Why overdraft fees persist

For more than two decades, U.S. consumer-protection writing has predicted the end of the overdraft fee. The CARD Act was supposed to end it; the 2010 Reg E opt-in was supposed to end it; the rise of online-first banks was supposed to end it; the CFPB's 2024 rulemaking was supposed to end it. Each time, the fee proved more durable than its critics expected. The reason is not that regulators have failed or that competition is weak. The reason is that the fee fits the architecture of U.S. retail-deposit banking in a way no substitute has come close to matching, and the industry has structured its other product offerings to depend on it. To understand why the fee persists is to understand a basic fact about U.S. retail banking: it is sold as a free service to a majority of customers and subsidized by fees on a minority who, the data show, are mostly the customers least able to pay them.

The argument here is in four parts. First, the fee is structurally large because the U.S. consumer-deposit business is structurally low-margin and the cross-subsidy it provides is structurally important. Second, the customer base that pays the fee is sticky in ways that make rate competition and product alternatives ineffective at moving them. Third, regulatory restrictions on the fee have repeatedly produced bank-side adaptation rather than fee elimination. Fourth, the recent fee reductions at large banks are real but represent a partial accommodation rather than a structural change, and the smaller institutions where the legacy economics persist will continue to operate them.

The cross-subsidy thesis

U.S. retail banks compete primarily on convenience, brand, and product breadth. The deposit base they accumulate from this competition pays minimal interest at branch-heavy institutions — often a few basis points against a federal funds rate that has spent much of the past two decades in the 0-5% range. The bank's net interest margin on the deposit base funds operating expenses, capital, and the long tail of consumer-facing services (branches, mobile platforms, customer service, dispute handling, free ATMs, free checking) that the depositor experiences as "the bank."

The economic problem is that the per-customer revenue from a typical branch-checking customer is small. A customer maintaining a $3,000 average balance in a non-interest-bearing checking account at a 4% deposit-rate-spread produces about $120 per year of net interest income for the bank. Add ten debit-card transactions per month at the regulated debit interchange rate of about 21 cents plus 0.05%, and the bank earns another $30 to $50 per year in interchange. Total per-customer revenue: somewhere between $150 and $170 per year, against an industry estimate of $300 to $400 per year in cost-to-serve a typical checking customer when branch network, technology, fraud protection, and compliance overhead are fully loaded.

The math does not work — at least not for the median checking customer. It works because of dispersion: a substantial minority of customers produces revenue well above the average, and a smaller minority produces revenue well above that. Wealth-management customers fund a substantial part of the business through advisory fees. Loan customers (mortgages, HELOCs, credit cards carrying balances) fund another substantial part. And overdraft customers — historically perhaps 10% to 20% of checking-account holders incurring the bulk of all overdraft fees in any given quarter — fund a third substantial part.

The CFPB's research on overdraft revenue, drawn from bank call reports and from voluntary data submissions, has consistently documented that a small minority of checking-account customers pay the majority of overdraft fees. The 2017 CFPB study (still the most-cited reference for the share concentration) reported that 9% of consumers paid 80% of overdraft fees in the studied period; subsequent studies have shown similar concentration patterns even after fee reductions. The structure has been stable across regulatory regimes.

Stickiness and selection

If overdraft customers paid materially more per year for their bank than non-overdraft customers, and if they could readily move to an institution that did not charge the fee, market competition should drive the fee toward the cost of providing the underlying short-term credit. The reason this has not happened is that overdraft customers are systematically sticky and systematically not the customer base that the fee-free alternatives are designed to attract.

Stickiness comes from several sources. First, switching banks is operationally burdensome — redirecting direct deposit, updating autopay, navigating a new bank's interface. The CFPB has documented switching rates for U.S. checking accounts that are low even when consumers express dissatisfaction. Second, customers who repeatedly overdraft are disproportionately customers with thin or impaired credit-bureau and deposit-bureau histories; opening a new account at a competing institution may be harder for them than for the average consumer (see ChexSystems and being denied an account). Third, the financially marginal customer-base — the one that most needs the short-term-credit function the overdraft fee economically charges for — also has the least time and bandwidth for sophisticated banking-relationship optimization.

Selection compounds the stickiness. Online-first banks and fintech apps that explicitly do not charge overdraft fees have grown rapidly over the past decade, but their customer acquisition has been heavily skewed toward higher-income, lower-overdraft-risk consumers who would not have been profitable for the original institution anyway. The frequently-overdrafting customer is harder to acquire, harder to underwrite, harder to keep within bounds, and harder to monetize through the alternative revenue streams (interchange, wealth management, lending products) that the new entrants depend on. The CFPB's 2024 overdraft rule, which would have substantially restricted overdraft fees at banks with more than $10 billion in assets, was structurally an effort to fix this by regulation — to force the institutions that have the overdraft-fee-paying customer base to give up the revenue regardless of competitive dynamics.

Regulation produces adaptation, not elimination

The history of overdraft regulation since the 1990s is a history of restrictions producing bank adaptation rather than fee elimination. The 2010 Reg E opt-in rule prohibited overdraft fees on one-time debit-card and ATM transactions without consumer opt-in; the bank-side response was a substantial opt-in marketing push that produced opt-in rates of roughly 20% of customers and 40% of frequent overdrafters, preserving the bulk of the revenue while shifting it to a smaller, more concentrated base. The CARD Act prohibited various credit-card fee practices and produced bank-side adaptation toward higher upfront APRs and richer rewards programs that produced comparable economics differently.

