Who Really Pays a Fee: Incidence in Consumer Finance
Who Really Pays a Fee: Incidence in Consumer Finance
The law decides who a fee is charged to. The market decides who pays it, and the answer is whichever side is less able to walk away. A late fee cap that removes $28.80 a year per account, recovered through interest, becomes 0.90 percentage points on every revolving balance. The people who paid late pay less. The people who paid on time and carried a balance pay more. The cost moved; it did not disappear.
What you'll learn
Who a fee is charged to is not who pays it
Economists distinguish between the party a charge is levied on and the party that bears it. The distinction is called incidence, and its central result is simple: the burden of a charge falls on whichever side of the market is less able to walk away, regardless of which side the law names.
The share borne by customers is roughly the supply side's responsiveness divided by the sum of both sides' responsiveness. Put three stylized markets through it.
| Market | Customers bear | Providers bear |
|---|---|---|
| Customers have few alternatives, providers can redeploy easily | 89% | 11% |
| Both sides equally responsive | 50% | 50% |
| Customers switch readily, providers are locked in | 14% | 86% |
The legal label on the charge does not appear anywhere in that calculation. A fee levied on a merchant and a fee levied on a cardholder end up in the same place if the elasticities are the same. Per our cross-subsidy analysis, this is why debates about who should pay for financial services so often miss: they argue about the label and the market decides the incidence.
What a fee cap actually does
Caps are the cleanest test, because they remove a charge by rule and then the market redistributes it. Take a card late fee capped from $32 to $8, the kind of cap the CFPB adopted in 2024 and a court vacated in 2025. With accounts averaging about 1.2 late payments a year, the issuer loses $28.80 per account annually.
- If the issuer recovers that revenue through interest, the APR on an average $3,200 revolving balance has to rise by about 0.90 percentage points.
- The late fee was paid by the accounts that paid late. The APR increase is paid by every account that revolves, late or not.
- So the cap does not remove the cost. It moves it from the minority who triggered it to the larger group who carry a balance.
Whether that redistribution is good is a separate question, and it has a real answer on each side. Late fees fall disproportionately on households in strain, and spreading the cost across all revolvers is arguably fairer. But a late fee is also a price on a behaviour the borrower controls, and moving it into APR removes the signal. Per our price discrimination analysis, the question is which characteristic the price is allowed to track. Incidence does not settle that. It says only that the cost does not disappear.
Four cases where the label misleads
| Charge | Levied on | Where it tends to land |
|---|---|---|
| Interchange | Merchants | Per our interchange analysis, largely in retail prices, paid by every customer including those paying cash |
| Interchange caps on debit | Issuers | Per our debit routing analysis, it was widely reported that free checking became less common afterward, moving cost to account holders |
| Card surcharges | Card-paying customers | Per our surcharging analysis, split with the merchant depending on how many customers walk away |
| Interest rate caps | Lenders | Per our rationing analysis, on the borrowers above the cap, who lose access rather than pay more |
The last row is the extreme case of incidence. When a price cannot rise to cover cost, the burden falls on quantity rather than price, and the people who bear it are the ones who are no longer served. Per our small-dollar lending analysis, a 36% cap is borne almost entirely by borrowers whose expected loss puts them above it, and it shows up as declines rather than as a line on anyone's statement.
The first row is the one with the widest reach. A cost that the merchant folds into its prices is paid by customers who never used a card, which per our rewards analysis is part of why card rewards are partly funded by cash payers.
Who is least able to walk away
If burden follows inability to leave, the practical question in any market is who is stuck. In consumer finance, a handful of features reliably identify them.
- Switching is costly. Per our exit analysis, existing customers on long-term products are captive in a way new customers are not.
- Alternatives are absent. A borrower with a thin file has fewer lenders to move to, so per our credit invisibility analysis any cost added to the product they can get lands on them.
- The charge is not salient. A fee nobody notices is a fee nobody responds to, and a market that does not respond bears the charge in full.
