Willow Audit Willow Club Willow Academy
Articles

Pillar Two TP Adjustment Case Study

This case study isolates a timing tension that becomes visible only because Pillar Two is accounting anchored. A transfer price can be adjusted through a tax procedure long after the transaction year, while entity financial accounts for the transaction year remain unchanged. The main question is not whether an arm’s-length price should be reflected for GloBE purposes (Article 3.2.3 assumes that problem), but in which Fiscal Year the Pillar Two computation must absorb the adjustment and how that timing interacts with the post-filing mechanics for Covered Taxes. In addition, the case study considers a specific practical variant that frequently arises in unilateral audit scenarios: the counterparty jurisdiction (here, the UAE) does not grant corresponding Corporate Tax relief, yet the group must still determine whether (and when) a symmetry adjustment to UAE Pillar Two Income is required for UAE DMTT purposes.

Facts

An MNE Group has two Constituent Entities:

  • DE Co located in Germany.
  • UAE Co located in the UAE.

Year 1 (transaction year): UAE Co provides marketing services to DE Co. The group’s accounts record a fee of 100: DE Co books expense 100; UAE Co books revenue 100.

Year 3 (audit year) (German unilateral adjustment): Germany audits DE Co’s Year-1 return and asserts the arm’s-length fee should have been 80, disallowing 20 of the Year-1 deduction (a unilateral adjustment). Germany assesses additional CIT of 6 (assume 30% of 20), and (per the fact pattern) the assessment is issued/recorded in Year 3.

On the UAE side, there are two possible paths:

  • No relief in the UAE. UAE Co continues to report revenue 100 and no UAE corporate tax reduction arises. So an adjustment in Germany is purely unilateral, at least initially.
  • UAE corresponding relief is later granted. UAE Co’s taxable income for Year 1 is reduced by 20 and (assuming 9% UAE CT) UAE corporate tax reduces by 8 for Year 1, recognised when recorded in the accounts.

Questions

  • For German minimum tax (DMTT) purposes, in which period should DE Co recognise (i) the +6 increase to Covered Taxes (numerator) and (ii) the 20 adjustment to the Pillar Two denominator (GloBE / Pillar Two Income)?
  • If the UAE later grants corresponding relief that reduces prior-year Covered Taxes, must UAE Co recompute the Year-1 ETR / Top-up? If it does, must the income side be adjusted in the same year?
  • Cross-border symmetry. If UAE Co does not obtain UAE relief for corporate tax purposes, should a symmetry adjustment still be made in the UAE DMTT computation (and, if so, which periods should be adjusted)?

Executive summary

Upon examination of the applicable legislation and guidance, the better view is as follows.

1. In Germany, both the tax increase and the corresponding GloBE-income adjustment should be reflected in Year 3, not by rewriting Year 1.

2. If UAE Co in Year 3 files a voluntary disclosure for Year 1, or equivalent corresponding relief is granted through MAP, APA or another accepted UAE mechanism, and that relief reduces Year 1 Corporate Tax, then:

  • the jurisdictional ETR and Top-up Tax for Year 1 must be recomputed, and
  • the Pillar Two denominator must be adjusted together with the numerator, that is, in Year 1.

3. However, if the UAE corresponding relief produces an aggregate decrease in Corporate Tax for the jurisdiction for Year 1 of less than EUR 1 million, UAE Co may elect to treat that decrease in the re-determination year rather than reopen Year 1. If the decrease is material, Year 1 must be recomputed.

4. For UAE DMTT purposes, a symmetry adjustment of Pillar Two Income may still be required even if UAE Corporate Tax is not reduced to neutralize the double-taxation concern. However, because UAE Covered Taxes do not change, Article 4.6.1 is not engaged on the UAE side, and the rules do not provide a single explicit timing rule for this denominator-only correction. Under the alternative readings considered, the timing splits into two camps:

  • Year 3 (current-year recognition): Variants 1 and 3 treat the UAE denominator-only adjustment as something taken in the year the issue is identified for Pillar Two reporting.
  • Year 1 (transaction-year recognition): Variants 2 and 4 treat the adjustment as a correction to the Year-1 transaction and therefore reflect it in Year 1.

The case study therefore presents the UAE no-relief timing as an interpretive choice, not as a settled rule.

