How Small Business Loan Decisions Get Made Now

Automated lending gives small businesses more options, not fewer

There's growing sentiment that automated underwriting is harmful to small business owners applying for a loan. The argument goes like this: community banks are based on relationships, and local economies need a lender who knows the owners, the local business scene, and a business's regional impact – it can't be all about the numbers.

But the argument gets three things wrong. First, it assumes local banks are now incapable of making ad hoc lending decisions. They aren't. Second, it takes a specific advantage of regional banks and generalizes it across the entire financial industry, collapsing several types of loans and underwriting criteria into one impenetrable category. That's an oversimplification. And third, it treats an automated decline as an error the software made, which no published data supports.

For most business owners, automated underwriting has resulted in faster decisions and more places to apply. More types of loans, more lenders, and more diverse underwriting criteria are a democratizing force, not a constraint. Let's get into what critics get wrong about automated lending, and how a business owner should approach it.

The first mistake: Assuming local banks stopped using judgment

The first premise against automated lending is that community banks have stopped underwriting on anything but numbers. The FDIC's Small Business Lending Survey, which drew responses from roughly 1,300 banks and remains the most complete picture of how banks underwrite small business credit, does not support that premise.

The FDIC's summary of the findings says technology has not replaced the staff-intensive, relationship-oriented practices banks use with small firms. Only one in ten banks has a credit-scoring system that can partially or fully automate underwriting for any lending outside credit cards. Fewer than one in thirty use one to auto-approve a loan at all, and fewer than one in a hundred will auto-approve a loan of $250,000. The banks that do fully automate, about 3%, generally keep it to their smallest loans.

A business owner whose loan application would depend on something outside the numbers can still sit down with a community bank and make that case in person. What has changed is that the in-person meeting is no longer the only route to a loan, and the shape of a small business loan is more varied than it used to be.

The second mistake: Treating every loan as the same loan

The second premise holds that automated systems weigh cash reserves too heavily and operating history too lightly. Sometimes that may be true. However, that is an argument for relationship banking, not an argument against lending automation.

A reserve requirement is a criterion for a certain type of loan, and reserves and cash flow are separate tests during underwriting. A business sitting on four months of steady deposits and almost no cushion may fail a conventional term loan's reserve test yet pass a cash-flow underwriter's test. A business that cannot yet show steady revenue may need to make their case through conversation, and a community bank is who should hear that case.

A conventional term loan from a large bank has never been friendly to a young business with a thin balance sheet, and that predates any algorithm by decades. The financing market segmented in the first place because one loan type could not serve every borrower. An SBA loan, inventory financing, working capital financing, and a short-term line of credit each carry their own underwriting criteria, their own terms, and their own intended use. These products are not four versions of the same loan competing on generosity.

Before a business owner concludes that lending is closed to them, they should ensure that they applied for multiple financing options designed for different types of borrowers.

The third mistake: Thinking that automated lending is rigged against you

The third and most common argument against automated underwriting is that the system may miss something material in a submission and decline a borrower who should have been approved. That claim is hard to evaluate, because no published data establishes how often it happens, or whether it happens more often than a human underwriter misreading the same file.

What has measurably changed is recourse. A declined applicant used to have someone to call, and the call sometimes resulted in a conversation that changed the answer of the lending decision. Many applicants no longer have that call. Some of those borrowers would have been declined either way, but the missing conversation still takes something from them: the explanation of what fell short and the knowledge that a person reviewed the file before the answer came back.

What automation offers in exchange is that the lending criteria are knowable in advance. An automated system evaluates a defined set of inputs, and a borrower can find out what those inputs are before an application is submitted: how long the business has been depositing revenue, how the bank statements look across a full seasonal cycle, whether the business credit file exists separately from the owner's personal one.

Knowledge is powerful here. A new owner financing their first business who learns what lenders read, and in what order, carries that knowledge into every financing decision that follows. If you’re a borrower returning to the market after a long stretch away, it is not wrong to find it unfamiliar; SBA applications go through a different process and ask for different data now.

What the research shows: Speed and consistency

Clearing away the false premises does not make the case for automated underwriting. The case in favor of automation rests on two findings, both drawn from large loan-level datasets.

Speed is up, defaults aren’t

A New York Fed study of market-wide mortgage data found that technology-based lenders processed applications about 20% faster than other lenders, controlling for loan, borrower, and location. The more important result: default rates did not rise with the increased speed. Those lenders also kept up when application volume spiked, rather than rationing credit the way capacity limits force traditional lenders to.

According to the FDIC, three in four banks take as long as ten business days on a decision for a typical loan. Slash financing, provided through Slope, returns the majority of credit decisions instantly.¹, ⁵ Over the last 90 days, the median time from application to funds in the account was about a day and a half, and the fastest payout landed in six minutes.

Consistency, with improved fairness

In an NBER study of equity in lending decisions to minority borrowers, one of the takeaways was that rate disparities for Black and Latinx applicants were roughly 40% lower among algorithmic lenders than among lenders deciding face to face, with less disparity in the approve-or-decline decision as well. The authors read that second result as a sign that algorithmic lenders may extend credit to borrowers whom face-to-face lenders turn away.

The approve-or-decline result is the one that matters most to a small business. A lender working from a uniform set of inputs is judging the business on its numbers rather than on whether its owner has a relationship to trade on, and the owners without that relationship are the ones with the most to gain.

Give the newer lenders a look, starting with Slash

A business owner right now may read a few stories about algorithmic lending and decide the new lenders are not worth an application. Yet a business that only ever applies in one place is limiting itself to the criterion of one underwriter. The decline may feel like a verdict on the validity of your business, but it may just be a mismatch with a single lender.

Alternative underwriting can be worth an application. A cash-flow lender looks at weekly deposits and payment history instead of the balance sitting in the account on the day you apply, so a business with steady revenue and a small cushion can qualify where a conventional term loan would turn it down. Instead of waiting a week or more to hear back about your decision, an automated lender could disburse your funds to your business on the same day that you apply.

Slash offers working capital financing through Slope, a short-term line of credit drawn from the Slash dashboard with 30, 60, and 90-day repayment terms. Slope underwrites on cash flow, building its view of a business from transaction-level data rather than from the reserve balance and credit file a conventional term loan weighs most heavily. If your business wants easy access to financing in the same dashboard you use to manage your money, consider Slash.