How to Spot a Fake or Edited Bank Statement
If you take bank statements from people you do not know, you are being sent edited ones. A rent application with the overdraft weeks deleted, an affordability pack with a salary line raised, a disclosure bundle missing the fortnight that matters: none of it requires skill any more, because a PDF of a statement can be opened in a word processor and typed over.
The good news is that editing a statement is much harder than it looks, because a statement is not a picture of numbers. It is an arithmetic chain, produced by one specific kind of software, on one specific day. Break any of those and the document says so. Below are the five checks that catch most of what actually gets sent, why each one works, and what a hit does and does not prove.
Start with the arithmetic, because it is unforgiving
Every line of a statement carries a running balance, and each one is the line above it plus the credit or minus the debit. That makes the whole document one long sum: take the opening balance, apply every transaction in order, and you must land exactly on the closing balance.
Someone deleting an embarrassing payment almost never repairs the chain behind it. Deleting a 240.00 debit leaves every balance after it 240.00 too low, so the closing figure no longer matches the transactions above it. Raising a salary credit does the same thing in the other direction. The only way to edit a statement and keep the arithmetic is to retype every balance from the change to the end of the document, and that is enough work that most people do not.
This is the single most productive check, and it is worth doing by hand on any statement that matters: opening balance, plus the credits, minus the debits, against the closing balance. A converter that recomputes the running balance line by line does it for you, and a gap is reported as an amount rather than a score, because the amount is what you go and ask about.
Look for rows that repeat exactly
The second thing people do to a statement is pad it. A thin account is made to look busier by copying a plausible line and pasting it a few times, usually a card payment or a transfer in.
Copied rows are identical in a way real rows are not: same date, same description, same amount, three or more times over. Genuine statements do repeat, and a daily coffee or a standing order will look similar, but they do not repeat the same date. Three or more rows matching on all three fields is worth a look, not because it is proof, but because it is what padding looks like when it is done quickly.
Count how many amounts are perfectly round
Invented numbers are tidy. Asked to make up a month of spending, people write 50, 200, 1,000 and 35, because those are the numbers that come to mind. Real spending is 47.63 and 12.99 and 1,842.17, because it is made of prices, interest and fees.
So count the amounts that land on a whole unit with nothing after the decimal point. Over a reasonable sample, twelve transactions or more, a genuine statement with more than about 80% round amounts is unusual enough to be worth explaining. A business account paying suppliers on round invoices is a real exception, which is exactly why this one is a prompt to ask rather than a conclusion.
Read the PDF metadata, which people forget exists
A PDF records what produced it. Open the document properties in any reader and there is a Producer field and a Creator field, filled in by whatever wrote the file, and nobody editing a statement thinks to change them.
A real statement is produced by a bank’s statement generator or a reporting engine. It is never produced by design or office software, so these names in the metadata of a bank statement are a serious finding:
- Photoshop, GIMP, Illustrator, Inkscape or Canva, which are for making images and documents from scratch
- Microsoft Word, LibreOffice, OpenOffice or Pages, meaning the statement was retyped or exported from a document
- iLovePDF, Smallpdf, Sejda or PDFescape, the online editors people reach for to change a PDF they cannot otherwise edit
- Nitro, PDF-XChange Editor, PDFelement, PDFfiller or Foxit PhantomPDF, desktop editors sold for exactly that job
Compare the creation date against the modification date
The same properties panel holds two timestamps: when the file was created and when it was last modified. A bank writes a statement once, in a batch, and never touches it again, so on a genuine document those two are the same moment or a few seconds apart.
A file created in April and modified in June was edited in June. That is not necessarily dishonest, and the innocent explanations are real: someone combined two statements into one file, or re-saved it to compress it for email. It is still a fact about the document you were handed, and it is worth knowing before you rely on it.
What none of this proves
Every one of these checks is a reason to look, not a verdict. Balance reconciliation fails on honest statements when a page was scanned badly and a row was misread. Metadata names a PDF tool when a landlord merged three months into one attachment out of courtesy. Round amounts cluster on a contractor account that invoices in hundreds.
Treat a finding as a question you can put to the person who gave you the document, and note that all five questions have easy answers if the statement is genuine: ask for the original download from online banking, or for the bank to send it directly. Someone with nothing to hide can produce that in a few minutes. The pattern worth acting on is not one flag, it is a flag plus a reason the original cannot be produced.
Running all five in one pass
Doing this by hand takes about ten minutes per statement, which is fine for one and not fine for a caseload. All five checks run automatically on every statement PivotBank converts, and each finding says what triggered it: the size of the balance gap, how many rows repeated, the percentage of round amounts, the exact producer string. The verification page sets out what each check looks at and where its limits are.
Because the parsing happens in your browser, the document does not leave your machine to be checked, which matters when it belongs to a client rather than to you. Mortgage brokers and law firms usually want the transactions out as well as verified, and the analysis view gives you the categorised spending alongside the integrity report on the same upload.