5. Materiality does not switch off the arm’s-length symmetry requirement. In the no-relief path, materiality is:

  • irrelevant for Variants 1 and 2 (their timing rules do not depend on materiality),
  • relevant in principle but practically not engaged for Variant 3 (because the 4.6.1 “material/immaterial decrease” machinery is not triggered when there is no Covered Taxes decrease), and
  • relevant as a design choice for Variant 4 (if the group chooses to import a threshold analogous to the EUR 1m election for negative adjustments, otherwise Variant 4 behaves like Variant 2).

Analysis

The governing architecture

1. Pillar Two is a jurisdictional ETR system. It compares Adjusted Covered Taxes (numerator) with GloBE (Pillar Two) Income (denominator). Where a transfer price is later adjusted through tax procedure (audit, unilateral APA, MAP), the tax outcome may be booked or paid in a later year (here Year 3), while the financial accounts for the transaction year (here Year 1) may remain unchanged.

2. Article 3.2.3 supplies the substantive direction: “Any transaction between Constituent Entities located in different jurisdictions that is not recorded in the same amount in the financial accounts of both Constituent Entities or that is not consistent with the Arm’s Length Principle must be adjusted so as to be in the same amount and consistent with the Arm’s Length Principle…”. So controlled transactions must be reflected consistently with the Arm’s Length Principle and, where necessary, symmetry in income of one party and costs of its counterparty must be achieved across counterparties for GloBE purposes.

3. Para 101 of the OECD Commentary explains that Article 3.2.3 is not an abstract ALP policing mechanism but a targeted anti-distortion rule: it requires adjustment to FANIL to avoid double taxation or double non-taxation where taxable income for one or more counterparties is determined using a transfer price different from the one used in the financial accounts. According to para 100, these differences can arise at filing or later on audit.

4. The remaining question that drives this case study is: when must the Pillar Two computation reflect the adjustment? Article 3.2.3 states the symmetry requirement, but it prescribes such a symmetry only for income of one counterparty and correspondent costs of another. It doesn’t establish a rule for recognizing income and local income tax simultaneously and doesn’t specify which Pillar Two tax period should be affected to adjust GloBE income.

5. According to Article 3.2.1(h), “a Constituent Entity’s Financial Accounting Net Income or Loss is adjusted for the following items to arrive at that Entity’s GloBE Income or Loss: … (h) Prior Period Errors and Changes in Accounting Principles”. Corrections of errors referred to in this Article should be reflected as an adjustment in the computation of GloBE Income for the fiscal year in which the correction is recognized in the financial accounts, rather than through a recomputation of GloBE Income for prior fiscal years. Under the structure of Chapter 3, GloBE Income is determined by starting with the financial accounting net income or loss of the Constituent Entity for the relevant fiscal year (Article 3.1) and then applying the adjustments listed in Article 3.2.1. As such, the adjustments under Article 3.2.1 operate as modifications to the accounting result of that fiscal year.

6. At this stage, it is not clear whether Article 3.2.1(h) applies to book-to-tax errors as Article 10.1 defines “Prior Period Errors and Changes in Accounting Principles” as “all changes in the opening equity at the beginning of the Fiscal Year of a Constituent Entity (a) a correction of an error in the determination of Financial Accounting Net Income in a previous Fiscal Year that affected the income or expenses includible in the computation of GloBE Income or Loss for such Fiscal Year…”. Financial Accounting Net Income id defined in Article 3.1.2 as “the net income or loss determined for a Constituent Entity (before any consolidation adjustments eliminating intra-group transactions) in preparing Consolidated Financial Statements of the Ultimate Parent Entity”. “Net income” describes “profit and loss”.[1]

An amount of profit and loss can is to be taken from “a statement of profit or loss and other comprehensive income for the period”.[2] Separately presenting revenue should be a part of this statement.[3] Under IFRS 15.46-47:

  • “When (or as) a performance obligation is satisfied, an entity shall recognise as revenue the amount of the transaction price…”.
  • “The transaction price is the amount of consideration to which an entity expects to be entitled in exchange for transferring promised goods or services to a customer, excluding amounts collected on behalf of third parties (for example, some sales taxes)” (para 47); and
  • “An entity shall consider the terms of the contract and its customary business practices to determine the transaction price” (para 46).

Therefore, under IFRS 15, revenue should be recognized at the transaction price, which is the amount the entity expects to receive for transferring goods or services. This transaction price may differ from the arm’s length price in related party transactions, as IFRS 15 does not require revenue to be recognized at arm’s length. Instead, it uses the actual agreed-upon price between the parties involved. This means that the revenue recognized in the financial statements will be based on the transaction price agreed by the parties and not necessarily the arm’s length price.

