The email from your bank says something like "as part of our ISO 20022 migration, MT940 statement delivery will be discontinued; please switch to camt.053." If your reconciliation process — a treasury system, an ERP import, or a well-worn spreadsheet macro — was built around MT940, that's not a trivial swap. The good news: camt.053 carries everything MT940 did and quite a bit more. Here's how the two formats line up, and what to watch when you make the move.
Why banks are switching
MT940 is a SWIFT message format from the 1980s: line-oriented text, colon-prefixed tags
like :61: and :86:, and hard limits on line length. It served for four decades, but
SWIFT is retiring the MT category for cash management in favor of ISO 20022 XML, and
SEPA schemes already standardized on the camt family. camt.053 — the
BankToCustomerStatement — is the designated end-of-day replacement. (For a full tour
of the new format, see What is a CAMT.053 file?)
How the fields map
Every MT940 concept has a camt.053 home:
| MT940 | camt.053 |
| ----------------------------- | ------------------------------------------- |
| :20: transaction reference | Stmt > Id |
| :25: account identification | Acct > Id > IBAN (or Othr > Id) |
| :28C: statement number | ElctrncSeqNb / LglSeqNb |
| :60F: opening balance | Bal with type OPBD |
| :61: statement line | Ntry (amount, dates, CdtDbtInd) |
| :61: N-code (e.g. NTRF) | BkTxCd domain/family/sub-family |
| :86: information to owner | NtryDtls > TxDtls (parties, refs, text) |
| :62F: closing balance | Bal with type CLBD |
The shape changes — colon tags become nested XML — but the statement is the same: opening balance, one record per transaction, closing balance.
What camt.053 adds
The :86: line was MT940's junk drawer: counterparty, references, and payment purpose
all mashed into a few 65-character lines, formatted differently by every bank. camt.053
gives each of those a structured field:
- Named counterparties. The debtor or creditor's name and IBAN live in
RltdPties— no more parsing them out of free text. - End-to-end references. The
EndToEndIdthe payer supplied survives the whole payment chain intact, which makes matching invoices dramatically more reliable. - Mandate IDs for SEPA direct debits, straight from
Refs>MndtId. - Real transaction codes. Instead of bank-flavored N-codes,
BkTxCdcarries an ISO classification likePMNT.RCDT.ESCTwith a published meaning. - Untruncated remittance info.
Ustrdtext isn't clipped to fit a line length. - Batch detail. A batch booking is one
NtrywhoseNtryDtlsholds oneTxDtlsper underlying payment — the individual transactions MT940 could only summarize.
What to watch when you migrate
A few things genuinely change behavior, not just syntax:
- Row counts can differ. If your bank sends batch entries with full
TxDtlsdetail, a statement that was 40 MT940 lines might legitimately become 90 camt.053 transactions. That's more data, not an error — but reconciliation totals built on line counts will move. - Pending entries exist. camt.053 entries carry a status;
PDNGitems may appear that MT940 never showed. Filter onBOOKif you only want booked movements. - Versions vary.
camt.053.001.02and.001.08structure a few things differently (party names, status codes). Confirm which version your bank sends and that your tooling reads it. - File size grows. XML is verbose; expect files ten times the size of the MT940 they replace. The content compresses well and parses fine — just don't be alarmed.
- Signs are explicit. Amounts are always positive in the XML; direction lives in
CdtDbtInd. Anything that inferred debit/credit from MT940'sD/Cmarks needs the equivalent logic.
Reading both formats without an ERP
During a migration you'll often hold both files for the same day — the old MT940 and the new camt.053 — and want to eyeball them side by side. Both LegacyFile converters run free in your browser with no upload: the CAMT.053 to CSV converter flattens the XML to one signed row per transaction with counterparties and references, and the MT940 to CSV converter does the same for the SWIFT format. Drop each file in, export CSV or Excel, and the comparison is a spreadsheet diff instead of an archaeology project.