GIR XML: the format, the schema, and the gotchas

Everything you need to know about the OECD GloBE Information Return XML format — from namespace declarations to Part E reconciliation.

The GloBE Information Return (GIR) is the OECD's standardised filing format for Pillar Two top-up tax. Tax authorities expect to receive GIR data in XML format following the OECD's published schema. Here is what you need to know if you are building a GIR parser, validator, or filing system.

Structure overview

A GIR XML file has two top-level sections:

MessageSpec — filing metadata: sending entity TIN, transmitting country, message type, reference ID, and reporting period.
GIRBody — the substantive content, divided into Parts A through E.

Part A contains the jurisdictional top-up tax computation. Part B covers standard entity information. Parts C and D address specific allocation rules. Part E contains the summary reconciliation.

Part E reconciliation

Part E is where many implementations fall over. The total top-up tax payable must equal the sum of IIR top-up tax and UTPR top-up tax, less any QDMTT credit. A common bug is computing these figures independently and allowing rounding differences to accumulate — the Part E total then fails to reconcile, and the filing is rejected.

Common validation errors

ETR > 1.0 — an ETR above 100% is impossible and indicates a data error (usually swapped numerator and denominator).
Duplicate jurisdiction codes — each ISO 3166-1 alpha-2 code must appear exactly once in Part A.
QDMTT contradiction — claiming the QDMTT safe harbour while QDMTTApplied = false is a logical contradiction that most validators will reject.
Non-numeric amounts — currency amounts must be integers (no decimals) in most schema versions.

Download the sample files

The Pillar2OS resources page offers four GIR XML files: a valid sample, an invalid sample with five annotated errors, a multi-jurisdiction sample (10 jurisdictions), and a large sample (50 jurisdictions) for performance testing.