Going further down the credit spectrum to find safe loan growth
With 2021 more than halfway behind us, planning for next year — and beyond — is well underway. Banks and credit unions have been racing to develop growth strategies to stay one step ahead of the surge in deposits and they're seeing mixed results.
During the last 12 months, for example, credit unions’ surplus funds rose 31 percent while assets grew only 13 percent. The closely watched loan-to-asset ratios have ticked up from their depths in April, but are at a mere 60 percent for the credit union industry at large.
Turbocharge your portfolio growth
One of the strategies to strengthen your loan growth we identified is to change the way lenders assess creditworthiness in a post-Covid world. Consumers are increasingly vocal about not wanting to be defined by a three-digit score.
In fact, 62 percent of Americans wish there was another way to prove themselves to lenders outside the standard credit score. Seven out of 10 Americans would switch to a financial institution that has more inclusive lending practices, according to a 2020 Zest AI / Harris Poll Consumer credit survey.
The nation’s biggest banks have already announced plans to augment credit reports and scores with real-time income or cash-flow data.
Lenders who have switched to ML underwriting typically get 15-20 percent higher approvals and a big jump in inclusion.
Unlocking potential by serving the underserved
This evolution will continue and is setting the stage for the next trillion-dollar growth opportunity in financial services. This growth comes with better serving the underbanked — a problem that credit unions are very well suited to solve given their mission to enrich the lives of their members and communities. At a recent credit union executive roundtable, Mike Dill, EVP and Chief Lending Officer at Royal Credit Union, said, “This is a big issue for us and we’re kicking around a lot of different technologies and strategies.”
The central challenge? Conservative lenders are wary of the risk of near-prime paper and people with thin credit histories. It doesn’t have to be this way.
A lack of credit history doesn’t necessarily make someone riskier than someone with a robust file. It just makes them harder to score using traditional underwriting, which depends heavily on a dozen or so factors such as credit score, income, and current debt outstanding. Limiting the factors ignores a good deal of information that can greatly impact a lender’s decision to approve a loan and unfairly penalizes millions of Americans.
Lenders don’t need to abandon this opportunity. Machine learning (ML) models can ingest hundreds or thousands of variables — all of which come from the credit bureaus you’re already in business with — to generate a truer risk-ranking of borrowers. With AI, underwriters can use trended data and credit-adjacent data from checking accounts, rental history, and utility bills to supplement borrower profiles. Technology has caught up to meet the moment.
Doing the math on AI-automated lending
The increased predictive power yields real economic gains. Lenders we’ve worked with that switched to ML underwriting typically generate 15-20 percent higher approvals and with that comes a jump in inclusion: More thin-file, underbanked, and protected-status applicants get approved — all with little to no impact on total portfolio risk.
Jenny Vipperman, Chief Lending Officer at Vystar Credit Union, used a Zest-built credit card model to increase approvals by 22 percent among her members: “That is thousands of people who otherwise would not have had access to a credit card.” Approvals went up even more for women and people of color. With ML models, the path to serving the underbanked is realistic and no longer a catch-22.
Download a copy of “Do The Math,” the Zest Guide to loan growth.