The arm’s length principle applies in taxation for transfer pricing purposes but is not a requirement for financial accounting under IFRS 15. Consequently, a book-to-tax correction of revenue does not qualify as “a correction of an error in the determination of Financial Accounting Net Income”.

7. In addition, Article 4.6 governs adjustments to amounts of Covered Taxes incurred with respect to a jurisdiction after the GloBE Information Return for the period has been filed.[4]

Pursuant to Article 4.6.1 “a Constituent Entity’s liability for Covered Taxes for a previous Fiscal Year recorded in the financial accounts shall be treated as an adjustment to Covered Taxes in the Fiscal Year in which the adjustment is made, unless the adjustment relates to a Fiscal Year in which there is a decrease in Covered Taxes for the jurisdiction. In the case of a decrease in Covered Taxes included in the Constituent Entity’s Adjusted Covered Taxes for a previous Fiscal Year, the Effective Tax Rate and Top-up Tax for such Fiscal Year must be recalculated under Article 5.4.1”.

Para 119 of the Commentary to this Article explains that these rules apply to “a change in the amount of income recognised for local tax purposes due to an examination of the tax returns by the tax authority of a jurisdiction or a review of the tax returns by the Entity’s management or tax advisers. In addition, the liability for Covered Taxes may increase or decrease due to errors or more accurate estimates of the tax liability after the GloBE Information Return is filed”.

8. At first glance, it may appear that the timing rules applicable to the denominator (income) and the numerator (tax) operate independently. Article 3.2.3 focuses on ensuring that the arm’s-length principle is applied to income, while Article 4.6.1 deals with the treatment of Covered Taxes. Furthermore, Article 3.2.3 on arm’s-length adjustment does not directly link to either Article 3.2.1(h) or Article 4.6.1, and neither Article 3.2.1(h) nor Article 4.6.1 refers back to Article 3.2.3. This structural separation may give the impression that the timing rules for tax and income adjustments are independent of each other. However, the OECD Commentary accompanied with a definition of Prior Period Errors clarifies that while these articles operate in separate domains, they work together in the overall framework to ensure proper alignment of tax and income adjustments for GloBE purposes.

Why Articles 3.2.1(h), 3.2.3 and 4.6.1 must be read as a system (not silos)

9. The definition of “Prior Period Errors and Changes in Accounting Principles”[5] contains a critical carve-out: error corrections are excluded “to the extent such error correction resulted in a material decrease to a liability for Covered Taxes subject to Article 4.6”.
That carve-out is a drafting signal: once 4.6 is engaged (material decrease), the system does not allow “income-side” adjustments to float on a different timing track.

The cited carve-out is designed to ensure that tax and income adjustments stemming from material changes in prior-period liabilities cannot be treated independently. They must be recalculated together, ensuring coherence in the overall Pillar Two adjustment process. This approach reflects the broader principle embedded in the OECD Commentary, which seeks to ensure that adjustments are made in a way that maintains the integrity of the GloBE framework and avoids distorting outcomes, especially when tax relief or corrections affect the tax base.

10. Paragraph 122 of the Commentary to Article 4.6.1 makes this linkage explicit: the purpose is commensurability of numerator and denominator. It states the connective policy sentence: the rule is linked to Article 3.2.1(h), and it is designed to ensure that GloBE Income or Loss and the Covered Taxes associated with it are taken into account for the same Fiscal Year in ETR / Top-up computations to avoid a distorted result. The same paragraph then explains operationally how that commensurability is achieved:

  • where a correction produces a material decrease, taxes (and, if relevant, GloBE income) are re-determined for the prior year;
  • where it does not, corrections are taken into account prospectively.

11. Para 99 of the Commentary to Article 3.2.3 states that, where authorities agree the price must be adjusted to the same arm’s-length price, each CE adjusts its GloBE Income or Loss, and “the adjustment … is taken into account … pursuant to Article 4.6.1”. This is the interpretive bridge:

  • the Commentary does not treat a TP audit as automatically rewriting Year 1;
  • it routes timing through the post-filing mechanics reserved for Covered Tax corrections.

