Who Will Fill the SMB Lending Gap? One Credit Union Says: We Can
By Matt Doffing, Senior Editor at The Financial Brand
Simple Subscribe
Subscribe Now!
Executive Summary
- Approval rates for small business loans have fallen to just 13% among the largest banks. Many institutions actively shun them, because of challenges in identifying and vetting viable prospects.
- Community banks and credit unions could be uniquely positioned to fill the gap but are hampered by laborious credit approval processes and the haphazard record keeping of many small businesses.
- One California credit union has broken the logjam and identified scores of prospects, by using AI to analyze customers’ financial behavior, rather than examining financial statements and tax returns.
Large nationwide banks long held the upper hand in SMB loans. Their branch networks let them scoop up small-business borrowers in bulk, underwriting loans under $100,000 with credit scores and then pooling them into massive, diversified portfolios that cushioned any losses.
That dominance is now fading. Nationwide branch footprints are shrinking, and the largest banks’ approval rates for small-dollar business loans have fallen to just 13 percent, according to data published by Experian from the U.S. Census Bureau.
At the same time, new business formation is elevated, with 400,000 to 450,000 new applications every month since 2020, Experian reports. That’s compared to about 300,000 a month before the pandemic.
Community banks and credit unions are a natural fit to fill that void, but to do it, they must overcome the same inefficiencies that once diminished their role. One institution, the $5-billion Travis Credit Union, based in Vacaville, Calif., is testing technology to scale SMB lending by identifying opportunities hidden among its consumer members, approving small-dollar lines of credit using its data-driven lending platform.
Cracking the Cost Barrier in SMB Lending
Lenders have long struggled with the economics of small business loans. Borrowers wanting less than $100,000 also tend to be companies that historically have had thin cash reserves and high churn. The underwriting process for that size of loan eats up nearly as much staff time as a multimillion-dollar commercial deal. For community institutions running on tight margins, the math hasn’t really worked.
Gord Baizley, CEO of JUDI.AI, sums up the dilemma: “Typically, what we hear is, we treat a $50,000 loan like a $5 million loan, and it kills the economics, and it kills the experience.” The result has been predictable: Most community lenders either avoided the segment altogether or left it to the biggest banks and fintechs.
At the same time, many small business owners lack an accountant beyond the one they use for tax season, creating a scenario where scaling small business lending requires a new, more efficient approach.
A Different Lens on Credit
That is exactly what Travis Credit Union has set out to tackle. Instead of relying on years of tax returns and reams of financial statements, many of which may be incomplete or outdated, the credit union turned to behavioral data. By working with JUDI.AI, it now evaluates applicants through real-time cash flow.
The tactic goes right after a quirk of microbusiness finance: Owners often minimize taxable income by running personal expenses through the business. On paper, profitability looks meager. But once discretionary spending is accounted for, many of these firms can, in fact, support debt payments.
Baizley describes how JUDI.AI’s model works in practice: “What we learned a long time ago was that collecting and spreading financial statements from small businesses wasn’t very effective. They were usually stale, certainly not audited, and it took forever to get them. So, we migrated to bank statements — now we ingest them digitally, categorize them, and feed the variables into predictive models. That data has proven to be highly predictive.”
During an initial pilot, Travis Credit Union ran a randomized test across its membership. The results: 628 small firms identified and 468 pre-qualified for credit lines of $100,000 or less. The program is still under review, but plans are already underway to extend the model to business credit cards.
The Power of Real-Time Data
In the past, collecting financial statements was hugely inefficient, says Baizley. What made it even more challenging was that they were not audited.
“You’d collect them, print them out, and go through them with a highlighter to see if you found anything interesting,” he says. All that time spent on tax returns from last year when they matter less than a tax installment payment withdrawn last week. “We have the transaction and behavioral data in real time; it’s the same data as is on the business’s account statement.”
By shifting to live cash-flow monitoring, the model does more than improve efficiency. It fundamentally transforms the borrower’s experience. Rather than waiting to learn if they qualify, many small businesses now hear directly from their institution that they already meet the lending criteria.
“Instead of waiting for an application, they can contact small businesses that are already prequalified,” Baizley explains.
Small Businesses Beneath the Surface
Some entrepreneurs manifest another idiosyncrasy that can make them hard to find for community institutions: When people start a new venture, they use consumer checking accounts to process payments and manage expenses because of convenience. Without intervention, it’s not as common for organizations to open formal business accounts without enticement, invisible to their institutions.
To address this, Travis Credit Union also tested anonymized transactional data through the FinGoal Insight Platform. By analyzing anonymized retail account data, the review found that roughly 7 percent of consumer accounts showed signs of business activity, such as repeated rent deposits, rideshare income, or even multiple mortgage payments.
David Nohe, CEO and cofounder of FinGoal, explains how the tool works in practice: “We looked at transactions from the core, cleaning them so we can make sense of the counterparties to transactions. We’re helping the institution interpret the meaning of transactions to identify unmet needs and cross-sell opportunities. We’ve found cases where 75 percent of an institution’s SMB customers were hiding in plain sight within retail accounts.”
SMB Lending Still Open for the Taking
The stakes are high. Travis Credit Union’s own analysis revealed that more than 74 million dollars in deposits were leaving the institution annually from these “hidden” businesses. Without engagement in the near term, it is possible fintech competitors will create relationships for SMBs’ merchant, credit card, and operating account needs. That also puts them in a good place to earn their first working capital loan as well.
The potential pool is substantial for Travis. In its 12-county footprint, there are approximately 281,000 small businesses, 70 percent of which are micro firms with five or fewer employees or less than one million dollars in revenue. For community institutions willing to identify and engage them early, these hidden SMBs represent not just deposits but long-term lending and service opportunities.
“Mortgage lending and auto lending have slowed,” Baizley says. “In many other areas of banking, margins have eroded. Small business lending is one of the last pieces of blue ocean in banking.”
