To lend or not to lend, that is the question.
With the increase in online lending, financial service companies face the daunting task of determining which consumers will repay their loans and, more importantly, which will not.
Traditionally, the risk was determined based solely on credit history. However, ValidiFI, a Florida-based company, provides lending companies an alternative means of evaluating risk, in part, through predictive modeling expertise from the data scientists of South Dakota State University’s Department of Mathematics and Statistics.
“We help financial service companies make better decisions when providing a financial product to a consumer,” explained ValidiFI Chief Operating Officer Jesse Berger. “We are a few years into it and already working with some of the largest personal lenders in the nation.”
Using vast amounts of complex data to detect patterns can help lenders determine which applicants will repay loans—and it’s one of the strengths of the SDSU data science program, explained associate professor Tom Brandenburger.
When the two met in 2016, Berger was interested in recruiting data scientists to work in Florida. Brandenburger suggested opening an office in the Research Park at South Dakota State University—and that’s what Berger did.
As ValidiFI’s chief data scientist, Brandenburger mentors data scientists who are also SDSU alumni at the Brookings office, which opened last year.
Providing unique decision-making data
Data science has dramatically improved financial technology and is vital to providing access to more consumers. As financial services migrate to the digital realm, the possibility for risk is greater and more sophisticated.
“Many consumers want to do their banking online. In the virtual world, we no longer sit across the table from a client. Consequently, there are more opportunities for fraud,” said Berger, who described himself as a serial entrepreneur. During the last 25 years, he has started multiple companies that provide data or technology services.
“What’s unique is the type of data we provide for the credit decision-making and underwriting,” explained Berger, who started Merchant Boost about four years ago. The firm was recently renamed ValidiFI. “We spent years building out the technology and went to market about two years ago.”
“We are using alternative data to verify that people who are taking out loans are who they say they are and that they are a decent credit risk,” Brandenburger said. “It’s data that credit bureaus do not have—and it’s helping people demonstrate they are good payers and weeding out the bad players who are purposefully trying to defraud the lender.”
For instance, he explained, “If the name, Social Security number and birthdate do not match the data provided by the applicant, our Bank Account Validation product concludes that person is not likely to repay a personal loan.” The same is true, for example, if a single address is listed on 37 applications for credit in the last seven days or if that address has 72 different people associated with it during the last 60 days.
The BAV product also “confirms ownership of bank accounts and debit cards to reduce or even eliminate some of the fraud that is happening,” Berger said.
Though Brandenburger describes predictive modeling as “a fancy version of math that established likelihood out of a large amount of data,” he also admits the difficulty lies in the complexity and the size of the data.
“Tom and his group in Brookings use machine learning and artificial intelligence to build different types of scores to help make sense of the data,” Berger said.
The Payment Instrument Risk Score, for instance, analyzes a variety of financial datasets to generate a risk score that places the applicant in a low, medium or high credit risk category. “We send the results in real time within milliseconds, giving the financial institution the score and details about the person based on the data we have,” Berger noted.
CEO Oscar DiVeroli, center, visits with the Brookings ValidiFI staff. Pictured are, from left, Basanta Chalise, Nikhil Pamidimukkala; Diveroli, associate professor Tom Brandenburger and Audrey Bunge.
Minimizing overdraft charges
ValidiFI’s Bank Aggregation technology monitors a borrower’s bank account. By explicitly authorizing this type of oversight, the borrower not only secures a loan, likely at a more reasonable rate, but also receives protection against a bounced or late payment—and in the end, a better credit rating.
However, many people don’t realize that bounced payments hurt both the lender and the borrower. “When you bounce a payment, you get an overdraft—but so do lending companies that use ACH (automated clearing house) payments,” Brandenburger explained.
“We can alert the lender ahead of time if the bank account balance is running low and if there is a high probability the client will not make the next payment,” Berger continued. The parties can then work out arrangements to avoid an overdraft. “This helps both the borrower and the lender.”
This is particularly beneficial for a client with little to no credit history—and something that young people find worthwhile, Brandenburger said. “It’s a generational thing—young people are happy to give to get. We are creating an environment with a safety net for everyone involved that reduces the risk associated with these monetary transactions,”
Through a partnership with First Bank & Trust of Brookings, ValidiFI will be offering a new service line for credit card and ACH payments.
“Essentially, we will provide a data solution to verify identity, detect fraud and determine credit-worthiness based on payment instruments like credit cards and bank accounts,” Berger said. This new service line builds on the company’s core payment data instrument.
The payments will be processed through ValidiFI’s partnership with First Bank & Trust. “It’s exciting to partner with such an innovative company, as ValidiFI’s products and proven results are a real game-changer in the lending industry,” said Cal DeJong, the bank’s President of National Products.
In the future, Berger hopes to expand into the mortgage, automobile and home loans market.