12. The key takeaway from the above is that Articles 3.2.1(h), 3.2.3, and 4.6.1 must be understood as part of a unified system, not as separate provisions. This interconnected approach ensures that tax and income adjustments are applied consistently within the Pillar Two framework. Once a material decrease in Covered Taxes is identified, the system mandates that both tax and income adjustments be recalculated together to maintain the integrity of the GloBE framework. The carve-out in Article 4.6 and the guidance from the OECD Commentary emphasize that adjustments cannot be processed on separate timing tracks, ensuring that both the numerator and denominator align for accurate ETR and Top-up computations. This coherence between tax and income adjustments prevents discrepancies and avoids distorting the Pillar Two results. Therefore, recognizing these provisions as part of a holistic system is essential for the proper implementation of the GloBE rules.

What Article 3.2.3 changes and what it does not

13. OECD Commentary paragraph 104 states that “Article 3.2.3 does not impose any requirements beyond an arm’s length price. Thus, it does not require the MNE Group to conform the timing of an item of income or expense for GloBE purposes to the timing of that item for local tax purposes”. The point is that, once the transaction must be brought to the arm’s-length amount for Pillar Two purposes, the recognition of the relevant income, expense and associated tax effects must be determined under the Pillar Two rules themselves rather than by mechanically following the local tax timing of the transfer-pricing adjustment.

14. Paragraph 104 real significance is that, once the arm’s-length amount must be reflected for Pillar Two purposes, the timing of the resulting GloBE consequences is not dictated by domestic tax law, but must be determined under the GloBE Rules themselves. Accordingly, para 104 is best read as confirming the autonomy of Pillar Two timing rather than as a mere disclaimer. Even if a transfer-pricing adjustment is recognised in another period for local tax purposes, the question of when the related Covered Taxes and the corresponding GloBE tax base are taken into account must be resolved by reference to the Pillar Two framework, including its specific rules for post-filing adjustments, and not by mechanically following local tax timing.

Application to the fact pattern

Germany: where does the +6 go, and does the Year-1 denominator change?

15. On the facts, Germany assessed additional Year-1 tax but (procedurally) recorded it only in Year 3. That is an increase in prior-year tax liability recognised in Year 3.

16. Germany’s Minimum Tax Act (MinStG) provides a direct analogue: Section 52(1) treats increases in prior-year tax liability reported in annual financial statements as an adjustment to covered taxes in the year in which the adjustment is made.

17. Accordingly, in this fact pattern, the clean technical treatment is:

  • Numerator (Germany Covered Taxes): +6 in Year 3.
  • Denominator (Germany GloBE Income): +20 also in Year 3.

So, there is no automatic “rewrite” of Year 1 merely because Germany changed the income (Covered) tax base for Year 1. Under the architecture emphasized above, the trigger for reopening the prior year under 4.6.1 is not satisfied by a mere increase. The OECD’s 4.6.1 discussion focuses on the recomputation requirement in the event of a decrease and treats other corrections prospectively.

UAE: if UAE grants corresponding relief, do we reopen Year 1 and move both numerator and denominator together?

18. If UAE grants corresponding relief and UAE Corporate Tax for Year 1 decreases by 1.8, the UAE DMTT text[6] squarely engages Article 4.6.1’s “decrease” branch: where there is a decrease in Covered Taxes included for a previous year, the ETR and Top-up Tax for that year must be recalculated. In that recalculation, Pillar Two Income for that year and any intervening years must be adjusted as necessary and appropriate.

19. This instruction answers “period alignment” point directly: once the UAE is in the recomputation branch, the statute itself demands denominator alignment as part of the recalculation (not merely a numerator correction).

Cross-border symmetry if UAE does not grant relief

20. If UAE does not grant corresponding relief under UAE corporate tax, the Group faces a different question: does it nevertheless adjust UAE DMTT income to 80 to mirror Germany?

21. Here Article 3.2.3’s doctrinal limiter becomes decisive. Paragraph 101 presumes the taxable-income transfer price is arm’s length and provides that GloBE income “should be adjusted accordingly… where necessary to prevent double taxation or double non-taxation”.
Paragraph 102 then explains the precise danger in the present case fact pattern: a local adjustment that increases taxable income in a high-tax jurisdiction (Germany) requires a corresponding decrease to the GloBE income of counterparties in an under-taxed jurisdiction (UAE). Otherwise, the income included in the high-tax jurisdiction under local law could be taxed twice: once locally (Germat Cortporate Tax) and again under the GloBE rules (DMTT in the UAE).

