The making of a movement

View This Article in BOSS Magazine

Transforming the credit union industry is just the beginning for VyStar

Driven to improve their members’ financial lives and with a mandate to push the credit union industry forward, VyStar Credit Union is blazing a new path in lending. By their nature, credit unions are not-for-profit financial institutions that are expected to tend to the needs of their members. However, VyStar has honed that baseline vision into a movement to strengthen the credit union industry as a whole. They’re pairing game-changing technology with a goal of making credit union lending more accessible, inclusive, and fair.

Established at Naval Air Station Jacksonville in 1952 as a credit union for military and civil service members and their families, today’s VyStar is a regional powerhouse serving more than 775,000 members throughout Florida and Georgia. With assets topping $13 billion, VyStar is one of the 15 largest credit unions in the country, offering a full range of financial services from deposit accounts to mortgages, auto and consumer loans, credit cards, and commercial and small business lending.

While they have the scale and heft to partner with marquee fintechs used by large banks, VyStar has instead launched two separate but integrated investment funds to support cutting-edge fintechs primarily serving the credit union industry. Their technology focus is trained on products that truly make a difference for all credit unions, and that will become part of the fit-for-purpose ecosystem they are working to build.

“We want to strengthen the entire credit union movement by funding fintechs that are not just going out to banks but that are coming to support our industry,” said Jenny Vipperman, VyStar’s Chief Lending Officer. “We look for fintech partners that provide a solution to problems faced in our industry in a unique and valuable way.”

Fund recipients include Zest AI, an artificial intelligence and machine learning company that creates more accurate and inclusive loan decisioning models, Payveris, creators of an open-API, cloud-based digital payments platform, and banking-as-a-service developer Nymbus, selected by VyStar as their digital lending provider. VyStar is also a member of the CURQL fund, co-created by Joel Swanson and several credit union peers. CURQL is a consortium of 55 credit unions that are pooling their resources to fund fintechs with great ideas and great execution that need funding to grow as they continue to create innovative credit union-specific products.

“The credit union movement is a cooperative with a shared mission: to serve the underserved. In fact, credit unions serve more rural, low-income, and minority areas than banks do,” Vipperman said. “It’s critical that inclusive services that improve the financial lives of every American become available to the credit union industry because we are the ones that serve those populations and geographies.” These investments represent a major step toward that goal.

Decisions, decisions

Rethinking lending criteria is a central part of VyStar’s quest to serve borrowers long neglected by both banks and credit unions. To that end, they leverage a pair of partners specializing in predictive analytics for financial institutions: Zest AI helps lenders develop fair, consistent, accurate, and compliant loan decision models, and Open Lending provides advanced loan decisions and loss insurance for non-prime auto lending. The implementation of these two fintechs allows VyStar to be a full spectrum lender with prime lender risk.

VyStar further partners with 2020 Analytics, which provides predictive modeling and economic scenario analysis, to additionally evaluate VyStar’s portfolio through a Fair Lending lens to identify opportunities to implement more inclusive lending practices for its members.

The goal, Vipperman said, “is to collaborate to create a more inclusive financial services environment where all consumers have access to products and services that meet their needs, providing dignity and positioning consumers for long-term financial wellness and success.”

Early on, Vipperman decided that the results of their efforts would be completely shareable, not only with VyStar members, but with the entire community of credit unions. “For a lot of credit unions there's more on the line if they get it wrong, if they go out and do artificial intelligence and machine learning and it doesn’t pan out,” she said. “That’s something that can hold them back from moving forward.”

That lack of opacity necessitated strong results worth sharing, and VyStar delivered. Vipperman shared all their results as a way to help other credit unions make decisions about the technology with confidence. Their openness also showcased how nimble credit unions can be when making tech moves. “I think that the credit union industry can move faster than the banking industry, which is bigger and takes a lot more to implement and is not as willing to share results amongst themselves because they are truly competing with one another,” she said.

VyStar’s credit scoring models are both regional and product-based, enabling a higher degree of accuracy in contrast with the “one-size-fits-all” scoring methodology used by FICO, which contains inherent bias based on age, race, and income, and whose decisions are based on a mere 10 to 25 factors captured at a single moment in time.

“Our models truly calculate the risk of a loan based on our lending footprint, for each individual product, taking into account the structure of the loan and other loan-specific criteria. This allows us to better serve our communities and members by saying yes to more people, being more inclusive in our lending, and more accurately assessing risk to ensure safety and soundness alongside fairness and growth,” Vipperman shared.

Zest AI’s processes relate credit scoring to game theory and the core principle that every time a factor changes, it changes the outcome of the entire game, so players must look at all of those factors together to make the best decisions.

While AI models can contain, and even increase bias, the Vystar model actually reduces bias by collecting hundreds of factors to create a score and back testing each factor over a year’s worth of applications to ensure accuracy. They also examine the results for consumers in protected and unprotected classes.

“In every single case and in every single model, we are finding that we are increasing approvals for those consumers that are in protected classes at higher levels than consumers in unprotected classes,” she said. “That does not mean we're making fewer loans to people in unprotected classes; in fact, we're increasing those loans as well.”

Increasing loans to the protected classes at a higher rate closes the human bias gap to make the models fairer and more consistent, delivering an accurate representation of risk. AI’s precision impacts the approval process as well, making it possible to decline borrowers that were initially approved for a loan but were actually found to be riskier than initially determined, and vice-versa.

Auto loan applications are run through Zest AI and then through Open Lending — a process that takes only seconds to complete — and gives VyStar an opportunity to closely review the loans. If there’s a subset of members whose risk is slightly higher than the credit union’s tolerance but can be insured, potential declines can be converted to approvals. “It's really kind of a cascade to bring in as many people as possible for the best rate and approval while keeping our organization safe and sound,” she said.

Increasing the accuracy and inclusiveness of their lending decisions is only part of VyStar’s formula to create a new movement within the industry. “Over time, over generations, we have created a scenario in which people who try to apply expect to be declined and therefore do not apply,” said Vipperman, who is deeply concerned about the painful experience of being declined. “It feels like your character, you as a person are being declined. That is a terrible feeling, and I want to remove that feeling entirely from everyone's lives in terms of credit. This is about dignity and respect.”

A recent partnership with CuneXus will give VyStar members an instant, reliable grasp on their buying power via direct offers and later as a presentment through the mobile app and online banking. Because members’ loans are pre-approved, the solution eliminates the need for an application or stressful visits to the branch to determine creditworthiness. “Members can see what their balance is and what they can afford for that week or that month, immediately see what they qualify for, what the rate is, what the payment is, and they can accept it and move on,” Vipperman said. “And instead of them saying, ‘Please see me as worthy of a loan,’ we get to come to them and say, ‘Please allow Vystar to earn your business.’”

As we went to press, a member with a credit card with an APR over 20% stopped into a VyStar branch, unsure if he would be eligible to consolidate that debt. “He was going to try for $5,000 of credit to see if we could approve it. Our Zest model gave him a $27,000 limit and he was able to consolidate debt at 0% for 12 months. That changes his life.”

He returned to VyStar to refinance his mortgage and family cars, and brought along his son, who thought he couldn’t get a credit card but was approved for a $7,000 rewards card. Vipperman added, “We are making a real impact on the average American struggling to pay off credit card debt. We are making a real impact in people’s lives, and we are opening up their relationship with us, and that's just one from today.” Results like this are powerful enough to reshape an industry, and VyStar is doing just that.