paymentswithpaul_

July 19, 2026 · 8 min read

Chargeback Ratio Math: How VAMP and ECM Thresholds Actually Work

How Visa's VAMP and Mastercard's ECM calculate your chargeback ratio, why timing mismatches inflate it, and what actually happens at each threshold.

Most merchants who get flagged for excessive chargebacks are surprised by it. Not because they ignored their disputes, but because they were tracking the wrong number. Visa and Mastercard each calculate your ratio differently, on different timelines, with different inputs. If you are managing to a single blended "chargeback rate" in your dashboard, you are not managing the numbers the networks actually enforce. Here is the real math, as of mid-2026.

The two ratios that matter

There is no single chargeback ratio. Visa runs the Visa Acquirer Monitoring Program (VAMP). Mastercard runs the Excessive Chargeback Merchant program (ECM) and its escalation tier, High Excessive Chargeback Merchant (HECM). Each network only sees its own card volume, so a merchant can be safely under one network's threshold while blowing through the other's. You need to compute both, separately, every month.

Visa's VAMP: fraud and disputes in one number

In 2025 Visa retired its two legacy programs, the Visa Dispute Monitoring Program (VDMP) and the Visa Fraud Monitoring Program (VFMP), and consolidated them into VAMP. The consequential change is that fraud and non-fraud disputes are now combined into one ratio.

The VAMP ratio is calculated as:

(TC40 fraud reports + TC15 non-fraud disputes) / settled card-not-present transactions

Three things to understand about the inputs:

  • TC40s are fraud reports filed by issuers. They count against you even when the cardholder never files a formal chargeback.
  • TC15s are non-fraud disputes, the classic chargebacks driven by Visa reason codes like 13.1 (merchandise not received) or 13.7 (cancelled services).
  • The denominator is settled CNP transaction count, not dollar volume.

VAMP thresholds as of mid-2026

Threshold Ratio Applies to
Merchant excessive 1.5% (was 2.2% until April 1, 2026) Merchants in US, Canada, EU, APAC, LATAM
Merchant excessive (CEMEA) 2.2% CEMEA region merchants
Acquirer above standard 0.5% Acquirer portfolio
Acquirer excessive 0.7% Acquirer portfolio

The threshold drop from 2.2% to 1.5% on April 1, 2026 roughly cut the allowable dispute rate by a third overnight. A minimum activity floor applies: as of June 2025, the ratio is only enforced once you have 1,500 or more combined fraud and dispute transactions in the month. Small merchants are below the radar; a merchant doing real volume is not. VAMP also tracks a separate enumeration ratio for card-testing attacks, which is a different problem with a different fix.

Note the acquirer thresholds, because they explain processor behavior. Your acquirer gets measured at 0.5% across its entire portfolio. That is why processors terminate merchants well before the merchant threshold: you are consuming their headroom. This dynamic is most of the reason processors shut down high-risk accounts that were technically still compliant.

Mastercard's ECM: counts and basis points

Mastercard still runs a chargeback-only program and it uses a two-part test. You must trip both the count and the ratio to be flagged.

  • ECM: 100 to 299 chargebacks in a month AND a chargeback-to-transaction ratio of 1.5% to 2.99% (150 to 299 basis points).
  • HECM: 300 or more chargebacks in a month AND a ratio of 3% or higher (300+ basis points).

Mastercard expresses the ratio in basis points: chargebacks received in the current month divided by transactions in the preceding month, times 10,000. That denominator choice matters, and we will come back to it.

The two-part test is why low-volume merchants rarely hit ECM even with ugly ratios, and why high-volume merchants can hit it with ratios that look modest. At 50,000 transactions a month, 150 basis points is 750 chargebacks. You will trip the count threshold long before then.

Count versus dollars

Both networks count transactions, not dollars. A $9 disputed trial subscription hurts your ratio exactly as much as a $900 disputed order. This has two practical consequences.

First, low-ticket, high-frequency billing models (trials, micro-subscriptions) generate ratio risk out of proportion to their revenue. This is a structural problem in verticals like nutraceuticals and supplements, where rebill models are standard.

Second, refunding is cheap insurance relative to disputes. Losing $60 of revenue on a preemptive refund is trivial next to one more unit in a numerator that gets you terminated.

The timing mismatch trap

Here is the failure mode that catches otherwise healthy businesses.

Chargebacks lag sales. A cardholder typically disputes 30 to 60 days after the transaction, sometimes as long as 120 days under network rules. But the ratio is computed on a monthly window: this month's disputes over this month's (or, for Mastercard, last month's) sales.

The numerator reflects decisions you made one or two months ago. The denominator reflects what you sold recently.