22. That is not hypothetical. It is exactly the structure of OECD Example 3.2.3-2: one jurisdiction’s taxable income increases relative to the accounts (because an expense is reduced for tax purposes), and Article 3.2.3 requires the counterparty to include “20 less income” in its GloBE computation to avoid double taxation under the GloBE Rules.

23. Applied to the present case:

  • Germany’s unilateral disallowance (+20) increases Germany taxable income.
  • If UAE remains at revenue 100 for GloBE purposes, the same margin remains embedded in UAE’s GloBE base and could be subject to top-up, while Germany has already taxed it.

Absent an Article 3.2.3 adjustment on the UAE side, the Pillar Two system risks precisely the double-taxation concern described in OECD Commentary paragraph 102: the income becomes taxed (or exposed to top-up) twice within the GloBE architecture because one counterparty’s base has been increased for tax purposes without a corresponding reduction on the other side for GloBE purposes.

24. OECD paras 102 and 103 set up two contrasting patterns, and they are the right starting point for analysing the present case.

24.1. Paragraph 102 addresses the double-taxation risk created by a unilateral adjustment in a higher-tax jurisdiction: where one jurisdiction increases taxable income (e.g., by disallowing part of a deduction) but the counterparty jurisdiction does not make a corresponding adjustment, the same margin can remain embedded in the counterparty’s GloBE base and therefore be exposed to Top-up Tax while it has already been subjected to local tax in the adjusting jurisdiction. To prevent that GloBE double-count, paragraph 102 contemplates that the counterparty’s GloBE Income or Loss may need to be reduced correspondingly under Article 3.2.3.

24.2. By contrast, para 103 states the veto: Article 3.2.3 adjustments are not made where they would themselves create double non-taxation or double taxation under the GloBE Rules. It illustrates this with:

  • unilateral reductions that would create a GloBE “hole” (income neither taxed nor topped up) and
  • unilateral increases in an under-taxed jurisdiction that would expose income to Top-up Tax in one jurisdiction while the income is already taxed and (or) topped up elsewhere.

24.3. Applied to the present case, the primary adjustment is made in Germany (a high-tax jurisdiction), increasing Germany taxable income by 20 via disallowance of deduction. If UAE does not grant corporate tax relief, the margin can remain embedded in UAE’s GloBE base and could be subject to Top-up Tax there, while Germany has already taxed it. That is the paragraph-102 pattern (high-tax increase should trigger corresponding decrease on the counterparty side to avoid GloBE double taxation), not the paragraph-103 veto pattern (which is concerned with adjustments that would create a GloBE hole or double charge).

25. Accordingly, even without UAE corporate tax relief, Article 3.2.3 can require a UAE DMTT denominator adjustment (i.e., reducing UAE Pillar Two Income by 20) to neutralise the double-taxation concern described in paragraph 102. This is exactly the structure illustrated by OECD Example 3.2.3-2, where one jurisdiction’s taxable income increases relative to accounts (expense reduced for tax purposes), and Article 3.2.3 requires the counterparty to include “less income” in its GloBE computation to avoid double taxation under the GloBE Rules.

26. The EUR 1 million immateriality election in Article 4.6.1 is an election to treat an immaterial decrease in Covered Taxes as a current-year adjustment. It presupposes a decrease in Covered Taxes for a prior year.

Where (as here) UAE corporate tax is not reduced in any year (no prior-year decrease and no current-year decrease), Article 4.6.1’s decrease-recalculation regime is simply not triggered, and the timing therefore remains prospective regardless of whether the underlying pricing difference is large or small.

The UAE denominator correction (required to prevent GloBE double taxation under paragraph 102) is therefore absorbed in the year in which the post-return outcome is “taken into account” through the post-filing mechanics—Year 3 on these facts.

Period for GloBE only income adjustment

27. We have earlier established that the Model Rules and the Commentary create explicit bridges that are highly relevant to the period of recognition of post-return outcomes, but they do so asymmetrically.

  • Article 4.6 is the regime’s post-filing timing engine for Covered Taxes, and the definition of “Prior Period Errors …” in Article 10.1 is drafted to yield to that timing engine.
  • The Commentary to Article 3.2.3 expressly routes post-return transfer-pricing outcomes into Article 4.6.1
  • The Commentary to Article 4.6.1 articulates the commensurability policy: the linkage is intended to ensure that GloBE Income/Loss and the Covered Taxes “associated with such amount” enter the ETR computation for the same fiscal year to avoid distortion.

