Overdraft, NSF, and Late Fees

Six times the price of the sandwich

A $35 overdraft fee on a $6 purchase means you paid nearly six times the price of the thing you bought — for the privilege of being short for a day or two. Stated that way, it sounds absurd, and it should: no one would accept "the sandwich costs $41 if your timing is off" as a posted price. But that is exactly the transaction, and it happens tens of millions of times a year — mostly to the same people, which is the real story of this chapter.

Penalty fees are the bluntest instrument in the extraction economy. There's no financial engineering here, no fine print requiring a law degree — just a fee, triggered by a timing failure, priced at whatever the institution decided. What makes them worth a full chapter isn't their mechanics; it's their distribution, which is one of the most precisely documented facts in consumer finance, and their fixability, which is nearly total.

The concentration

Overdraft and NSF (non-sufficient funds) fees ran about $12 billion in the U.S. in 2024. The striking fact isn't the total — it's how concentrated it is: CFPB research found that a small minority of accounts, roughly 9%, generate nearly 80% of all overdraft fees.¹ This is not a cost that everyone pays a little of. It's a cost that a few pay enormously, repeatedly, because the underlying condition — cash timing running close to the edge — doesn't fix itself between incidents. The same account that overdrafted in March overdrafts in May, and August, and the fees stack into hundreds of dollars a year for exactly the households with the least room to pay them.

And the distribution has a shape that upends the intuitive story. In the Federal Reserve's own 2024 survey, the group paying overdraft fees most often isn't the very poorest — who are more likely to be unbanked entirely, facing a different set of costs — but working households earning $25,000–$50,000 a year, at 19% annually, more than three times the 6% rate above $100,000.² This is a thin-margin problem, not strictly a poverty problem: households with income, obligations, and no slack, where every month is a timing exercise conducted without instruments.

Race and disability track separately and compound on top: Black adults pay overdraft fees at more than double the rate of white adults (21% versus 9%), and disabled adults at roughly double the rate of adults without a disability — gaps the Fed's own analysis states persist even after accounting for income.³ Single-parent households are more than five times as likely to be unbanked entirely as married-with-children households, which pushes them toward cash-and-check-cashing costs that run higher still. None of this means the fees are targeted — it means thin margins and thinner safety nets compound wherever they already exist, and they don't exist evenly. The premium lands where the cushion is thinnest; that's not a metaphor, it's the data.

Late fees, and the sting behind the sting

Credit card late fees run the same playbook with a nastier second act. The fee itself — up to about $40 under current rules (the federal safe harbor allows $30 for a first miss, $41 for a repeat; a 2024 attempt to cap fees at $8 was struck down in court in 2025)⁴ — is the visible cost, and it's the smallest part of the damage.

Behind it: a late payment can trigger a penalty APR near 30% on the balance going forward — a repricing of your entire debt, not a one-time charge. And at 30 days late, the miss becomes a mark on your credit report, where payment history is the heaviest factor in every scoring formula (Your Credit Report and Score) — a single mark that can cost 60–100+ points and raise the price of every future loan for years.⁵ One missed Tuesday, three concentric circles of cost: the fee, the rate, the record. The industry has softened somewhat — a few large banks have eliminated overdraft fees outright and others cut them to a fraction of the old price — but banks that still charge them typically bill $25–35 per incident, and the fees still land overwhelmingly on the least cushioned.⁶

Timing failures, not character failures

Here's the reframe this chapter exists to deliver: these fees are timing failures, not character failures — and timing is forecastable.

Nobody overdrafts because they're bad. They overdraft because rent left on the 1st, the card autopay hit on the 3rd, and the paycheck lands on the 5th — a sequencing problem (Timing Is Everything is devoted to it) that no amount of virtue solves and a simple forward view dissolves. The household paying $300 a year in overdraft fees isn't spending more than the household paying zero; it's seeing less. The fee is a tax on operating blind.

Run the numbers with the income-context rule. Four overdrafts a year at $35 — still the going rate at some of the biggest banks — is $140, and for a household in that $25,000–$50,000 band, four is a conservative count, not a worst case; the frequent-overdrafter tail runs to hundreds per year. Against a high income, $140 is a rounding error and this chapter is preventive maintenance. Against $40,000 with no slack, the overdraft habit plus one late-fee-and-penalty-APR incident can consume a full percent of gross income — which, recall from The Quarter-Trillion Fleecing, is the entire interest-and-fee budget of a financially healthy household, spent on nothing at all.

What visibility does here

This is the extraction category where visibility pays fastest, because the fix is purely informational. Plenee projects your daily cash position ahead — every scheduled obligation, every expected inflow, laid on a calendar — so a shortfall shows up before it happens, while it's still a Tuesday problem (move a payment, shift a date) rather than a $35 problem. Due-date tracking and autopay planning keep late fees structurally off the table: not "remember better," but "arrange the system so remembering isn't load-bearing" (Late Fee Elimination builds the full architecture). The 9% of accounts paying 80% of the fees aren't different people; they're people without instruments. The instruments exist.