The Consumer Financial Protection Bureau tried to cap overdraft fees at $5. Congress blocked it. Banks lobbied hard. The rule never went into effect.
Before that: years of public pressure, congressional hearings, voluntary commitments from the largest banks to reduce overdraft fees. Chase reduced theirs. Bank of America reduced theirs. Wells Fargo added a grace period. The stories got written. The reform press cycle ran its course.
The average overdraft fee in the United States in 2026: $26.61. Down from $33 a few years ago. Still $26.61.
They haven't fixed it because they don't want to fix it.
The customer who triggers an overdraft is, by definition, the customer with the least margin. The fee is largest relative to income for the people least able to absorb it. They pay it anyway — not because they chose to, but because the alternative is worse. The rent check that bounces has its own consequences. The fee is not a choice. It's a toll on an already-narrow road.
Richard Cantillon observed this dynamic in 1730: financial benefit reaches the center of the system first and the edges last — often as extraction without the corresponding service. The overdraft fee is a small, contemporary version. The bank is positioned between the customer and the money they need. The bank charges for the position.
Overdraft is the retail version. Commercial lending is where the architecture becomes fully visible.
The Mystery of the ISO Channel
Here's something that doesn't get explained plainly.
A small business owner walks into their local branch and gets declined for a loan. They go home. A week later, they're talking to an ISO — an independent sales organization, a commercial finance broker — and the ISO gets them the loan. From a bank. Often the same kind of bank.
That's not a contradiction. It's a channel distinction.
The branch is a retail interface. It's staffed for volume, calibrated for consumer products, and optimized for customers who already have good deposits at that institution. Its underwriting is conservative and its relationship with the back-office lending desk is limited. When the branch declines you, it's not necessarily saying the institution won't lend to you. It's saying this channel won't.
The ISO has a different relationship. They're calling the underwriting desk directly, not the branch. They have a track record with that desk. They know which products the bank is actively trying to place, what credit profile fits the current appetite, and how to package an application to get it through. The borrowers they bring are pre-screened — the ISO doesn't present deals that won't close, because their reputation depends on it.
The result: a borrower who was declined at retail, who is still actually bankable, gets a loan. The ISO made it happen. The ISO also charges a professional services fee for making it happen. Typically around 10% of the total loan amount.
Ten points. On a $200,000 loan, that's $20,000. For a phone call and a relationship.
That's not scandalous — the ISO provided real value. They navigated a channel the borrower couldn't access alone. But it illustrates how the friction in the bank's retail process functions: it doesn't prevent the loan. It creates an intermediary layer that extracts rent from borrowers who need access to the underwriting relationship the branch won't give them directly.
The friction is the product. It generated $20,000.
The Cliff Edge
The ISO channel works for a thin slice of the market — borrowers who are marginal by the branch's standards but still genuinely bankable by any real credit analysis. That's a real population, and ISOs serve them.
Below that line is a different world.
If your paper isn't in that marginal-bankable zone, the ISO can't help you get a bank product. What you get instead is a merchant cash advance. The MCA is structured as a purchase of future receivables, not a loan, which means it's not subject to usury laws and carries no APR disclosure requirement in most states. The effective APRs run from around 40% at the competitive end to well above 100% at the predatory end. The payment is typically a daily or weekly debit directly from your business account, before you decide what else to do with that revenue.
The MCA market exists because there are creditworthy businesses that genuinely need capital and can't access it through any bank channel — ISO-connected or otherwise. Revenue-based lending at a premium is a legitimate product for legitimate circumstances. But the market is also where borrowers in distress end up regardless of whether an MCA is appropriate for their situation, because it's the only channel that said yes.
Grade A paper goes to the bank. Marginal-bankable goes to the ISO, pays 10 points, and gets the bank loan anyway. Below that: the jungle, and you hope you're not eaten.
What Happens to Your Application Data
There's a fourth tier that doesn't get discussed, and it's where the system becomes genuinely hostile.
When a commercial borrower applies to a traditional lender and gets declined, that application data — business name, owner name, revenue figures, contact information — moves. It gets sold. Not always, not by every lender, but often enough that it's a standard feature of the experience. The borrower who needed capital and didn't get it becomes a lead in someone else's database.
The ISO community runs partly on this data. A business that was declined by a bank six months ago is a warm prospect for a broker today. Multiple brokers buy from the same sources. They all call.
Fifty calls a day is not an exaggeration.
The borrower who applied for a loan because they needed to grow their business is now spending two hours a day declining calls from people they've never heard of, all claiming to have been referred. The stress of capital need, compounded by the harassment of the lead economy that feeds on it. This is a known and predictable feature of the commercial lending market — not an accident, not a regulatory gap that nobody noticed. A feature.
Something is being built to address exactly this problem. It doesn't have a public launch yet.
What the Marketplace Does
FundScout routes around the branch channel, the ISO fee layer, and the data harvest.
A marketplace submission is one application that reaches multiple lenders simultaneously — lenders who have structured their operations to serve borrowers the bank routes around, who price for real risk rather than portfolio allocation convenience, and who operate in a competitive environment that keeps that pricing honest.
This is not charity. Alternative lenders charge more than banks when they're taking on more risk, and they should. A responsible borrower should understand the full cost of what they're taking — the commercial finance glossary exists for exactly that. But the borrower deserves to see the full market clearly, not just the channel that's available to them based on which phone calls they happened to receive.
The bank's overdraft fee and the bank's branch decline are the same architecture at different scales: friction as a revenue mechanism rather than a service failure. The ISO fee is a third layer of the same structure. The data harvest is what the structure does to you after it's finished.
The friction is the product.
The marketplace exists because it shouldn't have to be.
Related: Congress Gave Banks Permission to Charge 85,000% APR — the APR on a $5 overdraft held three days. The same architecture, the smallest scale.
Related: Tell Me the Price — the American Factoring Association is taking its fight against MCA disclosure laws federal. The fight is the tell.
Sources
- Why Some Banks Still Charge High Overdraft Fees — New York Times
- CFPB Overdraft Rule — Consumer Financial Protection Bureau
- Overdraft/NSF Fee Revenue Trends — CFPB research
- Small Business Lending Survey — Federal Reserve Small Business Survey