There are 32 million small businesses in the United States. Half of the U.S. workforce owns or is employed by a small business. And small business owners create two-thirds of net new jobs annually.
Yet despite their importance to the economy, small business owners are chronically underserved regarding access to capital.
But now more than ever, financial institutions are using new sources of data to better identify and serve more small and medium businesses (SMBs) without increasing risk.
For our panel discussion at Fearless in Fintech 2021, experts from across the financial ecosystem joined Madeline Ross, Enigma’s VP of Marketing, to share their perspective on the state of the SMB economy, how data is used in lending decisions, and why collaboration between banks and fintechs is an advantage to both lenders and borrowers:
Here’s an overview of their discussion.
For our State of SMB Economy Report, Enigma analyzed card revenues from a sample of more than 16 million U.S. businesses to measure the impact of the first 12 months of the pandemic on the SMB economy (starting in March/Q2 2020). We found that many small businesses saw a steep decline in sales over that period, with the lowest point in card revenues occurring in April 2020. Recovery was slow, especially in the restaurant and travel industries.
But there were signs of hope too: a boom in “pandemic entrepreneurs” with thriving new businesses. According to the Peterson Institute for International Economics, businesses launched between March 2020 and March 2021 increased by 23% over the same period between 2019 and 2020. In our sample, the survival rate for businesses launched in 2020 was approximately 46.5% higher than those launched in 2019.
Meanwhile, in small business lending:
The lending landscape is evolving, with new entrants in the space and traditional banks using technology in new ways. According to the Federal Reserve, 42% of businesses that applied for financing in 2020 sought funding from a large bank, up just slightly from 40% in 2019. But 43% of businesses applied to a small bank, up from 36% in 2019.
According to Mills’s book Fintech, Small Business & the American Dream, small business loans amounted to about 20% of banks’ lending in 2017, compared to 30% prior to the 2008 financial crisis.
Small business owners and entrepreneurs need access to a variety of credit sources. Short-term credit matters for the day-to-day management of cash flow, while longer-term credit is essential for capital investments. Yet less than half of small businesses report that their credit needs are met, according to a Federal Reserve survey. The Fed also reports that the share of applicant SMB firms that received all the financing they sought declined from 51% in 2019 to 37% in 2020.
So why aren’t small businesses getting the funding they need?
Mills says there are a number of “frictions and barriers” that hinder small business lending. The most prominent of these are information opacity and heterogeneity.
“It's hard to see inside a small business and know if they're creditworthy, which is what we mean by 'information opacity,'” she explains. Making that determination requires more data points than banks can typically access or have the expertise to evaluate.
The second “friction,” heterogeneity, means that all small businesses are different.
“One day you're lending to a dry cleaner, the next to a funeral home, the next day to a cafe, the next day to a parts supplier,” says Mills. So it’s difficult for a lender to know what the credit profile for a creditworthy dry cleaner would look like.
But if that lender had credit data for 1,000 dry cleaners, they’d have an idea of what makes one risky or promising. It would be fairly simple to tell whether the 1,001st dry cleaner is worthy of lending to. That’s what big data can do — the sheer volume of information, which machines can parse much more quickly than people, can produce insights that reduce these historical frictions.
Data can thus “drive more access and opportunity, particularly for underserved small businesses [like] women-owned businesses and minority-owned businesses, which are most subject to these frictions and barriers,” Mills says, noting potential benefits to the economy including more diverse businesses and more jobs.
There are a number of 'frictions and barriers' that hinder small business lending. The most prominent are information opacity and heterogeneity. —Karen G. Mills, Senior Fellow at Harvard Business School and President, MMP Group
The Paycheck Protection Program (PPP) created a burst of small business lending at a frenzied pace, and even banks experienced in serving small businesses rushed to keep up.
Customers Bank participated in nearly 350,000 PPP loans with a value of $9.6 billion. But the bank’s approach to PPP was very different by the time the program ended in May 2021.
Bell said the Customers team wanted to enable access to PPP loans for as many clients as possible. To do that, the bank needed to compete on a bigger scale — and be nimble enough to do things differently.
That nimbleness requires “pipes of data,” Bell notes. So the bank partnered with fintech companies, including Enigma, to rapidly parse that data and efficiently process applications — enabling even the smallest businesses, like sole proprietors, to get better access to PPP funding.
