CbC Reporting XML Schema Errors: Common Filing Mistakes and How to Avoid Them

CbC reporting XML schema errors rarely come from one cause. A file that fails validation outright and a file that passes validation but still gets flagged by a tax administration break for different reasons, at different stages of the same process — and most troubleshooting lists collapse both into one undifferentiated pile of “common mistakes.” Splitting the failures into three layers makes the difference actionable: a structural error means the parser never got past the file’s shape; a business-rule error means the shape is fine but the logic inside it isn’t; a content error means the file is technically compliant and still wrong in substance. Tax authorities automate detection of the first two. Only some catch the third — which is exactly why it causes the most damage downstream.

Why “schema valid” doesn’t mean “correct”

CbC reports move through a network the OECD describes as covering 110+ jurisdictions with a CbC reporting filing obligation and more than 3,300 bilateral relationships for exchanging those reports, reviewed annually across 136 jurisdictions in the current peer review cycle (OECD, BEPS Action 13). At that scale, a national tax authority’s XML validator is doing one job: catching what can be checked mechanically before the file is accepted into that exchange network. It cannot catch everything, because plenty of CbC reporting mistakes are judgment calls about what number goes where — not malformed XML. A multinational that treats “the file validated” as “the report is correct” is checking the wrong thing.

Layer 1: structural errors — the file doesn’t parse

These are failures before the tax authority’s system even reads the numbers. HMRC’s CbC XML User Guide lists several as named business rules that fire at this stage, including CbC-MNE-BR-006 (failed schema validation against the published .xsd), CbC-MNE-BR-012 (more than one CbcBody element in a single file), CbC-MNE-BR-013 (the CBC_OECD version attribute left unpopulated) and CbC-MNE-BR-014 (the file is not encoded in UTF-8) (HMRC, Country-by-Country XML User Guide V2.0). Hong Kong’s Inland Revenue Department enforces the same category of checks on its own portal: the CBC_OECD root element’s version attribute must read “2.0”, every file must be UTF-8 without a byte-order mark, and the upload is capped at 50 MB (Hong Kong IRD, CbC Return XML Schema v2.0 User Guide). None of these are tax questions. They are produced almost exclusively by whatever system generates the XML — a stale schema version hard-coded into an export script, an encoding default left at the OS locale instead of UTF-8, or a template that duplicates a body element when a filer merges two entities’ data by hand. Groups that have already automated their CbC data collection tend to see fewer of these errors simply because the export step is no longer a manual, one-off script.

Layer 2: business-rule errors — the file parses, the logic doesn’t

This is where most rejected CbC filings actually fail, and it’s the layer with the most jurisdiction-specific detail. The Australian Taxation Office publishes a full table of its own CbC lodgment error codes, separate from the OECD schema’s own validation: E013 fires when the same DocRefId is used on more than one record; E033 requires the ReceivingCountry field to read “AU”, with both transmitting and receiving country set to AU for domestic lodgers; and E035 catches a DocTypeIndic that doesn’t match the MessageTypeIndic sent — a CBC401 message (new information) may only carry OECD0 or OECD1 records, while a CBC402 (corrections or deletions) may only carry OECD0, OECD2 or OECD3 (ATO, Guidance on lodgment errors). HMRC’s own business rules cover equivalent ground for UK filings: CbC-MNE-BR-009 requires ReceivingCountry to be “GB”, CbC-MNE-BR-008 checks that the SendingEntityIN matches the filer’s own CbCID, and a family of rules numbered CbC-MNE-BR-005a through 005g governs the MessageRefId alone — covering whether it is populated, matches the filename, follows the required format, is a duplicate, matches the filer’s CbCID, matches the reporting year, and carries a valid timestamp (HMRC, CbC XML User Guide V2.0). None of this logic is visible from reading the OECD’s base schema. It is enforced entirely by the receiving jurisdiction’s own systems, on top of the schema.

JurisdictionRule / codeWhat it actually flags
Australia (ATO)E013The same DocRefId used on more than one record in the file
Australia (ATO)E029Incorrect name, ABN or TFN given for an Australian constituent entity
Australia (ATO)E035DocTypeIndic doesn’t match the MessageTypeIndic sent (CBC401 vs CBC402 rules)
United Kingdom (HMRC)CbC-MNE-BR-009ReceivingCountry must be “GB”
United Kingdom (HMRC)CbC-MNE-BR-012More than one CbcBody element in the file
United Kingdom (HMRC)CbC-MNE-BR-005 seriesMessageRefId missing, duplicated, mismatched to the CbCID/reporting year, or wrongly formatted
Hong Kong (IRD)Portal validationDocRefId must be unique and no longer than 200 characters; MessageRefId no longer than 170 characters

Layer 3: content errors that pass validation anyway

This is the layer no XML validator can see, because the file is structurally sound and the business rules all pass — the numbers themselves are simply wrong in substance. HMRC’s own internal manual on completing the CbC report flags several of these explicitly: dividend payments received must be excluded from both Revenue and Profit (Loss) Before Income Tax for accounting periods beginning on or after 1 January 2020, with a further HMRC interpretation applying to periods beginning on or after 1 January 2024 that depends on how the payor treated the distribution; Table 1 must use full, un-rounded figures, never shortened or abbreviated numbers; each jurisdiction should appear on exactly one line in Table 1; Number of Employees, Tangible Assets and Stated Capital are all expected to be positive values, with a Table 3 explanation required wherever they aren’t; and a single CbC report should never mix more than one reporting currency (HMRC, International Exchange of Information Manual, IEIM300140). HMRC’s own manual page also points filers to the OECD’s running compilation of errors MNEs make in preparing CbC reports as a separate, periodically updated reference — a signal, from a tax authority itself, that this content layer is treated as a distinct and recurring compliance problem, not an edge case.

