Bank statements as proof of income for a rental application

The application form asks for your last three months of bank statements. You open the first one and it is a record of everything you did with money for a month: the pharmacy, the bar tab, the transfer to your brother, the gym you stopped going to in week two. The landlord asked one narrow question, whether you can pay the rent. The document in front of you answers a far broader one.

The pack you send has two jobs at once. It has to be something the letting agent or referencing firm can actually verify, and it should not hand over the parts of a statement that are payment credentials rather than evidence of income. Those goals conflict in exactly one place, and knowing where saves you a round trip.

What the agent is checking

What a landlord may ask for, what a referencing firm checks, and what you can decline to provide all vary by jurisdiction and by agency, so start from the wording on the application in front of you. The concern underneath is usually affordability, and these are the things a statement can speak to.

• That the income arrives. A credit matching the employer name and the figure on the application form.
• That it arrives on a rhythm. The same credit in each requested month, landing on roughly the same day. A single large deposit with nothing before or after it has the shape of a loan rather than a salary, and it invites a question you would rather answer up front.
• That the account clears the rent. Whether what remains after existing credit commitments and regular direct debits leaves room for the rent you are applying for.
• That payments do not fail. Returned direct debits, unpaid item fees, and days spent below zero. These carry weight out of proportion to their size, because they bear directly on whether the rent will clear, and they are visible in the balance column whether or not anything else is.

None of that needs a debit description. All of it needs dates, amounts, and an unbroken balance column.

Send the bank's own statement documents

Send the statement the bank issues, complete: page 1 of 4 through page 4 of 4, with the header intact. The header carries the account holder name, the address the bank holds, the statement period, and the opening and closing balances. Those four items are what make the pages checkable as a set. Deliver the periods requested and stop there.

A screenshot from a banking app is not a statement. It carries no statement period, no opening or closing balance, and nothing tying the figures to the institution, so there is nothing in it for anyone to check. Where the form asks for statements, expect a screenshot to come back. If your bank only offers the current month as a download and older periods sit in an archive, the archive is where the requested months live; the transaction export usually reaches back much less far than the statement archive does.

Attach a converted spreadsheet alongside the statements, never instead of them. The spreadsheet makes the income obvious in one screen. The statements are the evidence it summarises, and a converted file is a derivative that has to be checkable against its source.

What is genuinely sensitive, and what only feels sensitive

The distinction that matters is not how private something feels. It is whether the item lets somebody move money or impersonate you.

Take these out

• The full account number and the sort code or routing number. Together these are the credentials for setting up a direct debit against your account. Mask all but the last four digits of the account number; those four are enough for the agent to confirm all the pages belong to one account. If the tenancy later requires your details for a standing order, supply them separately, in writing, once the agreement is real.
• A card number printed in full. Some statements print the card PAN on the header of a linked card section. There is no reason for that to be in an application pack.
• Your online banking customer number or user reference. It is a login identifier. It proves nothing about income.
• The name and address of a joint holder who is not on the application. Those are identifiers, along with their name where it appears beside a row, and masking them takes nothing out of the numbers. Their credits are a different matter and have to stay. A credit sits in the running balance column, so removing one either leaves a hole anyone can fill by subtracting one balance from the next, or forces you to remove the balance column too and destroy the only check the agent has. There is no version of that redaction that works. If a joint account exposes more than you are willing to share, the honest remedies are to ask the agent whether a joint account is acceptable for the application at all, or to supply a sole account covering the same months.
• The machine readable strip or barcode at the foot of some statements. It encodes the account identifiers you just masked, so removing the digits and leaving the barcode removes nothing.

Leave these alone

• Your name and the address the bank holds. This is the identity match and, for many agents, part of the address history check. Redacting it defeats the purpose of sending the document.
• The statement period, the bank name, and the page numbering. Without them nobody can tell whether the set in front of them is complete.
• The opening balance, the closing balance, and the running balance column. These are load bearing. Nothing can be checked without them.
• The date and amount of every credit, and the payer on the credits you are relying on. Those are your case. The credits that are not yours stay too, amount and all, because they are part of the arithmetic that makes the rest checkable, even where the name beside one is masked.
• The date column, on every row. Gaps in dates read as removed pages.

That leaves one genuinely discretionary category: the descriptions on your debit rows. Masking a merchant name is defensible, and an agent who needs to see that rent leaves the account each month does not need to see the name of your therapist. Mask descriptions if you want to. Do not touch the numbers beside them.

