Chargeback Reason Codes Explained: Visa and Mastercard

Last updated: September 2026

TL;DR

A reason code is not a description of what happened. It is the card network's classification of what somebody asserted, and that determines which evidence rebuts it and how long you have to produce it. Read one at a time, codes are administrative. Read as a distribution across a book, they tell you where your controls are failing and at which point in the flow — which is a different and more useful question.

What are chargeback reason codes?

A chargeback reason code is the code a card network assigns to a dispute to classify the basis on which it was raised. Every network maintains its own set, and the same underlying dispute carries a different code depending on which network processed the transaction.

The distinction that matters is between the code and the reality. The code records the claim — what the cardholder told their issuer, or what the issuer concluded — not a finding about what actually occurred. A cardholder who does not recognise a legitimate charge and a cardholder whose card was genuinely stolen may generate the same code.

Two things follow, and they are the practical content of this page. The code determines what evidence is capable of rebutting the dispute. And it starts a clock whose length depends on the network and the code, not on the dispute.

Who is who in a dispute

Five parties, and each one uses the reason code differently.

• Cardholder — raises the dispute with their issuer. Does not choose the code.

• Issuer — the cardholder's bank. Assesses the claim, selects the reason code, and initiates the chargeback.

• Acquirer — the merchant's bank. Receives the chargeback and passes it on.

• Merchant — supplies the evidence to rebut it, if there is any.

• Card network — defines the codes, the evidence rules, and the deadlines.

This page is written from the issuer and risk-team side of that list. If you are selecting codes and reading them in aggregate rather than fighting individual disputes, the code mix is a diagnostic about your own book — and that is the angle most published material on reason codes does not take.

The four families every network's codes fall into

Underneath the different numbering schemes, the networks agree on the shape of the problem. Codes group into four families, and the family tells you more than the number does.

That final column is the reason to organise the taxonomy this way. Fraud codes and consumer-dispute codes are usually owned by different teams and fixed by different work, and a book weighted toward one calls for a completely different response from a book weighted toward the other.

How Visa numbers its dispute conditions

Visa organises disputes under Visa Claims Resolution into four numbered families that map directly onto the four above: 10 for fraud, 11 for authorisation, 12 for processing errors, and 13 for consumer disputes.

The most frequently looked-up condition is 10.4, Other Fraud — Card-Absent Environment, which is the card-not-present fraud code and the one that dominates e-commerce dispute volume. The fraud family also covers the EMV liability-shift conditions (10.1 counterfeit, 10.2 non-counterfeit), 10.3 for card-present fraud, and 10.5 for disputes arising through Visa's fraud monitoring programme.

The processing-error family includes conditions such as 12.1 Late Presentment, 12.5 Incorrect Amount and 12.6 Duplicate Processing. The consumer-dispute family covers 13.1 Merchandise or Services Not Received, 13.2 Cancelled Recurring Transaction, 13.6 Credit Not Processed and others.

A caveat that belongs in the copy rather than in a footnote: the Visa condition names and windows above are drawn from an acquirer-published reason-code guide dated 2019 and last revised 2021. The four-family structure is stable and current. Individual condition names and day counts should be checked against Visa's own current published rules before you rely on them operationally, and that is true of any third-party table, this one included.

How Mastercard numbers its chargebacks

Mastercard's Dual Message System chargebacks fall into four groups that map onto the same four families, per Mastercard's own Chargeback Guide (Merchant Edition, 13 May 2025):

• Authorisation-related Chargeback — 4808

• Cardholder Dispute Chargeback — 4853, plus 4850 (Installment Billing Dispute — Participating Countries) and 4854 (Cardholder Dispute — Not Elsewhere Classified — United States Domestic)

• Fraud-related Chargebacks — exactly four codes: 4837 (No Cardholder Authorization), 4870 (Chip Liability Shift), 4871 (Chip Liability Shift — Lost/Stolen/Never Received Issue (NRI) Fraud), and 4849 (Questionable Merchant Activity)

• Point-of-Interaction Error — 4834

4837, No Cardholder Authorization, is the Mastercard code most people arrive looking for. It is the rough counterpart to Visa's 10.4, though the two are not interchangeable.

The two-digit codes are not obsolete

Mastercard's guide prints its codes in dual notation — 4808/08, 4853/53, 4837/37, 4834/34 — and this trips up a lot of secondhand documentation.

The four-digit code is the Dual Message System code. The two-digit code is its Single Message System counterpart. They are parallel schemes for different message systems, not a current code and a retired one. If you have seen both and assumed the shorter form was legacy, that is a reasonable inference and it is wrong.

Two related distinctions worth keeping straight. 4804, 4809 and 4811 are Mastercard-Initiated Disputes, where the network rather than the cardholder initiates — a group most published reason-code lists omit entirely. And 4901, 4902 and 4903 are not chargeback reason codes at all; they support second presentment and arbitration, and seeing them listed alongside chargeback codes is a reliable sign a table was assembled without reference to the source.

Discover and American Express each maintain their own numbering as well. Both are outside the scope of the sources used here, so this page names them rather than guessing at their codes — check each network's own published rules.

Time limits: the deadline is set by the code, not by the dispute

The single most common question about chargeback deadlines contains a false premise, so it is worth correcting directly.

120 calendar days is the Visa window, not the Mastercard one. Mastercard's standard Dual Message System chargeback window is 90 calendar days from the Settlement Date or Central Site Business Date, as applicable. A team carrying one number in its head is wrong about one of the two networks it processes.

