Cash flow underwriting — the use of consumers' bank and other financial transaction data in making decisions about whether to lend them money — has
Understanding how money flows in and out of a consumer's bank account or brokerage account, many feel, gives a clearer perspective on a potential borrower's financial picture and ability to repay a loan than does the traditional bank loan decision system, which relies heavily on a person's FICO score and select pieces of credit bureau data that are often a few months old. Lenders could use cash flow underwriting to make loans accessible to the estimated 100 million Americans who have no credit file, or a thin one, and no FICO score.
"In our current credit system, there's a catch-22," said
Nova Credit offers a cash flow underwriting system called
Yet the idea is still foreign to many banks that stick to their FICO-based lending decision software for a host of reasons, including real and perceived cultural, regulatory and technology hurdles.
The challenge of getting leaders to accept the idea
The biggest obstacle to adopting cash flow underwriting seems to be getting leadership to agree to this newish way of making loan decisions.
"Finding a stakeholder to come along on this ride with us was our first hurdle and our first challenge," said Nancy Berger, assistant vice president of credit strategy and administration at AT&T, which uses cash flow data in its device-financing decisions. In her case, she was able to convince others that cash flow underwriting would lead to more sales.
The traditional way of pitching the use of new types of data in loan decision engines is to buy a new dataset and test it, said Richard Franks, global head of risk strategy at PayPal. In this case, there isn't an archive of data to purchase.
"What I've seen successful is to say, 'Hey, let's not make the story solely about the underwriting case,'" Franks said. "Let's make it about the whole thing. So let's make it about income verification, verification of bank accounts, disbursement of funds." Cash flow data can be used for these and other purposes.
Persuading others to use cash flow underwriting is a matter of explaining the benefits of it, according to Tim Hong, chief product officer at MoneyLion, which announced this week that it partnered with Nova Credit to activate cash flow underwriting within its hosted decisioning engine. This collaboration will let lenders on MoneyLion's platform integrate Nova Credit's cash flow data analytics with their loan decision models. Subprime lender Concora Credit is the first to use it.
"Ultimately we were motivated by this notion of consumer empowerment and user permissioned data" that could be used to help make loan decisions, Hong said.
Former Comptroller of the Currency Eugene Ludwig pointed out the dangers of the status quo — loan decisions based largely on FICO scores — for individual banks and for the industry.
"What worries me is when one tool becomes the favorite of a regulator and becomes the tool by which things get measured, because people are different and their circumstances are massively different," Ludwig said. "This one-size-fits-all approach, I think, is really dangerous and damaging. The more we can get robust data to make a credit decision, that's better from an economic perspective and financial institution perspective and actually doing good for the borrower."
Tools like cash flow underwriting "are a way of getting as robust a picture technologically as one can, using data as one can in terms of who the borrower is and what their capacity is," Ludwig said.
Compliance risks
One compliance challenge to using cash flow underwriting is getting consumers to give consent for a lender to suction up their bank account data and to share their passwords. This obtaining of consumer permission and data is sometimes called "conversion." The rules around this will become clearer when the Consumer Financial Protection Bureau releases its rules implementing section 1033 of the Dodd-Frank Act.
MoneyLion did a recent survey of its customers, and 90% said they would be willing to opt in to data sharing, Hong said.
Other compliance challenges include the need to create and manage adverse action codes when consumers are denied loans and the need to manage disputes when this turns into an argument.
A recent FinRegLab paper on cash flow underwriting concluded, "Cash-flow data underwriting holds substantial promise." But it also noted there are risks.
"While the increasing use of cash flow data in credit underwriting is providing benefits for consumers and small businesses, it also raises privacy tradeoffs and potential concerns about fairness, accuracy, data security and transparency," the paper stated. "These issues are not limited to loan origination but also can arise in connection with loan servicing and firms' re-use of cash-flow data for other commercial purposes. Although some positive developments are occurring, uncertainty about the application of existing laws and inconsistency among market actors could become an increasing source of risk as the market continues to expand and evolve."
Esipov believes the CFPB's 1033 rule will "clear the fog" around regulatory compliance and cash flow underwriting.
"There's some regulatory ambiguity that's creating hesitation," he said in an interview. "It's been anticipated for so long. I think seeing that final rule come through in the next few weeks will further catalyze the adoption of cash flow underwriting."
Working with multiple data aggregators
Another challenge is the need to work with multiple data aggregators to obtain the customer bank account data from the 13,000 U.S. financial institutions. Plaid, Finicity, Akoya, MX and Envestnet Yodlee are among the companies banks can work with.
MoneyLion started with a single data aggregator and realized over time that "a multi-aggregator approach makes sense for redundancy, for routing, there are data quality differences, even institution by institution," Hong said. "But it doesn't have to be this scary thing because there's now off the shelf multi-aggregator routers that are out there. So more and more of this is being done on your behalf, rather than something you have to build."
Another risk is that consumers will share only positive data and exclude derogatory accounts for lending consideration.
"This is a big danger," Ludwig said. "But it is one of those many things that I think you can work around and that regulators should allow you to work around." Banks may end up charging a bit more or lending less when consumers refuse to provide full access to their data, he said.