At the same time, the text does not provide a direct hinge between Article 3.2.3 and Article 3.2.1(h). The latter is designed for the accounting errors affecting P&Ls. That definition is difficult to extend to a pure book-to-tax adjustment under Article 3.2.3. The absence of a direct cross-reference therefore blocks the syllogism that is sometimes suggested in practice: if Article 4.6 does not apply, then Article 3.2.1(h) must determine the timing of a 3.2.3 adjustment.

Competing constructions on timing of post-return Article 3.2.3 adjustments

28. The period of recognition for a post-return Article 3.2.3 adjustment can be construed in three ways.

28.1. Variant 1 – “4.6 only periodisation”.

On this view, para 99’s instruction that post-filing transfer-price adjustments are taken into account “pursuant to Article 4.6.1” is read as a complete timing rule for TP outcomes:

  • where the local tax adjustment affects Covered Taxes, Article 4.6.1 governs year allocation (current-year treatment as a default, recomputation for the relevant prior year in the decrease case);
  • where the TP outcome does not change Covered Taxes in the jurisdiction (denominator-only change), the Model Rules simply do not supply a specific periodisation rule. In the absence of a specific reopening rule, the adjustment would be reflected prospectively in the fiscal year in which it is identified for Pillar Two reporting, unless and until administrative guidance prescribes a different allocation.

This construction is internally coherent, but it makes the period of recognition of the denominator turn entirely on whether a tax-side trigger exists, and it provides no principle-based explanation for why a pure income-side correction should be locked to the year of discovery rather than the year of the transaction.

28.2. Variant 2 – “Transaction-year purity”: substantive-year allocation.

On this view, an adjustment under Article 3.2.3 is inherently a correction to the measurement of the controlled transaction in the fiscal year in which the transaction occurred. The adjustment therefore belongs in the Year-1 GloBE base (denominator) because it is the Year-1 intercompany price that is being corrected to an arm’s-length amount. The conceptual attraction is that it treats the GloBE base as “transaction-year truth”: once the correct arm’s-length measurement is known, the Year-1 Pillar Two base should be computed as if the transaction had been priced correctly in Year 1.

A practical virtue of Variant 2 is administrative stability. By anchoring the denominator to the transaction year, the group avoids repeated “re-allocations” of the base as the controversy evolves. If subsequent developments occur (e.g., MAP correlative relief, a revised settlement, or a later reversal) the tax-side effects can be dealt with under the applicable post-filing tax mechanics without repeatedly moving the denominator to a new fiscal year. In that sense, Variant 2 is designed to prevent the denominator from becoming a “floating” number that migrates between years depending on when the dispute is resolved.

The vulnerability of Variant 2 is not that it is conceptually incoherent, but that it can be administratively demanding: it assumes the regime tolerates retroactive recomputation of the Pillar Two base for the transaction year even where no explicit “reopen prior year” trigger is engaged on the Covered Taxes side. It therefore places weight on the proposition that Article 3.2.3 is itself an adjustment rule that speaks to the correct measurement for the year of the underlying transaction, rather than being merely a current-year reporting adjustment.

28.3. Variant 3 – “Systematic commensurability”.

On this view, the regime’s architecture reveals a general organising idea: corrections that must be brought into the Pillar Two base but are not reflected in the relevant prior year P&L should, by default, be recognised prospectively in the current year, unless a decrease in Covered Taxes triggers a prior-year recomputation mechanism. The argument is not that Article 3.2.1(h) governs Article 3.2.3, but that 3.2.1(h) and 4.6.1 together reveal an implicit periodisation policy (current-year recognition as a default; reopening only for defined decreases) that can be extended by analogy to denominator-only 3.2.3 adjustments.

Variant 3 has explanatory power in the “no tax change” scenario, but it can require repeated re-periodisation of the denominator as subsequent developments occur (for example, if the post-return outcome changes again through MAP or another relief tool), which Variant 2 avoids by fixing the denominator to the transaction year.

