The Rise of Alternative Data in the Lending Market

EXCLUSIVE –  There were conflicting views on the importance of the FICO score at the PayThink Conference in Phoenix this week. While only some professionals were bold enough to say the standardized credit scoring has become marginalized, others maintained that it was here to stay. However, all agreed on the increasing potential of alternative data.

For fintechs and FIs, alternative data is typically gathered through machine learning and artificial intelligence. This data collects a wide spectrum of information, ranging from someone’s job stability to their financial profile based their bank account balance.

According to the myFICO blog, a FICO credit score is calculated using the person’s payment history, length of history, amount owed, new credit and types of credit used.

The problem with the FICO Score is its inherent limitation: a good FICO score is typically 700 and above. The market below this number is diverse and vast.

There’s a lot complexity in the subprime market that the 680 or below score doesn’t acknowledge, Marc Stein, CEO of Underwrite.ai said on the Credit Scoring panel at the conference.

The Boston-based startup applies artificial intelligence and biologically-based machine learning techniques to provide lenders with non-linear, dynamic models of credit risk for their customers (both consumer and small businesses).

Stein’s company focuses on third-party validated data that works with over 2,400 attributes through different criteria. One example is analyzing numbers as a time series: meaning recognizing there’s a difference when,  for instance, someone misses 3 payments in 3 months and someone misses 3 payments in 18 months.

Stein said this alternative data serves not as a score but as a probability.

There are many others like Stein who have companies founded on the use of alternative data. Take the more recent launched of Petal. Founded by Jason Gross, the company released its first product earlier this month: a credit card that doesn’t require a credit score.

The idea behind this product is to cater to the “credit invisible,” according to Gross.

Instead of focusing on credit scores, Petal looks at an applicant’s financial behavior. The three main drivers that shows a customer’s payback behavior are income, stability of income and willingness to pay their expenses. These are the indicators Petal considers.

Being able to determine the likelihood that a person will pay their bill based on certain behavioral patterns, their cash flow and their current bank statement is just as reliable a benchmark as a FICO score, according to these fintech thought leaders.

“Someone’s history of their ability to pay is not the best, and certainly not the only indicator that a person is financially responsible,” another fintech CEO, who requested not to remain unnamed, told Bank Innovation at the conference.

For instance, the fintech ‘darling demography’ that is the millennial or younger generation might not have enough years to establish a good credit score.

Point-of-sale lender Affirm’s Rob Pfeifer would know.

In fact, Pfeifer, Affirm’s chief risk officer, explained at the conference how his company uses alternative data when underwriting a loan to one of its many users.

The San Francisco-based Affirm needs the buyer’s full name, date of birth, email, and mobile number to run a “soft credit check,” and can underwrite a loan within seconds. Affirm brands itself as a loan provider for millennials, but also caters to other sections of the underbanked such as immigrants or students. These are the people that are excluded under FICO.

There are over 65 million people in the U.S. with no or misrepresented credit scores due to factors like age, veteran, or immigration status.

And indeed, the fintech market is exploding with startups catering to the credit invisible. For those companies that do not want to ignore this market, FICO makes little sense, Underwrite.ai’s Stein said.

Mike Armstrong, president of Zest Finance (another platform that uses MI to provide data to consumer lenders), referred to the Consumer Financial Protection Bureau issued “no action letter,” last week as a feat for the alternative data model.

Essentially, he said, the letter signifies that the bureau recognizes the potential for AI and machine learning in accessing the credit invisible.

The CFPB’s letter was issued to San-Francisco fintech Upstart Network last week. Upstart uses alternative data in making credit and pricing decisions. The letter requires the company to  report lending and compliance information on a regular basis to the bureau to mitigate risk to consumers. These reports will also help the Bureau understand the impact of alternative data on the market.

Kathy Boden Holland, the executive vice president of bank products at Elevate looks to the 2008 recession as a major catalyst for the rise of alternative data.

With the recession, many financially responsible people lost jobs and fell behind on things that they ordinarily wouldn’t and this has destroyed their credit score, she said. So, to her, acknowledging the needs of the subprime market is a not just a social matter  but also an important business strategy for banks to embrace.

Based in Texas, Elevate is a publicly traded online lender that makes loans to non-prime Americans.

For someone like Greenlight Financials’ Tim Sheehan, the rise of alternative data is indeed an interesting phenomenon he told Bank Innovation. Sheehan, who co-led Yahoo Finance, launched Greenlight Financial in January. The digital startup is focused on the financial relationship between child and parent.

To him, the phenomenon is only logical in that it acknowledges a diverse set of people, many of whom are financially responsible (despite their FICO score).