Redacting an amount does not hide it

This is the redaction that backfires, and it costs you twice over. Take four rows from a March statement, with the balance column the bank printed:

Opening balance 2,140.00. Salary credit of 3,250.00 takes the balance to 5,390.00. Rent of 1,450.00 takes it to 3,940.00. A card payment of 610.00 takes it to 3,330.00. Groceries of 184.20 take it to 3,145.80.

Now black out the 610.00. The row still shows a balance of 3,940.00 above it and 3,330.00 below it. Anyone who can subtract knows the amount was 610.00. You have hidden nothing, and you have handed the agent a document with a visible alteration on it, which is a worse first impression than the transaction ever was.

The alternative is worse still. Black out the balance column as well so the subtraction no longer works, and now nothing on the page can be checked. Nobody can confirm that the salary credit is real, that the pages join up, or that a page is not missing. That leaves two options, taking the document on trust or asking for a clean copy, and the second one is free.

Build the summary the agent can check

The summary is not a substitute for the statements. It is what stops a busy agent from having to read twelve pages to find three salary credits.

1. 1

   Download every requested period in full

   Take complete statement documents from the bank, all pages, for exactly the months the form asks for. Check the page numbering on each one before you go further.
2. 2

   Convert the set into one spreadsheet

   Get every row into a single sheet with date, description, amount, balance, and a column naming the source statement so any figure can be traced back to a page.
3. 3

   Rebuild the balance column before you touch anything

   Confirm each row's rebuilt balance matches the balance printed beside it, that the last row lands on the printed closing balance, and that one statement's closing balance is the next one's opening balance.
4. 4

   Isolate the income rows

   Filter to credits and tag the salary lines. Total them per month. Leave the other credits visible and labelled rather than deleting them.
5. 5

   Redact the identifiers on the statements themselves

   Mask the account number and the other credentials on the statement documents, leaving every date, amount and balance untouched.
6. 6

   Send the statements and the summary together

   Attach both. State in the covering message which months are enclosed and what the summary totals.

What the numbers should look like when you check them

Run these five checks on the summary before it leaves your outbox. Each one is arithmetic you can redo in front of the agent if they ask.

Three salary credits of 3,250.00 come to 9,750.00, and with the 480.00 transfer in, total credits across the three months are 10,230.00. If someone adds the credits and gets a different figure from your summary, the summary is what they stop trusting. Check the closing-to-opening join between consecutive months too, and check the statement periods run end to end with no gap between them. A break in either is cheap to spot here and expensive to be told about later.

What a reconciling summary does and does not prove

Two different checks get called the same thing, and only one of them is worth relying on. The weak one totals a month: opening balance plus every credit minus every debit should land on the printed closing balance. Any two errors that cancel defeat it, and the commonest pair is not exotic. A row lost at a page break and another row of the same value read twice leave the total exactly where it belongs, because they cancel.

The strong one rebuilds the balance column. Start from the opening balance, apply one row at a time, and compare each rebuilt figure against the balance the bank printed beside that row. The March salary credit read as 325.00 instead of 3,250.00 fails on its own row and on every row after it. The lost-and-duplicated pair fails too: the rebuilt column goes out at the row that vanished and stays out until the duplicate cancels it, so every row in between disagrees with the page. That is the check to run on your summary, and it is the reason the balance column has to survive the redaction pass intact.

Even then it only speaks for the amount column. A row dated 3 March that actually fell on 30 March rebuilds perfectly and still misrepresents when your salary lands, and the same goes for a merchant name, a payer name, and a row the conversion recorded with an amount of zero. A balance misread in a way that agrees with the amount misread beside it survives as well. Where a statement prints no running balance there is nothing to rebuild against, and you are back to reading the page. That is why the original statement documents go in the same envelope, and why any line of the summary has to be checkable against the page it came from.

As above, what a landlord may lawfully require and what you may decline to provide vary by jurisdiction, so check the rules where you are renting before you refuse a request outright. The arithmetic holds regardless of where you are. If the same pack is going to a lender rather than a letting agent, underwriters read the same statements for a different set of signals, and the sourcing questions they ask about large deposits are worth anticipating. And before you upload statements anywhere to prepare them, the retention and handling questions are worth settling first.