So when sales shrink, your ratio spikes even if nothing about your dispute rate changed. Run the numbers. You process 10,000 transactions in March and generate disputes at a true 0.9% rate, which is 90 disputes that mostly arrive in April and May. Then April sales fall to 4,000 transactions because you paused an ad account, hit seasonality, or lost a traffic source. April's measured ratio is roughly 90 / 4,000 = 2.25%. You just breached VAMP with a true dispute rate under 1%.

The corollary is worse: scaling down after a bad month is exactly the wrong move. The disputes from the bad month are already in flight. Cutting volume shrinks the denominator right as the numerator peaks. Merchants who panic-pause their stores routinely convert a marginal month into a program flag.

Manage to forward-looking exposure: estimated disputes from trailing 60 days of sales, divided by projected current-month transaction count. That is the number that predicts a breach before it happens.

What happens at each threshold

The consequences escalate in a predictable sequence.

  1. Fees and fines. Merchants flagged under VAMP are assessed $8 per fraudulent or disputed transaction, as of the current schedule. Mastercard ECM fines start around $1,000 per month and escalate with consecutive months in the program, reaching six figures for merchants stuck in HECM, plus an issuer recovery assessment of $5 per chargeback above 300 in a month.
  2. Remediation plans. Your acquirer must submit a plan to the network explaining how you will get the ratio down. Expect demands for alerts coverage, refund policy changes, descriptor fixes, and sometimes volume caps.
  3. MID termination. If the plan does not work, or the acquirer decides you are not worth the portfolio risk, they terminate. Most terminations happen here, at the acquirer's discretion, not at any published network line.
  4. MATCH listing. Terminated for excessive chargebacks usually means a MATCH list entry (reason code 04), visible to every acquirer for five years. That is the outcome that follows you. Getting a domestic merchant account with a MATCH entry is somewhere between difficult and impossible, which is how merchants end up needing offshore acquiring on bad terms.

Exiting the programs requires sustained performance, typically three consecutive months below thresholds for Mastercard.

How alerts and refunds interact with the ratio

Prevention alerts (Ethoca on the Mastercard side, Verifi CDRN and RDR on the Visa side) let you refund a transaction before it becomes a formal chargeback. They are useful, but the accounting is more nuanced than vendors imply.

  • For Mastercard, an Ethoca alert you refund in time generally never becomes a chargeback, so it stays out of your ECM count.
  • For Visa, RDR and CDRN resolutions keep the TC15 dispute out of your VAMP numerator, but they do not remove the underlying TC40 fraud report. A fraud case resolved through RDR still counts toward VAMP.
  • Compelling Evidence 3.0 (CE3.0) is currently the only mechanism that removes qualifying fraud disputes from both counts, and it only applies where you have the required prior-transaction data.

Translation: alerts are strong protection against Mastercard's count-based program and against Visa non-fraud disputes, but they do not neutralize fraud-tagged volume under VAMP. If your problem is true fraud rather than friendly fraud, alerts alone will not save your ratio. You also pay per alert, typically in the $35 to $40 range per resolved case, so model the cost against the fine schedule, not against the refund.

A liberal refund policy works upstream of all of this. Every dispute that becomes a support ticket and a refund instead is a unit that never enters any numerator. For high-risk merchants, refund friction is ratio risk.

Multi-MID architecture and ratio exposure

Since ratios are computed per MID, the architecture of your processing setup determines your exposure.

A single MID concentrates everything: one bad campaign, one fraud ring, one affiliate sending garbage traffic, and your only processing relationship breaches threshold. Spreading volume across multiple properly underwritten MIDs, each with its own descriptor and its own honest underwriting file, contains the blast radius. A dispute spike from one traffic source lands on one MID while the rest of your volume keeps settling.

To be clear about the line: multi-MID architecture is legitimate when every MID is disclosed, underwritten for the business it actually processes, and carries accurate descriptors. Load balancing volume to keep any single MID's denominator healthy is risk management. Opening MIDs under different names to hide a bad history is transaction laundering, and it ends in MATCH or worse. The mechanics of doing this correctly, including routing rules and per-MID caps, are covered in the MID load balancing guide.

Practical takeaways

  • Track VAMP and ECM separately, per network, per MID. A blended dashboard rate is a vanity metric.
  • Compute your forward exposure: trailing 60-day sales times your true dispute rate, over projected current-month count. This predicts breaches; the monthly ratio only reports them.
  • Never cut volume sharply while disputes are in flight. If you must scale down, taper.
  • Use alerts with clear eyes: they protect the Mastercard count and Visa non-fraud disputes, but TC40 fraud survives RDR under current VAMP rules.
  • Make refunds frictionless. It is the cheapest numerator control you have.
  • Verify thresholds quarterly. Visa moved the merchant line from 2.2% to 1.5% in April 2026, and there is no reason to assume the current numbers are final.

About the author

Paul Madut designs payment infrastructure for high-risk ecommerce brands: token vaults, MID load balancing, and offshore routing for merchants processing $50K+/month.