28.4. Variant 4 – “Sign-contingent periodisation”: transaction-year for negative corrections; current-year for positive corrections

A fourth construction periodises post-return Article 3.2.3 adjustments by reference to their direction and their capability to affect Covered Taxes if and when relief is obtained, using the post-filing mechanics as the organising template:

  • If the adjustment is negative (i.e., it reduces the counterparty’s GloBE Income and is of a kind that, in a correlative relief scenario, would tend to decrease Covered Taxes for the transaction year), the adjustment is recognised in the transaction year. This mirrors the logic of Article 4.6.1’s decrease branch: decreases are the paradigm case for reopening and recomputation.
  • If the adjustment is positive (i.e., it increases GloBE Income and corresponds to an increase in taxable income), the adjustment is recognised prospectively in the current year. This mirrors the general post-filing design for increases, which are typically taken into account in the year recorded rather than by reopening the transaction year.

Variant 4’s attraction is that it is direction-sensitive and therefore entirely reduces the “flip-flop” risk that can arise under a pure current-year default. It attempts to allocate the denominator in a way that anticipates how the numerator would be handled if relief is later granted or denied, thereby preserving commensurability more predictably.

28.5. Variants 2, 3, and 4 reflect different answers to the same question: whether Article 3.2.3 adjustments should be treated:

  • as substantive transaction-year remeasurements (Variant 2),
  • as current-year reporting corrections unless a tax-side reopening trigger exists (Variant 3), or
  • as direction-sensitive allocations that track the post-filing architecture ex ante (Variant 4).

Variant 2 maximises conceptual purity and administrative stability of the denominator by fixing it to the transaction year, but it presupposes that retroactive recomputation of the base is acceptable even absent a tax-side reopening event. Variant 3 maximises fidelity to the explicit post-filing architecture by using a current-year default and allowing prior-year movement only when Article 4.6.1 compels it. Variant 4 seeks to reduce the practical instability of other Variants by aligning periodisation with the sign of the adjustment and the post-filing treatment that would naturally follow if relief later materialises.

29. The four variants lead to the following UAE outcomes in the no-relief path.

29.1. Variant 1 (4.6 only periodisation) treats paragraph 99’s routing (“taken into account pursuant to Article 4.6.1”) as the only explicit timing logic for post-return outcomes. Since there is no UAE Covered Taxes adjustment (no increase, no decrease), Article 4.6.1 is not engaged on the UAE side. The Model Rules also do not provide a separate “reopen transaction year” mechanism for denominator-only 3.2.3 adjustments. On this narrow reading, the adjustment is therefore reflected prospectively in the fiscal year in which the post-return outcome is identified for Pillar Two reporting (here, Year 3) because nothing in the rules directs a recomputation of Year 1 for UAE.

If this variant is elected, UAE Co reduces DMTT Income by 20 in Year 3 with UAE Covered Taxes unchanged.

29.2. Variant 2 (transaction-year purity) treats Article 3.2.3 as a substantive correction to the measurement of the Year-1 controlled transaction. The arm’s-length price implied by the adjustment is 80. Thus, UAE’s Year-1 service revenue should be treated as 80 for GloBE purposes. Because the adjustment corrects the transaction year, it is recognised in Year 1, even though UAE corporate tax and the financial accounts remain at 100. The virtue is conceptual purity: the UAE GloBE base reflects the “transaction-year truth” and does not migrate between years depending on when the dispute is resolved.

Electing this variant results in a reduction of the denominator (Pillar Two Income) by 20 in Year 1, with UAE Covered Taxes unchanged. All else equal, the jurisdictional ETR for Year 1 increases mechanically (the numerator is unchanged while the denominator is smaller). If the UAE is in an under-minimum position for Year 1, that increase in the ETR correspondingly reduces the Top-up Tax (DMTT) for that year. 

29.3. Variant 3 (systematic commensurability) treats the 3.2.1(h)/4.6.1 architecture as evidence of a general periodisation pattern: current-year recognition is the default, and prior-year recomputation occurs only when the post-filing mechanism is triggered by a decrease requiring recomputation. In the no-relief path there is no UAE Covered Taxes decrease and therefore no UAE reopening trigger. Variant 3 therefore assigns the denominator-only adjustment to the current year, i.e., Year 3, without trying to pull it back into Year 1.

Result will be the same as in Variant 1: UAE Pillar Two Income (denominator) is reduced by 20 in Year 3, while UAE Covered Taxes remain unchanged. The jurisdictional ETR for Year 3 increases.

29.4. Variant 4 (sign-contingent periodisation) allocates timing based on the adjustment’s direction and its coherence with the post-filing architecture. The UAE symmetry adjustment is negative as it reduces UAE GloBE Income. On Variant 4, negative corrections are treated as transaction-year items because they are of the kind that, if relief were later granted, would naturally align with a “recomputing decrease” logic. Even though relief is not granted here, Variant 4 anchors the denominator adjustment in Year 1 to avoid later “flip-flop” if relief is subsequently obtained.