Data is only useful to financial institutions if the people who work for those institutions can use it.
“It's universally agreed that data has a ton of value, but extracting that value can be quite challenging,” says Kornhauser. “Ensuring that lenders, whether fintech or bank lenders, have the right technology in place to be able to extract the information they need from the data is hugely important.”
That’s why collaboration across a number of “core competencies and expertise areas” — like bank lenders and fintechs — is what’s necessary to solve this problem. In other words, it’s not enough to gather the data, categorize it and put it in context; lenders need the tools to analyze it.
“It's not just figuring out who's creditworthy, it's doing it for the smallest borrowers,” says Mills. “The people who have gotten left out are not just the underserved, but the small-dollar borrowers.”
She notes that about 75% of small business borrowers want a loan of less than $150,000. But loans of less than $100,000 don’t represent significant revenue for most banks. It isn’t economical to assign a banker to a small business owner who seeks a $7,000 loan — the potential profit just isn’t that significant.
But there's a big demand for small-dollar loans, says Mills. And companies like commerce solution Square have found success meeting that demand. Square’s average loan is $6,000, which it is able to do because it uses an automated system to approve or disapprove applicants quickly.
The average PPP loan disbursed by Customers Bank was under $30,000 — which “was an eye-opener for me,” says Bell. “To take on a loan size of $30,000 or less to a mass audience ... and to do it with speed, is quite a task.”
He agrees that the sub-$100,000 level is “a loan size that is often forgotten; sometimes small business owners just need a little bit to get to that next level ... [or] keep the lights on.”
It's not just figuring out who's creditworthy, it's doing it for the smallest borrowers. The people who have gotten left out are not just the underserved, but the small-dollar borrowers. —Karen G. Mills, Senior Fellow at Harvard Business School and President, MMP Group
As technology enables more borrowers to access capital, we hear a lot about “transparency” in lending terms, risk modeling, and data itself.
Mills says “transparency in all forms of lending is crucial, not just for the lenders and borrowers, but also for external parties like regulators that are involved in these markets.”
Kornhauser cautions the industry to remember that defining what we mean by transparency in lending is important. We need to ensure that transparency “goes beyond just folks that are really comfortable with and familiar with data” and aim instead toward democratizing access to and knowledge of how loan approval decisions are made — especially if that loan decision is made by an algorithm instead of a banker.
Mills says it’s also critical for the industry to ensure financial data has integrity.
“This is a place where I've really been delighted to get engaged with Enigma and see the quality of the data they’re bringing to the data aggregation world,” she adds. “Because that data can't just be anything. Data is tricky stuff, and you have to be sure that whatever you think it's representing, it is truthfully representing.”
When the PPP program rolled out, stories about who deserved those loans, who actually got them and those who committed fraud were all over the media. Does technology help or hinder financial institutions in identifying malicious actors?
Ultimately, “trustworthy, accurate, truthful data, like the data Enigma provides, is a huge piece of addressing this problem and doing so proactively,” says Kornhauser, who also thinks we need “human-plus-data ... Especially when it comes to issues of fraud, there are certain things people would be able to sniff out before data alone could tell us.”
We must ensure transparency goes beyond folks that are familiar with data and aim instead toward democratizing access to and knowledge of the loan approval process — especially if it's done by an algorithm instead of a banker. —Laura Kornhauser, CEO & Co-Founder, Stratyfy
What would the experts suggest for lenders looking to become more data-centric?
“Just get started,” says Kornhauser, who thinks many lenders can get overwhelmed by the process of adopting new technology. But even “small, incremental changes can be really meaningful and impactful, and can snowball in all the best ways.”
Bell says that he and his colleagues at Customers Bank know they need to keep iterating and innovating.
“One check is not enough for a lot of these businesses,” he explains. “So how can we start to offer multiple products or services that accommodate those that need it the most? … Thinking into the future has to happen, while we're still standing up some of these environments and data assets.”
Mills thinks lenders should “move faster — because this is a difficult problem, but this is not rocket science.”
According to Mills, technology will enable lenders to “look inside a small business and see whether they're creditworthy,” while the business themselves will benefit from “an accurate forecast of their cash flow.”
While data promises to continue to be “transformative” for businesses’ access to credit, the bigger picture is even more revolutionary, Mills adds.
“We'll have more opportunity in our economy for more people to have the American dream.”