None of this will stop a file from validating. It will, however, get the report queried, resubmitted, or flagged in a peer review — and because CbC reports move through the bilateral exchange network described above, a content error caught by one jurisdiction’s manual review doesn’t stay contained to that jurisdiction’s copy of the file.

Why a file that’s valid in one country can still be rejected in another

The OECD publishes one base CbC XML schema, but every receiving tax authority layers its own portal-level validation on top of it, and those layers don’t match. Hong Kong’s IRD runs a two-stage check — schema validation, then a separate database validation against the filer’s registration details — before posting results to the filer’s Message Box, and enforces field-length limits (170 characters for MessageRefId, 200 for DocRefId) that aren’t stated the same way in every jurisdiction’s own guide (Hong Kong IRD, CbC Return XML Schema v2.0 User Guide). The ATO’s E0-series codes and HMRC’s CbC-MNE-BR codes aren’t interchangeable either; they’re each tax authority’s own implementation of OECD-level concepts like DocRefId uniqueness and country-code matching, expressed as that jurisdiction’s own numbering system. The practical consequence for a group filing in multiple jurisdictions, which is the normal case given the thresholds that trigger a CbC obligation in the first place: a template built and tested against one country’s portal is not safe to assume valid everywhere else it gets submitted, even when every figure in it is identical.

A pre-submission checklist worth running every time

  • Confirm the schema version attribute matches what the receiving portal currently expects — not what last year’s template used.
  • Force UTF-8 encoding without a byte-order mark at export, rather than trusting a default that may vary by workstation.
  • Check DocRefId and MessageRefId for uniqueness and length against the specific jurisdiction’s limits before submission, not just against the OECD base schema.
  • Match MessageTypeIndic to DocTypeIndic deliberately for every record — new filings and corrections carry different allowed combinations, and mixing them is one of the most common rejection triggers across jurisdictions.
  • Re-check Table 1 for dividend treatment, rounded figures, mixed currencies and negative values even after the file validates clean — these are content errors no schema check will ever catch.
  • Treat each jurisdiction’s portal as its own validator, with its own field limits and error codes, rather than assuming a file accepted in one country will be accepted everywhere.

Most CbC XML errors start upstream of the file itself. If Table 1 and Table 2 figures are inconsistent before they ever reach your XML generator, no amount of schema validation fixes that — the CbC Report Data Pack gives you a structured way to collect, convert and reconcile that underlying data against the OECD’s own column definitions before it goes anywhere near a schema. €39.

FAQ

What are the most common CbC reporting XML schema errors?

They split into three layers: structural errors that stop the file from parsing at all (wrong encoding, wrong schema version, duplicate elements), business-rule errors where the file parses but breaks a jurisdiction’s own logic (duplicate or malformed DocRefId/MessageRefId, mismatched country codes, wrong MessageTypeIndic/DocTypeIndic combinations), and content errors that pass every automated check but are substantively wrong — dividend treatment, mixed currencies, rounded figures, unexplained negative values.

Does passing XML schema validation mean my CbC report is correct?

No. Schema and business-rule validation only confirm that the file is structurally and mechanically sound. Several substantive requirements — how dividends are treated in Revenue and Profit Before Tax, using one currency throughout, not rounding Table 1 figures — are enforced by HMRC’s own completion guidance and the OECD’s compiled error list, not by the XML validator itself.

Why does the same CbC XML file get accepted in one country and rejected in another?

Because each tax authority runs its own validation layer on top of the shared OECD base schema, with its own error codes and field limits — the ATO’s E0-series codes, HMRC’s CbC-MNE-BR codes and Hong Kong IRD’s two-stage schema-then-database check are three different implementations of broadly similar OECD-level concepts, not the same check running three times.

Where does the OECD publish a list of common CbC reporting errors?

The OECD maintains and periodically updates a compiled list of common errors made by MNEs when preparing CbC reports, referenced directly from HMRC’s own completion manual. Several national tax authorities, including the ATO and HMRC, also publish their own jurisdiction-specific error-code guidance covering the same ground in more technical detail.

Further reading: Guide to International Transfer Pricing: Law, Tax Planning and Compliance Strategies (Duff & Phelps, Kluwer Law International) is a standing reference for the documentation and compliance framework CbC reporting sits inside of, beyond the XML mechanics covered here.

Leave a Reply