The usual pitch for crypto payments is simple. No chargebacks, so no lost disputes and no dispute fees. That is true. It is also only half the picture.

Card fees pay for a dispute system that most merchants never see itemized. When you remove chargebacks, you also remove what that system did. Someone still has to handle refunds, screen for fraud and give a nervous customer a reason to trust you. After the switch, that someone is you.

What card fees actually buy

Start with a detail that is easy to miss. Visa says merchants do not pay interchange directly. They pay a merchant discount to their own financial institution. That discount bundles many services into one rate.

Interchange itself moves between banks. One industry analysis describes it as paying for a bundle. To justify higher interchange tiers, issuers must meet obligations such as customer service standards, liability allocation, fraud controls and consumer protections, and they also fund rewards. Online sales cost more because the risk is higher. Card-not-present transactions carry higher interchange rates because the fraud risk is greater.

So the fee is not a pure toll. Part of it pays for the machinery that decides who gets their money back.

The connection most people miss: you already pay twice

US debit cards show this most clearly, because a regulator wrote the fraud cost into the price. When the Federal Reserve set the debit interchange cap, it included an allowance for fraud losses, capped at about the median of reported issuer fraud losses, or 5 basis points. The base amount, meanwhile, covers allowable issuer costs excluding fraud losses. In other words, the fee was built to cover what the issuer loses to fraud.

Now look at who actually loses. The Fed’s 2023 data shows that fraud losses to all parties were 17.6 basis points of transaction value. Then it splits the bill:

• Merchants absorbed 49.9 percent, up from 38.3 percent in 2011.
• Issuers absorbed 28.3 percent, down from 59.8 percent in 2011.
• Cardholders took the rest, rising from under 1.8 percent to 21.8 percent.

Do the arithmetic. Issuers’ 28.3 percent of 17.6 basis points is about 5 basis points. That is roughly the allowance built into the fee. Our calculation, not the Fed’s, but it shows the shape of the deal. The interchange fee covers the issuer’s slice of fraud. The merchant’s slice arrives separately, as chargebacks. For online sales it is bigger still. Merchants absorbed 65.1% of card-not-present fraud losses in the 2023 debit data.

Two limits apply. This is debit only, and only for large issuers. Credit is a different story. No federal law caps credit card interchange. The credit-side split of costs is not published in the same way. Still, the debit data corrects a common assumption. The chargeback system is not free insurance that merchants lose when they leave. Merchants already carry much of its cost.

That changes how to read the trade. You are not swapping a service you fully paid for against a service you now build yourself. You are moving the same work from a shared system to your own team.

What actually moves onto you

Refunds become new payments

On a card, a refund reverses a transaction. On a chain, an agreed refund generally means sending a new payment to a verified destination. The original stays on the ledger. That raises questions the card world answers for you. Law firm Venable says merchants must design bespoke refund workflows, including whether refunds go out in stablecoin or fiat and how timing and valuation will be set.

Mistakes are also permanent. Irreversibility applies to sellers too, so a refund sent to the wrong wallet address is lost for good.

Fraud screening moves in-house

On cards, the issuer looks at each authorization. It checks funds, runs fraud scoring, and approves or declines. On a chain, the sender signs the transfer from their own wallet and nobody scores the order for you. Because payments are generally irreversible once confirmed, disputes shift to a merchant-side refund, review or support workflow. The decision to ship or release access is yours. So is the risk of being wrong about it.

Trust has to be earned in advance

A cardholder can pay a stranger because the bank will step in if it goes wrong. That is the point of rules like Regulation Z for credit and Regulation E for debit. Regulation Z binds the issuer, and the issuer’s investigation is what sends a chargeback to you. Most issuers use the card associations’ chargeback procedures to help a customer in a dispute with a merchant. Without that backstop, your own policy is the only protection the customer has. They will read it before they pay.

Is the trade worth it?

The case for it is real. Disputes are expensive. One compilation of Javelin data says merchants report $46 in third-party fees and $82 in internal costs per chargeback, about $128 in all. One analysis argues that the total cost of card fraud management already exceeds the cost of building voluntary refund processes for crypto payments.

