TL;DR: A fake merchant statement is a processing summary, either edited after the acquirer produced it or generated to look like one, that reports card volume, fees, and what the merchant was paid. The FBI's 2025 Internet Crime Report put reported losses at $20.877 billion. That is a broader fraud figure, not a rate for merchant statements. One sample layout fails as a visual check, because real statements differ by processor, and a missing interchange table is normal when the pricing is not interchange-plus.
What a Fake Merchant Statement Actually Is
A merchant statement is issued by the acquirer or the processor. It summarizes a period of card activity: what was submitted, what third parties took, what came back as chargebacks, what fees were assessed, and what amount was processed for the merchant. The card network does not issue it. A bank statement is a different document, covered in how to spot a fake bank statement.
Fiserv's interchange-plus guide is a concrete example of one real layout, not a universal template. The guide marks the page THIS IS NOT A BILL. It walks through a merchant ID, the period, amount submitted, third-party transactions, chargebacks, fees, and amount processed. Other acquirers use other labels and other page orders. A statement that does not match that PDF can still be the processor's ordinary output.
Mastercard publishes interchange that acquirers generally pay issuers. Mastercard is not a party to the price the merchant pays the acquirer. A page that presents itself as a bill from Mastercard, or from Visa, has the wrong issuer. Interchange may appear as a line the acquirer passed through. The absence of an interchange table means the merchant is not on interchange-plus pricing, which is a common real arrangement, not a defect.
The upload arrives in two production classes. Edited-after-creation starts from a real processor PDF, then changes the merchant ID, the period totals, the fee lines, or the amount processed. Generated-from-scratch builds a summary that reconciles on its face, with no processor file underneath. FinCEN's September 9, 2024 analysis of check fraud tied to mail theft found that 44 percent of the checks were altered and then deposited, and 26 percent were used as templates for counterfeits. Those percentages describe checks. They illustrate the two classes, a genuine instrument changed after it existed and a new instrument built from a template, and they are not a rate for merchant statements.
Tells That Still Work, and the Ones That Do Not
The arithmetic still works, on a statement built like the Fiserv interchange-plus example. Amount submitted, minus third-party transactions, chargebacks, adjustments, and fees, equals amount processed. Labels move from processor to processor. The relationship is the tell: the volume in has to survive the deductions and land on the amount the statement says was processed. A total that cannot be rebuilt from the lines on the same page is a content failure. A total that rebuilds cleanly can still be false, because a generated page is written so the lines agree.
The merchant ID still works, and only if it is confirmed with the processor through a channel the reviewer looks up. The MID printed on the file is part of the document under test. A number that the processor does not recognize, or that belongs to a different legal name, is the case. A number the processor confirms still leaves open whether the totals were edited after the statement was produced.
One sample layout fails. Hunting for the Fiserv sentence THIS IS NOT A BILL fails on every other acquirer's design, and a forger who has seen that guide can print the sentence. A missing interchange table fails as a tell, because the table shows up when pricing is interchange-plus and drops out when it is not. A statement offered as proof that cash is on hand is the wrong document for that claim. Cash position is a different file, covered in how to spot a fake proof of funds letter.
How Detection Actually Works
The file's story is that this acquirer or processor summarized this merchant ID for this period, and that submitted volume, minus third-party transactions, chargebacks, adjustments, and fees, produced the amount processed. Detection asks whether the bytes support that story.
Production method asks how the PDF was made. A processor download, a later edit of that download, a spreadsheet printed to PDF, and a generative pipeline do not share a path, even when the column headings match. Edited-after-creation leaves a real statement under a later save. The totals are the usual last change. Generated-from-scratch has no processor export underneath, and the lines reconcile because they were composed together. The case note should name the class.
Timestamp trail asks whether creation and modification fit the statement period. A March statement downloaded in April is ordinary. A March statement whose file was created in September, with a modify time minutes before upload and no earlier processor export, needs a story. Portal reprints and accountant merges add noise, so a new save date is evidence, not an automatic fail.
Issuer matching asks whether the page comes from an acquirer or processor that could have billed this merchant. The card network is the wrong issuer. Mastercard sets interchange that acquirers generally pay issuers, and it is not a party to the merchant's price. The MID, confirmed with the processor on a number the reviewer looks up, ties the page to an account. The legal name on that account should match the merchant in the application. A formation record that does not match that name is a separate check, covered alongside how to spot a fake certificate of good standing.
Consistency is the arithmetic turned inward, then outward to deposits. The lines on the statement have to rebuild the amount processed. Fees, chargebacks, and third-party totals have to be the same numbers in the summary and the detail. A generated statement passes that test by construction. Model artifacts are processor chrome, fonts, and section order that do not belong to that acquirer's export. Recycled patterns are the same template, the same MID, or the same fee table on an unrelated merchant. One reviewer with one month will not see the cousin file. The same classes of signal on other banking PDFs are covered in document fraud in banking.
What to Do with a Suspect File
Hold the upload. Rebuild the period from the lines on the page: amount submitted, minus third-party transactions, chargebacks, adjustments, and fees, should equal amount processed. If the labels differ, map them before declaring a break. A total that cannot be rebuilt is enough to stop. A total that rebuilds is not enough to accept.
Confirm the merchant ID with the processor, using a phone number or portal the institution looks up, not the number printed on the statement. Ask whether that MID belongs to the legal name on the application, and whether the period totals match what the processor shows. A mismatch is the case. A match leaves an edit-after-creation question open, because a real MID can sit on a statement whose totals were changed later.
Do not reject the file for lacking an interchange table, and do not reject it for failing to match one public sample. Do not read it as a bill, and do not read it as a bank statement. When the totals match the processor and the file still looks rewritten, ask which figures changed after the statement was produced. For a scored read of that history, check the file. The x-ray is production method, timestamps, issuer match, consistency, model artifacts, and recycled patterns. The processor's own total remains the outside check.
Where Sphinx Fits
Watchdoc scores the statement and shows the x-ray a reviewer can point to. Sphinx's published product figures describe the check as 54% less expensive, priced at $0.45 per document. The Watchdoc playground is free for the first file. Run the totals against the file's history, and against the processor, before the volume is treated as real.
Frequently Asked Questions
How can you tell if a merchant statement is fake?
Rebuild amount processed from amount submitted minus third-party transactions, chargebacks, adjustments, and fees, then confirm the merchant ID with the processor using a number you look up. Visual similarity to a sample statement is a weak test, because layouts differ and a missing interchange table is normal outside interchange-plus pricing. Competent fakes fail when totals were edited after the processor produced the file, or when the page was generated and the processor has no matching MID.
Is a merchant statement a bill or a bank statement?
The acquirer or processor issues it as a summary of card activity. Fiserv's interchange-plus guide prints THIS IS NOT A BILL on that layout. Other processors do not use that sentence, so the sentence is not the test. A merchant statement is not a deposit account statement, and it is not an invoice from Mastercard or Visa.
Why would a real merchant statement omit the interchange table?
Interchange-plus pricing itemizes the interchange the acquirer passes through. Other pricing bundles those costs into a discount rate or a flat fee, so the table never appears. Mastercard publishes the interchange acquirers generally pay issuers, and Mastercard is not a party to the merchant's price. A missing table does not show the statement is fake.
Who should confirm the numbers on a merchant statement?
The processor that owns the merchant ID, reached through a channel the reviewer looks up rather than a phone number or link printed on the file. The card network did not issue the statement. A confirmed MID and confirmed period totals still leave open whether the PDF was edited after the processor generated it.
.png)