- The need is immediate. Per our present bias analysis, a household choosing under time pressure has, in effect, no demand elasticity at that moment.
The four features cluster on the same households, which is why policies that shift charges around a market tend to land hardest on the same group whichever way they are aimed. Per our overdraft analysis, the households paying overdraft fees are also the ones with the fewest alternatives to the account that charges them.
The case that incidence analysis is used to block every reform
Incidence arguments are frequently deployed by providers against any change: cap our fees and you will pay through higher rates; cut our revenue and free products will disappear. Some of those warnings are correct and some are not, and the honest response is that incidence is an empirical question with an answer that can be estimated. Where providers face real competition and have little room to raise other prices, a cap can be absorbed in margin. Where they do not, it will be passed through. The argument to reject is the one that treats pass-through as automatic and total. Per our rate transmission analysis, even a change everyone expects to pass through fully, a policy rate, passes through unevenly and with lags that depend on competition.
The same logic, used deliberately
Incidence is usually discussed as a constraint on regulators. It is equally a tool for anyone setting a price, because it says where a charge will end up before it is introduced.
- A cost a business cannot avoid will be paid by its least mobile customers unless the business decides otherwise. Per our exit analysis, that is the default outcome and the loyalty penalty is its visible form.
- A cost removed from one line will reappear in another unless margin absorbs it. A provider that cuts a fee should decide where the revenue comes from rather than let the most captive segment supply it by default.
- A charge customers can see and respond to is shared with the provider; one they cannot see is paid by them in full. Making a charge visible changes its incidence even if its size stays the same.
The third point is the one with an ethical edge. A provider that chooses opacity is choosing who pays, and the choice falls on the customers least able to notice. Per our price discrimination analysis, that is the second-row pricing that borrows legitimacy from risk-based pricing without being tied to risk. A provider that chooses visibility accepts bearing more of the cost itself, which is a competitive disadvantage only until customers can compare.
Testable implications
- After a fee cap, the affected providers should raise other prices most in the segments with the lowest switching and least in the segments where customers are most mobile.
- Pass-through of an interchange change to retail prices should be faster and more complete in concentrated retail markets than in competitive ones.
- A rate cap should reduce the number of loans in the band just above the cap far more than it reduces the average rate on loans that remain.
- Charges that are less visible to customers should show higher pass-through to customers than equivalent visible charges.
The fourth implication is the one with the most practical consequence for anyone designing a fee. Per our complexity analysis, an opaque charge does not just avoid complaint; it changes who pays it, because the side that cannot see a charge cannot respond to it, and per our denominator analysis the way a cost is reported can hide it from the people bearing it.
Frequently asked questions
The difference between who a fee is charged to and who bears it. The burden lands on whichever side is less able to walk away.
The cost moves. Recovering $28.80 per account through APR takes about 0.90 points, shifting cost from late payers to all revolvers.
Merchants are charged it, but it largely lands in retail prices paid by all customers, including those paying cash.
No. It depends on competition and alternatives. Under a rate cap the burden falls on access rather than price.
Key takeaways
- The burden of a fee falls on whichever side of the market is less able to walk away, not on the party named.
- Where customers have few alternatives, they bear most of a charge; where they switch readily, providers do.
- A late fee cap removing $28.80 per account becomes about 0.90 APR points if recovered through interest.
- That moves cost from late payers to every revolver, which is a redistribution rather than a saving.
- Interchange is charged to merchants and lands largely in retail prices paid by all customers.
- Under a rate cap the burden falls on access: borrowers above the cap are declined rather than charged more.
- High switching cost, few alternatives, low salience and immediate need cluster on the same households.
- Pass-through is an empirical question, never automatic and total, and opaque charges pass through most.
This analysis is educational and does not constitute financial, legal, or policy advice. All figures are stylized illustrations using assumed elasticities, fee levels, late payment frequencies, and balances, and do not describe any particular provider or market. The regulatory status of fee and rate caps varies by jurisdiction and changes over time. Consult a qualified adviser before relying on any framework described here.