And the window is set per code, measured from different starting points. Depending on the condition, the clock may run from the transaction date, the processing date, the expected delivery date, or the date a credit was received. Two disputes filed on the same day can have genuinely different deadlines. Regional rules compound it further — Mastercard's guide records Mainland China domestic transactions at between 5 and 90 calendar days, and Costa Rica domestic at 10.

This is why a single generic deadline table is not merely imprecise but misleading. The window is something you derive from the network and the code together; it is not a property of the chargeback.

As above, the Visa figures come from a 2019/2021 acquirer guide and the Mastercard figures from the 13 May 2025 edition of Mastercard's own guide. Both are dated deliberately.

What the code mix tells you that no single code can

Here is where reason codes stop being administrative.

A book weighted toward fraud codes and a book weighted toward consumer-dispute codes are two different problems, with two different owners, and neither is solved by getting better at disputes. The first points at authorisation-time decisioning. The second points at the merchant, the product, or the customer experience. Reading codes one at a time cannot tell you which you have.

One consumer fintech described a pattern that makes the point concretely. Chargeback abuse was running across two accounts: funds were disbursed to one bank account the user controlled, while payments arrived from a second, and the chargebacks were lost on the second. Taken individually, those disputes looked like ordinary consumer disputes and were coded accordingly. Taken as a pattern across linked accounts, they were first-party fraud.

That is the general shape of it: the code is a label attached to a single event, and the fraud is only visible as a distribution across linked entities. In practice reason codes sit alongside bust-out patterns, excessive chargebacks and returns, and velocity anomalies as the typologies a fraud model is built to separate — which means the code is an input, not a conclusion.

There is a related blind spot worth knowing about. In research sponsored by Mastercard and conducted by Javelin Strategy & Research — a survey of 100 chargeback executives at US card issuers in September 2025 — 20% of financial institutions reported that they do not track how many transactions each accountholder disputes. Among those that do not, 67% said it was because they did not want to inconvenience customers with the follow-up. If the pattern is the signal, a fifth of institutions in that sample are not positioned to see it. The same research found 83% of issuers attributing rising dispute volume to growing customer awareness of the dispute process.

Responding to a code

The family determines what evidence is capable of working. A fraud code is rebutted by evidence the cardholder participated; a consumer-dispute code is rebutted by evidence of delivery or of what was promised. Misclassifying the family wastes the entire response, because the evidence assembled will not be responsive to the claim actually made.

From the issuer and risk-team side, three things are within your control: the authorisation decision that created the exposure, the signals feeding that decision, and whether the disposition feeds back into policy.

The honest limitation is worth stating. By the time a reason code arrives, the loss on that transaction has already happened — a chargeback is better treated as a trigger for a downstream action than as an evaluation in its own right. Its value is almost entirely in what it changes about the next transaction, which is an argument for treating the code stream as a feedback loop into transaction fraud decisioning rather than as a queue to be worked down.

Where to start

Pull last quarter's chargebacks and group them by family before doing anything else. Not by code — by family. The distribution usually answers the question of where the problem actually is, and it takes an afternoon.

Then check whether you can attribute disputes to accountholders across linked entities. If you cannot, that is the gap worth closing first, because it is the one that hides first-party fraud inside ordinary-looking dispute volume. Where those linked accounts turn out to have been opened by the same entity in the first place, the problem started earlier than the dispute — that is onboarding fraud, and it is cheaper to catch there than to litigate a chargeback later.

Reason codes are one input into a broader dispute and chargeback operation, and the decisioning side of that — how the code stream changes the next authorisation — is where the leverage sits.

Frequently asked questions

What are chargeback reason codes?

A chargeback reason code is the code a card network assigns to a dispute to classify the basis on which it was raised. It records what the cardholder or issuer asserted rather than a finding about what occurred, and it determines both the evidence capable of rebutting the dispute and the deadline for producing it.

Why do chargeback reason codes matter for financial institutions?

Individually they are administrative, but in aggregate the distribution of codes across a book is a diagnostic. A concentration of fraud codes points at authorisation-time decisioning; a concentration of consumer-dispute codes points at the merchant or the product. The two call for different work from different owners.

How are Visa and Mastercard reason codes structured differently?

Visa uses four numbered families under Visa Claims Resolution — 10 fraud, 11 authorisation, 12 processing errors, 13 consumer disputes. Mastercard uses a separate four-digit scheme with four Dual Message System groups that map onto the same families, each four-digit code shown alongside its two-digit Single Message System counterpart. The families correspond; the numbers do not translate.

What is the difference between a fraud reason code and a consumer dispute code?

A fraud code asserts that the cardholder did not authorise the transaction. A consumer-dispute code asserts that they authorised it but did not receive what was promised. They are rebutted by different evidence and prevented by controls at different points — authorisation decisioning for the first, merchant and product quality for the second.

How long do you have to file a chargeback?

It depends on the network and the specific code. Visa's dominant window is 120 calendar days; Mastercard's standard Dual Message System window is 90 calendar days from the Settlement Date or Central Site Business Date. Individual codes are measured from different starting points, and some regions differ, so the window has to be derived from the network and code together.

Where can teams find the primary card-network source for reason codes?

Each network publishes its own rules, and those are the only authoritative source. Mastercard publishes a Chargeback Guide, Merchant Edition; Visa's conditions are defined in its core rules and Visa Claims Resolution documentation; acquirers also distribute consolidated guides to merchants. Any third-party table, including this one, is a summary of those documents and should be checked against them.