, any lender should have been willing to lend to any qualified applicant.
It didn’t work out that way. Many banks limited their loans to their current customers, which was a hurdle for owners who lacked business checking accounts or loans. But even Black owners who had accounts were noticeably more likely than those of other races to end up with a fintech loan, Howell and her co-authors found.
This was not the case, they found, at the nation’s biggest banks. After researchers controlled for those elements, Blackowned businesses appeared to be just as likely as any other to get a loan from Bank of America, Citibank, JPMorgan Chase and Wells Fargo.
What the big banks and the fintechs had in common was automation. In one especially striking example, the authors studied a group of smaller banks that shifted partway through the Paycheck Protection Program to using automated systems from several fintech companies, including Biz2Credit. Their share of loans to Black-owned companies noticeably increased after the switch.
“You’d hope to find zero evidence of discrimination in PPP because the banks largely faced no credit risk at all,” Howell said. “What happens when banks are putting their own money on the line?”
The community bankers’ group called its members the “unequivocal leaders” of the economic recovery and said its lenders accounted for nearly 60% of the loans made by the program a point that underscores one fuzzy part of the study, which has not yet been formally peer-reviewed but was circulated Monday by the National Bureau of Economic Research.
There can be some overlap between what counts as a small bank and what counts as a fintech. Some banks — most prominently Cross River Bank in Fort Lee, New Jersey, which was the program’s sixth-biggest lender — are community banks that operate almost exclusively as service providers for online lenders. In Howell’s data, her team counted Cross River as a fintech lender; the Independent Community Bankers of America counts it as a small bank.
Shaundell Newsome, the chair of the Urban Chamber of Commerce, an alliance of Black businesses in Las Vegas, said that the Paycheck Protection Program was not an exception to the long history of Black business owners being hobbled by discriminatory lending.
“It’s a never-ending cycle,” Newsome said. “There should have been a stipulation for lenders to help the most affected and vulnerable groups first.”
Howell said she hoped her study’s results and the growing body of research on racial bias in lending would spotlight the ways technology may help banks make fairer credit decisions.
“There are times where there may be real benefits to removing humans from the process,” she said.