Petal offers alt-data underwriting tech to banks. Do they want it?

Petal, an online credit card provider, has started offering its cash-flow-based underwriting technology to banks and fintechs. It’s also providing to other lenders its own version of a credit score, CashScore, which is based largely on customers’ cash-flow data and bill-paying habits.

“We believe the credit underwriting of the future must look at the holistic financial picture of the consumer when we make these decisions,” says Jason Gross, CEO of Petal. “And we can no longer exclude the tens of millions of people in the United States that have a bank account and have income, but do not have credit history.”

The software, called Prism Data, will be available through an application programming interface. It takes raw bank account data and translates it into a picture of potential borrower’s credit risk, identity, financial status and more. It does use credit report data where it’s available. Petal does not share default statistics, but says consumers with no credit file underwritten using the CashScore perform like consumers with prime credit scores. Petal was one of the fintechs FinRegLab studied for a report that showed cash flow data can predict creditworthiness.

Petal’s technology could help banks provide credit cards to people who may be creditworthy but who lack a credit history, perhaps because they’re young or recently moved to this country. It could help banks compete with finechs like Petal and like TomoCredit, which recently launched a FICO-free credit card. But whether or not banks are ready and able to disrupt the way their credit card underwriting technology and processes is open to debate.

“For decades, companies and entrepreneurs have tried to build and sell a better tool for measuring and scoring creditworthiness, and there are lots of reasons that none have come close to overtaking the traditional system,” said Peter Wannemacher, principal analyst for digital banking at Forrester.

“Executives at banks have been using roughly the same methods for a really, really long time, so a new solution will need to be clearly better,” he continued. “It doesn’t just need to be provable that these models work, it needs to be proven. I think a lot of traditional providers will hold off to see how this plays out. But I do think a number of financial services companies will take a hard look at this, including fintech players and neobanks.”

In offering its underwriting software to others, a step it announced this week, the New York-based Petal is following in the footsteps of Upstart, Zest, Numerated and other fintechs that started out as direct lenders and pivoted to become technology providers to banks. But Petal will also continue to offer its own credit card, according to CEO Jason Gross. As of February it had issued 100,000 cards and it will have hundreds of thousands of cardholders by the end of the year, he said. WebBank, which is based in Salt Lake City, is the lender behind Petal.

Gross said he’s gotten calls from several fintechs and banks over the past year asking how they could use consumer-permissioned, transactional data in their underwriting, because they have been looking for reliable sources of signal during the pandemic.

“In the first few conversations, we were very flattered, but didn’t have any way to work with another business in that capacity,” Gross said. “But as we started to see more and more demand for this, we began thinking about how we could create a whole new line of business that’s focused on making bank account transaction data work for more complex and higher-level use cases.”

The maturity of the data aggregation technology platforms like Plaid and Finicity, which make consumers’ bank account data easy to obtain, and the bank regulators’ acceptance of the use of alternative data in underwriting have softened banks’ resistance to the idea of using cash-flow data in credit decisions, Gross said.

“Maybe the most powerful factor here is that COVID created a strong business need for more real-time, holistic underwriting data, because traditional credit scores became unreliable through the course of last year,” he said.

Will banks go for it?

Industry observers have mixed views on whether traditional banks are ready and willing to use Petal’s alternative methods of scoring and determining creditworthiness.

“Petal was already making a strong business case for the use of cash-flow underwriting and got a major assist from the economic volatility we experienced this past year,” said Leslie Parrish, senior analyst at Aite Group. “It makes sense that other lenders would want to benefit from an already tested model rather than trying to create one on their own.”

Parrish agreed with Gross that the joint statement bank regulators issued in 2019 about the use of alternative data in underwriting provided a green light for banks.

“Most banks are thinking about using account transaction data as a supplement to — rather than a substitution for — traditional credit scores,” Parrish said. “While those scores are highly predictive based on historical data, many lenders now want to pair that with data that provides a more real-time look at income and expenses.”

The most established lenders are not ready for this new approach, Wannemacher said.

“But for this to be successful, Petal doesn’t need the majority of traditional banks to sign up right away,” he said. “It needs to make and strengthen the business case for modeling and predicting creditworthiness in a new way.”

What Petal’s software can do

One of the things Petal’s Prism Data can do is make the bank account data gathered by account aggregators like Plaid and Finicity readable by banks’ underwriting systems.

The data Plaid makes available is raw, messy and unstructured, Gross said, and therefore unsuited to cash-flow-based underwriting.

“We know this better than most because these have been the challenges that we have faced in building our business over the last five years,” Gross said. “We’ve had to invest years and years and tens of millions of dollars in building these capabilities. We think we can save others a lot of time, a lot of money, a lot of effort in providing these capabilities off the shelf.”

The Prism Data software also categorizes bank account transactions.

“That’s a challenging problem to solve,” Gross said. “We do it using machine learning and artificial intelligence.” Other tech companies, including MX and Personetics, also provide categorization.

Prism Data also handles income verification and offers “insights” into a potential borrower’s financial life. One insight, for example, is a person’s cash inflows and outflows over the past three months.

The CashScore is based on the sum of all those parts, Gross said. Banks could use it the same way they use FICO and VantageScore numbers. It relies on bank account transaction history, so a CashScore can be generated for people who have no credit file, Gross said, and it can provide more insight into the behavior of people who have a credit score.

“A traditional credit score is based on the liability side of a consumer’s balance sheet,” Gross said. “It’s their use of debt over time.”

CashScore is based on money that a consumer makes, saves and spends on a monthly basis and is a real-time assessment of risk, whereas a traditional score lags behind real-time events by 90 days or more, he said.

“We believe the credit underwriting of the future must look at the holistic financial picture of the consumer when we make these decisions,” Gross said. “And we can no longer exclude the tens of millions of people in the United States that have a bank account and have income, but do not have credit history. This score solves for that inequality in access directly.”

Most Related Links :
Business News Governmental News Finance News

Source link

Back to top button