Every Small Business Loan Has 2 Exits, but Most Banks Measure 1

Every small business loan leaves the portfolio in two ways: the borrower fails, or the borrower pays it off early. Banks measure the first exhaustively. Most do not measure the second.

A default triggers a charge-off and a line in the asset quality report. A voluntary payoff costs no principal, so it never reaches the credit committee. Most core systems record it as “paid in full” and stop there, losing the part worth knowing. A borrower who sold the business has graduated. A borrower refinanced away by a competitor has been taken. One is a credit outcome, the other a growth problem and the ledger shows them identically.

Default counts the borrowers who ran out of options. Voluntary prepayment counts the ones who still had a choice, having refinanced or sold the business. A payoff is closer to a graduation than a loss, and the only routine signal a bank gets about healthy borrowers.

So how does an industry study a number it does not collect?

One Book Reports Both Exits
Loans guaranteed by the U.S. Small Business Administration report both exits at the loan level, going back three decades. Little else in small business lending is reported that way, which makes the guaranteed book the practical place to look.

It also tends to move early. The credit elsewhere test means the program serves, by statute, borrowers the conventional market will not, so the portfolio sits nearest the edge of bankability and moves first when conditions tighten. The same borrowers sit inside conventional books, which are less concentrated. Aggregated data picks it up later: business loan delinquency at commercial banks sat at 1.3% in first-quarter 2026, with a different population.

Through both healthy expansions of the past thirty years, 2004 to 2006 and 2013 to 2019, voluntary prepayment held between 11% and 13% a year while defaults stayed contained. That pairing is what health looks like.

By March 2026, prepayment had fallen to 8.4% and defaults climbed to 4.8%. They have moved in opposite directions only once before in thirty years of data — briefly, during the 2008 financial crisis.

Rates explain part of this. Prepayment tracks refinancing incentives, so some of the gap reflects an environment unfriendly to refinancing, not weaker borrowers. That accounts for the level, not for defaults rising alongside it.

The Survivorship Trap
A rising default rate invites a comfortable explanation: Healthy borrowers refinanced out, the pool shrank and the rate rose for arithmetic reasons, not credit ones.

The payoff data settles that. If survivorship were the cause, the worst vintages would have prepaid fastest. They prepaid slowest. By year three, the 2016 cohort had retired roughly 27% of its loans and was defaulting at 3.1% a year; the 2023 cohort has retired 17% and was defaulting at 8.0%. Every cohort from 2010 through 2021 sits in a tight band, between 2.4% and 3.5% default on payoffs of 23% to 29%. The 2022 and 2023 vintages sit outside it. Failing more and leaving less leaves fuller pools, not hollowed ones.

Two Doors, Two Borrowers
The split runs along collateral lines, because collateral decides whether a borrower can leave at all.

Smaller borrowers without real estate are trapped. Their payoff rates fell as defaults rose. They have nothing to pledge, and small firm credit is tight: the Federal Reserve’s January 2026 Senior Loan Officer Opinion Survey On Bank Lending Practices found banks expecting loan quality to deteriorate for small firms while holding steady for larger ones.

Larger, real estate-secured borrowers are still moving. Their payoffs rose even as defaults climbed.

Bank Boards should ask three questions:

1. 1. What is the bank’s voluntary payoff rate by segment, and which way has it moved? If no one can answer, the reporting is incomplete.
2. 2. Does the core technology capture why a loan paid off, or only that it did?
3. 3. Does asset-liability modeling still apply one portfolio-wide prepayment speed? A blended assumption is wrong in both directions at once.

Asset quality reporting, shaped by decades of Federal Deposit Insurance Corp. call report convention, is built around the involuntary exit. That gap is fixable. Portfolio analytics can score both exits across a bank’s own loans, guaranteed and conventional, without waiting for vintage curves to fill in.

The default rate counts the borrowers who failed. The payoff rate counts the ones who still had a choice, and it moves first. Boards get a precise count of the failures but should also get details on who paid off and why.