The result is the same as Variant 2. UAE Pillar Two Income (denominator) is reduced by 20 in Year 1, while UAE Covered Taxes remain unchanged. All else equal, the jurisdictional ETR for Year 1 increases, and  the DMTT for Year 1 correspondingly decreases.

Materiality in the “no UAE relief” path: how it interacts with Variants 1–4

30. In the “no UAE relief” path, UAE Covered Taxes do not change, so the EUR 1 million immateriality election in Article 4.6.1 is not directly engaged. Materiality therefore matters differently across the four variants.

30.1. For Variant 1 (textualist current-year default), materiality is irrelevant. The adjustment is recognised in the current fiscal year because, in the absence of an express reopening mechanism for a denominator-only correction, the computation proceeds prospectively. The size of the adjustment does not alter the year allocation.

30.2. For Variant 2 (transaction-year anchoring), materiality is likewise irrelevant. The adjustment is treated as a correction to the Year-1 measurement of the controlled transaction and is therefore recognised in Year 1. The size of the adjustment does not affect the year-of-recognition rule.

30.3. For Variant 3 (systematic commensurability), materiality is relevant in principle because the approach is built around the Article 3.2.1(h) and Article 4.6.1 dynamic with current-year recognition as a default, with prior-year recomputation only where a decrease in Covered Taxes triggers Article 4.6.1. The EUR 1 million election determining whether the decrease is dealt with prospectively. In the present no-relief path, however, there is no Covered Taxes decrease at all. Accordingly, the Article 4.6.1 “material vs immaterial decrease” machinery never comes into operation, and materiality becomes practically irrelevant.

30.4. For Variant 4 (sign-contingent), materiality is relevant as a design choice because this variant deliberately mirrors the Article 4.6.1 logic ex ante. A coherent implementation of Variant 4 can therefore import the EUR 1 million concept as an internal threshold: negative denominator corrections that would correspond to an immaterial decrease (if relief were granted) may be recognised prospectively, while material negative corrections are anchored to the transaction year. If no such threshold is adopted, Variant 4 (in the no-relief path) produces the same result as Variant 2 regardless of magnitude.

Conclusion for the no-relief path (UAE)

31. In the UAE no-relief path, the regime supplies a strong basis for the existence of a denominator-only symmetry adjustment (to avoid the paragraph-102 double-taxation concern), but it is less explicit on the fiscal year of recognition because there is no UAE Covered Taxes adjustment to activate Article 4.6.1 mechanics.

32. The interpretive choice therefore reduces to whether one prioritises (i) transaction-year measurement truth (Variants 2 and 4: Year 1) or (ii) a current-year reporting default in the absence of a reopening trigger (Variants 1 and 3: Year 3). Variant 4 has the additional attraction that it pre-commits negative corrections to transaction-year treatment, reducing the risk of later re-periodisation if relief is subsequently obtained.

[1] IAS 1.8.

[2] IAS 1.10(b).

[3] IAS 82(a).

[4] OECD Commentary to Article 4.6 of the GloBE Model Rules, para 119.

[5] GloBE Model Rules, Article 10.1.

[6] UAE Cabinet Decision No (142) of 31 December 2024 on the Imposition of Top-up Tax on Multinational Enterprises mirrors Articles 3.2.3, 3.2.1(h) and 4.6.1 rules.

The disclaimer

Pursuant to the MoF’s press-release issued on 19 May 2023 “a number of posts circulating on social media and other platforms that are issued by private parties, contain inaccurate and unreliable interpretations and analyses of Corporate Tax”.

The Ministry issued a reminder that official sources of information on Federal Taxes in the UAE are the MoF and FTA only. Therefore, analyses that are not based on official publications by the MoF and FTA, or have not been commissioned by them, are unreliable and may contain misleading interpretations of the law. See the full press release here.

You should factor this in when dealing with this article as well. It is not commissioned by the MoF or FTA. The interpretation, conclusions, proposals, surmises, guesswork, etc., it comprises have the status of the author’s opinion only. Furthermore, it is not legal or tax advice. Like any human job, it may contain inaccuracies and mistakes that I have tried my best to avoid. If you find any inaccuracies or errors, please let me know so that I can make corrections.