The case against is also real. Venable notes that this approach can increase customer service complexity and create friction compared with well-established card dispute frameworks. Customers who are used to a safety net may hesitate. That is a conversion cost, and it does not show up in a fee comparison.

There is also a quieter loss. Card networks police merchants. Visa cut its merchant threshold for its acquirer monitoring program from 2.2% to 1.5% on 1 April 2026. That is harsh, but it works as an outside alarm. Take the networks away and nobody tells you your dispute rate is rising. You need your own alarm.

One more option sits between the extremes. Circle’s Refund Protocol is a smart contract that holds USDC payments in escrow. The recipient can refund directly, and if the merchant refuses, the customer appeals to a designated arbiter. It has costs too. Adoption challenges include potential abuse by a malicious arbiter, the difficulty of specifying a refund address in custodial wallets upfront, and rising gas costs. It is a tool for higher-value sales, not a default.

A refund and dispute policy to build in its place

Treat this as a working draft. Adjust the numbers to your margins and order sizes.

1. Publish the rules before payment. State whether refunds are available, what asset will be returned, whether partial refunds are possible, and whether the buyer must provide a designated wallet. Show it at checkout, not in a footer.
2. Choose a valuation rule and say it out loud. Prices move between payment and refund. One common approach is to fix the refund in US dollars at the order price. Another is to refund the same coin amount. Each shifts price risk to a different party. Pick one and write it down.
3. Refund to the original wallet only. If the customer lost access, ask for a new address in writing, from the account email. Use the same coin and network the customer paid with. Log the refund transaction hash next to the order.
4. Keep refund funds ready. Maintain adequate stablecoin or fiat liquidity for refunds after the original payment has settled. A refund policy you cannot fund is a promise you will break.
5. Decide screening rules before you need them. Set how many confirmations you require before releasing goods. Tier your checks by order value. Route large or unusual orders to a person. Clear network instructions on the payment page also cut errors. Automated invoices, real-time transaction tracking and a clear refund policy help reduce disputes.
6. Handle wrong-network and underpayment cases explicitly. Say who pays the network fee on a return, and how long you will take to sort out a mistaken transfer. The business decides whether refunds go out in crypto or fiat, which price reference is used, and how network fees are handled. Say it in advance.
7. Set a dispute path with clocks. Acknowledge every claim within a day or two. Give a decision date. Name who reviews it and who hears appeals. For scale, credit card issuers must acknowledge billing disputes within 30 days and resolve them within 90. Aim to beat that by a wide margin. Customers with no chargeback option will judge you on speed.
8. Offer escrow where the stakes justify it. For custom work, large orders or new customers, route payment through an escrow arrangement such as Circle’s protocol or a processor feature, and say so on the order page.
9. Measure what the networks used to measure. Track refund rate, disputes per 100 orders and days to resolve. Set an internal alert well below the 1.5% card-network line. Note that refunds and won disputes do not remove anything from the card counts, so count honestly: every complaint, not only the ones you lose.
10. Train support to explain mechanics. Venable advises checking that customer support teams can explain transaction mechanics, refunds and delays in plain language. Most disputes start as confusion.

What to remember

Dropping chargebacks does not delete the costs of disputes. It changes who pays and how. The Fed’s debit numbers suggest merchants were carrying a large share already, on top of the fees. The work that remains is real, but you can write it down, staff it and measure it. A merchant with a clear policy, a funded refund balance and a fast answer will win trust that no card network supplies for them.

Before you launch, test your own policy. Ask a colleague to buy something, request a refund, then send a deliberately wrong network payment. Where they get stuck is where your policy needs work.

Sources

• Federal Reserve Board: 2023 Interchange Fee Revenue, Covered Issuer Costs, and Fraud Losses (Regulation II)
• Federal Register: Debit Card Interchange Fees and Routing (2011)
• Visa USA Interchange Reimbursement Fees
• Venable: Stablecoins, An Emerging Consideration for Merchants
• Chargebacks911: Crypto Chargebacks
• Crypto Economy: Circle Introduces New Refund Protocol for ERC-20 Payment Disputes
• Upshift: How to accept stablecoin payments
• Paysight: Visa VAMP vs Mastercard ECM in 2026
• Spark: Dispute Resolution Without Chargebacks
• Red Compass Labs: Visa and Mastercard agree to merchant shake-up