Machine Learning Times
Machine Learning Times
Why You Must Twist Your Data Scientist’s Arm To Estimate AI’s Value
 Originally published in Forbes, June 11, 2024. If you’ve...
3 Ways Predictive AI Delivers More Value Than Generative AI
 Originally published in Forbes, March 4, 2024. Which kind...
AI Success Depends On How You Choose This One Number
 Originally published in Forbes, March 25, 2024. To do...
Elon Musk Predicts Artificial General Intelligence In 2 Years. Here’s Why That’s Hype
 Originally published in Forbes, April 10, 2024 When OpenAI’s...

9 years ago
Banking Start-Ups Adopt New Tools for Lending


SAN FRANCISCO — When bankers of the future decide whether to make a loan, they may look to see if potential customers use only capital letters when filling out forms, or at the amount of time they spend online reading terms and conditions — and not so much at credit history.

These signals about behavior — picked up by sophisticated software that can scan thousands of pieces of data about online and offline lives — are the focus of a handful of start-ups that are creating new models of lending.

No single signal is definitive, but each is a piece in a mosaic, a predictive picture, compiled by collecting an array of information from diverse sources, including household buying habits, bill-paying records and social network connections. It amounts to a digital-age spin on the most basic principle of banking: Know your customer.

“We’re building the consumer bank of the future,” said Louis Beryl, chief executive of Earnest, one of the new lenders.

And in that bank, whether a customer uses proper capitalization and spends time reading terms and conditions of a loan may make him or her more creditworthy.

Yet the technology is so new that the potential is unproved. Also, applying the modern techniques of data science to consumer lending raises questions, especially for regulators who enforce anti-discrimination laws.

None of the new start-ups are consumer banks in the full-service sense of taking deposits. Instead, they are focused on transforming the economics of underwriting and the experience of consumer borrowing — and hope to make more loans available at lower cost for millions of Americans.

Earnest uses the new tools to make personal loans. Affirm, another start-up, offers alternatives to credit cards for online purchases. And another, ZestFinance, has focused on the relative niche market of payday loans.

They all envision consumer finance fueled by abundant information and clever software — the tools of data science, or big data — as opposed to the traditional math of creditworthiness, which relies mainly on a person’s credit history.

The new technology, proponents say, can open the door to far more accurate assessments of creditworthiness. Better risk analysis, they say, will broaden the lending market and reduce the cost of borrowing.

“The potential is there to save millions of people billions of dollars,” said Rajeev V. Date, a venture investor and former banker, who also was deputy director of the Consumer Financial Protection Bureau.

Investors certainly see the potential; money and talent are flowing into this emerging market. Major banks, credit card companies and Internet giants are watching the upstarts and studying their techniques — and watching for the perils.

By law, lenders cannot discriminate against loan applicants on the basis of race, religion, national origin, sex, marital status, age or the receipt of public assistance. Big-data lending, though, relies on software algorithms largely working on their own and learning as they go.

The danger is that with so much data and so much complexity, an automated system is in control. The software could end up discriminating against certain racial or ethnic groups without being programmed to do so.

Even enthusiasts acknowledge that pitfall. “A decision is made about you, and you have no idea why it was done,” Mr. Date said. “That is disquieting.”

The data scientists focus on finding reliable correlations in the data rather than trying to determine why, for instance, proper capitalization may be a hint of creditworthiness.

“It is important to maintain the discipline of not trying to explain too much,” said Max Levchin, chief executive of Affirm. Adding human assumptions, he noted, could introduce bias into the data analysis.

Regulators are waiting to see how the new technology performs. The Consumer Financial Protection Bureau wants to encourage innovation but is monitoring the emerging market closely, said Patrice A. Ficklin, head of its fair lending office.

The data-driven lending start-ups see opportunity. As many as 70 million Americans either have no credit score or a slender paper trail of credit history that depresses their score, according to estimates from the National Consumer Reporting Association, a trade organization. Two groups that typically have thin credit files are immigrants and recent college graduates.

Affirm’s office in San Francisco looks nothing like a bank, occupying a couple of floors in an old red brick building. The work space is open with high ceilings, bare wood floors and rows of benchlike tables, where workers are hunched over computers.

The start-up began its credit card alternative for online purchases in July, but it is growing fast and has ambitious plans.

Affirm says it is on track to lend $100 million during its first 12 months. More than 100 online merchants are now using its installment loan product, Buy With Affirm. Next up, the company says, will be student loans.

These are the first steps in a larger plan. “The long game is to use data and software to chew up and revolutionize the financial ecosystem,” said Mr. Levchin, co-founder of PayPal, the leading Internet payment service.

Mr. Beryl of Earnest got turned down for a loan to pay for education expenses when he was getting both an M.B.A. and a public policy degree at Harvard. By then, Mr. Beryl, who majored in financial engineering at Princeton, had worked for a few years on Wall Street. As a graduate student, he was adding to a résumé that screamed earning potential, investing in himself.

The lesson he took from the loan rejection was that traditional banks take a narrow view of loan applicants, and that loans are too hard to get and too expensive for many Americans.

Earnest was founded in 2013, and began lending last year. In 2014, its loans reached $8 million, growing gradually. By December the month-to-month growth rate was 70 percent, Mr. Beryl said. The typical Earnest loan is for a few thousand dollars, though they can range up to $30,000. Many of the loans are for relocation expenses and for professional training.

So far, Earnest’s borrowers are mainly college graduates, ages 22 to 34. The youth focus, Mr. Beryl said, also reflects the best business opportunity. “The most mispriced group in the loan market is financially responsible young people,” he said.

Early customers of the new data lenders speak of the speed and simplicity of the borrowing experience, as well as low rates. They are often young adults who are comfortable with buying online and sharing information.

Ananta Pandey, 22, used a loan from Affirm in August to buy an $850 mattress from Casper, an online retailer. Ms. Pandey graduated last year from Barnard College, where she majored in computer science, and she now works as a software engineer for a start-up in New York. The mattress was for a move into an apartment that she shares with three roommates.

Ms. Pandey has only one credit card that she got not long ago, and it has a low limit of about $1,000. The setup costs for the apartment were a temporary spike in her expenses, and the Affirm loan, Ms. Pandey said, gave her financial flexibility.

Ms. Pandey also appreciated the “very seamless” Affirm loan process. Affirm asks borrowers for their cellphone number, their name, date of birth and the last four digits of their Social Security number. Affirm then asks the applicant to reply to a text message, to authenticate the person’s identity.

Affirm makes the underwriting decision almost immediately. The entire process is generally completed in two minutes or less.

By: Steve Lohr
Originally published at

Leave a Reply