Every Lender in a Stacked Position Underwrote It Correctly

The position that defaults is usually one no single underwriter ever saw, because it does not exist in any one file.

A merchant cash advance repays through a holdback, a share of daily card receipts or bank deposits taken before the merchant sees them. Typical holdbacks run 10 to 20 percent, most commonly around 15. On a single advance that is a workable structure, which is the entire reason the product exists.

Take a second advance while the first is outstanding and the arithmetic changes shape. Two or more advances routinely consume 30 to 40 percent of daily receipts, and in the worst cases more than half of gross deposits are gone before the merchant pays a supplier or makes payroll.

The coverage collapses faster than the burden grows

The burden itself rises in a straight line. What matters to the merchant does not.

Set a merchant’s operating margin against the cumulative holdback and you get a coverage ratio, how many times over the margin covers the daily financing draw. At one advance taking 15 percent against a 15 percent margin, coverage is roughly one: the financing is being paid out of the margin with nothing spare. At two advances, coverage halves. At three it falls to a third.

Coverage falling below one is not a warning sign. It is a description of a merchant who has begun funding repayments out of capital, because there is no longer enough margin to cover them. Everything after that point is a countdown whose length depends only on how much cash they started with.

That is arithmetic rather than a survey finding, and it holds whatever numbers a specific merchant has. Substitute a 22 percent margin and a 12 percent holdback and the shape is identical, only the crossing point moves.

What this costs a lender

Industry estimates put merchant cash advance default or non-completion at roughly 15 to 20 percent, against 1 to 2 percent for SBA loans and 3 to 7 percent for conventional bank lending. That spread is understood and priced.

The relevant figure is what stacking does to it. Borrowers carrying multiple stacked advances are estimated to default at three to five times the rate of single-advance borrowers. On a book already running at 15 to 20 percent, a segment defaulting at three to five times that rate does not need to be large to dominate the loss line.

Run it forward on a book of 400 funded positions. If one in six ends up stacked, roughly 67 positions carry a default probability several times the portfolio assumption. Those positions were priced as though they were the other 333. The pricing error is not in the credit box; it is in a fact that arrived after funding and that nothing in the process was watching for.

Why correct underwriting produces this outcome

Here is the part worth sitting with. In a stacked position, every funder involved usually underwrote correctly.

The first advance was assessed against clean receipts and performed as expected. The second funder saw the same receipts, possibly slightly stronger, and had no reliable way to see the first position. Bank statements show the holdback as an outflow, but a debit to a processor looks much like any other recurring debit, and a merchant who wants the second advance has every incentive not to volunteer it.

So the failure is not analytical. Each decision was defensible on the information available. The position only exists in the space between three companies’ files, and none of them owns that space.

That reframes what a lender can actually do about it. Sharpening the credit model does not help, because the model was not wrong. What changes the outcome is whether anything looks again after funding, and how quickly.

How CXO Solves This

Custom Agentic Workflow makes the check continuous rather than a moment. The firm’s own history and any consortium or bureau data it subscribes to become queryable at the point of decision, and open positions get re-examined on a defined schedule against the merchant’s ongoing deposit behaviour. A new recurring debit appearing three weeks after funding, at a size and cadence consistent with a second advance, is a detectable event. It is only undetectable if nobody is looking.

Client Onboarding Automation is where the first half of that has to live, because a check that runs outside the application flow becomes a separate queue and gets skipped under volume.

Neither of these stops a merchant taking a second advance. They change when the funder finds out, from the first missed remittance to the week it happens, and that interval is the difference between a restructure conversation and a workout.

In most operations, far more work can be automated than leadership realizes. One discovery call is enough to size what automating it would return to your bottom line. Book it at https://cxocorporation.com/contact.