The 2024 CFPB rule (still subject to litigation as of this writing) proposed three options for covered banks: stop charging overdraft fees, cap the fee at a $5 benchmark, or treat overdrafts as TILA-covered credit with full APR disclosure. The bank-side response, anticipated by industry observers, was likely to follow the same pattern: many large banks would pre-emptively eliminate the fee on entry-level accounts (already happening before the rule), some would adopt the benchmark cap, and others would introduce explicit small-dollar credit products that price the credit transparently rather than through a per-item fee. The aggregate revenue collected from the overdraft-vulnerable customer base would not vanish; it would migrate.

This is not a criticism of regulation but a recognition of structural reality. The economic function of the overdraft fee — generating revenue from a defined customer segment to subsidize free service to a broader one — does not disappear because the rule changes; it finds another implementation. Regulators have moved the fee around, made it smaller, made it more transparent, and reduced its incidence; the function has been more resistant.

The recent reductions and what they actually represent

The past five years have seen substantial overdraft-fee reductions at most large U.S. banks: NSF fees largely eliminated, per-item overdraft fees reduced from around $35 to around $10 or $0, grace periods introduced, de-minimis thresholds added, and in some cases the fee eliminated entirely on flagship checking products. The aggregate dollar reduction in overdraft revenue has been substantial — CFPB research has documented declines of more than 50% from the pre-2022 peak at the largest banks.

These reductions are real and have produced real consumer benefit. They are also, structurally, a partial accommodation rather than a fundamental change. The largest banks have reduced the fee while retaining the basic product (checking accounts paying near-zero interest with broad branch and ATM access). The smaller and mid-size institutions where the legacy fee structure persists continue to operate it; CFPB research has shown that the dispersion in overdraft pricing across institutions has substantially widened since 2022, with large banks moving aggressively and many smaller banks not. The aggregate decline in industry overdraft revenue is largely an artifact of the largest institutions' decisions.

For the consumer most affected by the fee — the frequent-overdrafter at a small or mid-size bank — the reduction at the largest institutions is largely irrelevant unless they switch banks. The friction to switching, as discussed above, is part of why this customer typically does not.

The counter-argument

The strongest counter-argument to the structural thesis is that the recent reductions show that overdraft fees were not as load-bearing as the structural argument implies, and that further reductions are likely to continue under competitive and regulatory pressure. Bank profitability at the largest U.S. institutions has held up despite the overdraft-fee reductions, and several major institutions have eliminated the fee entirely on flagship accounts while remaining substantially profitable on those customers. This suggests that the cross-subsidy can be replaced by other revenue lines (interchange growth, wealth-management cross-sell, lending-product attach rates) and that the overdraft fee is not as essential to the U.S. retail-banking business model as the structural argument claims.

There is force to this. The Bank of America transition to a $10 overdraft fee with substantial program restrictions, the Capital One elimination of overdraft fees on consumer accounts, and similar moves at JPMorgan and Wells Fargo have not produced visible profitability degradation at any of them. If the largest banks can run profitable consumer-banking franchises without the overdraft fee, the rest of the industry presumably can as well.

The structural argument's response is that the largest banks operate at a different scale and with different customer-base composition than the mid-size and community-bank tier where the legacy fee structure is most concentrated. The big banks have wealth-management, mortgage-origination, capital-markets, and other revenue streams that smaller institutions do not. A community bank without overdraft revenue and without those other lines is a different financial proposition. The reductions at the top of the industry may not generalize to the rest of it; the dispersion in current overdraft pricing across institutions is consistent with this.

What this implies

If the structural thesis is right, three things follow.

First, overdraft-fee reductions at the largest banks should continue but at a slowing pace, with the residual fee on those institutions stabilizing rather than approaching zero. The 2022-2024 reductions captured most of the easy gains; further reductions will face more resistance from the underlying economics.

Second, the fee will remain meaningful at smaller and mid-size institutions, with regulatory pressure (the now-litigated CFPB rule, potential state-level action) the most likely catalyst for change. Competitive pressure at this tier of the industry is weaker than at the large-bank tier because the customer overlap with high-yield-savings and online-first alternatives is smaller.

Third, the substantive consumer-protection question for the most affected customers — the frequent-overdrafter at a small-to-mid-size bank — is increasingly about whether the explicit small-dollar credit products that some institutions are introducing as overdraft-fee replacements are actually a better product for those customers, or merely a re-pricing of the same economic transaction under a more sympathetic label. The early evidence is mixed; the structural challenge of profitably serving the population most prone to overdraft remains unresolved.

The overdraft fee has been declared dying for thirty years and has been substantially reduced multiple times. Each time, the structural function it served has reasserted itself in a different form. The 2022-2024 reductions are the most material to date, but they do not represent the end of the cross-subsidy; they represent its migration to a smaller, more concentrated, less visible base. The fee persists because the system it is part of persists; reforming it more deeply requires reforming more of that system than the political coalition has, so far, been able to assemble.