Case study · Pillar Two · Saudi Arabia
The four‑day calculation
A Saudi-headquartered energy company, with 300+ entities and 25+ joint ventures in 30+ countries, met the GloBE Rules in spreadsheets. I led the work that moved the calculation into one system, and cut it from days to hours.
Time to produce the Pillar Two calculation
Working hours, drawn to scale
01 · The scale
Three hundred entities. Twenty-five joint ventures. Thirty-odd rulebooks.
Energy groups like this are built from project companies: one for each plant, often held with partners, often under a tax holiday a host country granted to win the investment.
Under Pillar Two, each of those entities feeds a calculation for its jurisdiction, and each incentive becomes a question about the effective tax rate. The 25+ joint ventures add a second layer: the rules compute each joint venture group on its own, apart from the rest of the group.
02 · The problem
The rules were new. The tool was Excel.
When the GloBE Rules arrived, the calculations lived in spreadsheets, and advice came from several Big 4 firms at once.
Producing one set of figures meant pulling data from hundreds of entities, re-keying adjustments and reconciling workbook versions, for days.
The desk as it was. Drag any file, email or note anywhere.
entity,account,description,amount SA-011,400100,Revenue,-48210554.12 SA-011,510200,Staff costs,6004211.40 SA-011,160000,PPE at cost,420118902.00 SA-011,890100,Income tax,6204118.75 SA-014,400100,Revenue,-9120044.00 SA-014,160500,Other assets,84002310.00
- From
- Adviser
- To
- Group Tax
- Subject
- RE: SBIE position on a class of assets
Hi team, our current reading supports including them, but further guidance may change that. We’d suggest modelling both positions until then.
Regards
Illustrative files, figures and notes
03 · The sessions
First, the rules had to make sense to the people running them.
The rules were new to the team, so the work started in the room. I led the Pillar Two sessions with a team of two to three, then turned what we learned into a data specification and a build.
- Step 1
Pillar Two sessions
The GloBE Rules, worked through with the tax team.
- Step 2
Data requirements
Every input the rules ask for, named and traced to a source system.
- Step 3
Technical sessions
The data and the mechanics, with the people who own the source data.
- Step 4
The build
The Pillar Two solution, on Oracle EPM Tax Reporting.
- Step 5
First run
The full calculation in hours.
What didn’t hold
One calculation was never going to be enough.
The rules were still being clarified, and the group had not settled its positions. A single answer would have been out of date the moment a decision changed.
So we built several scenarios of the calculation side by side, to show where each direction would take the group. The substance-based income exclusion moved the answer most.
Each position trades tax now against risk later. Pick one to see the trade.
The lowest top-up tax now. If a tax authority later takes a narrower view of which assets qualify, the carve-out shrinks on review, and the difference comes back with interest and penalties for every year already filed.
Illustrative figures. The walk-through below follows the full carve-out.
04 · One number
Follow one jurisdiction from its accounts to its top-up tax.
A simplified, illustrative calculation for one jurisdiction, in USD millions. Each line names the article of the GloBE Model Rules it comes from. For the detailed calculations on a fictional group, sign in to Ambit.
11.4%Annex A, Ch. 1P2-TCBCR-2026.1
First, the shortcut.
The transitional CbCR safe harbour can switch the full computation off. Tax expense of 6.20 against profit before tax of 54.20 gives a simplified ETR of 11.4%, against a 15% threshold for FY2024. The test is not met as computed, so the full computation applies.
48.00Art. 3.1.2P2-FANIL-2024.1
Start from the accounts.
Financial accounting net income of the jurisdiction's entities: 48.00.
+6.20Art. 3.2.1(a)P2-NTE-2024.1
Put the tax back.
Income tax expense of 6.20 is added back, because the rules measure tax against profit before it.
(2.40)Art. 3.2.1(b)P2-EXDIV-2024.1
Take out the dividends.
Dividends from other holdings leave the base: 2.40 under the excluded dividends rule.
51.80Art. 5.1.2P2-ETR-2024.1
One figure for the jurisdiction.
Every entity in the jurisdiction is blended into net GloBE income of 51.80.
5.00Art. 4.1P2-ACT-2024.1
Then the taxes.
Current tax of 4.10 and a deferred tax adjustment of 0.90 under Article 4.4 give adjusted covered taxes of 5.00.
9.65%Art. 5.1.1P2-ETR-2024.1
Divide.
5.00 over 51.80 is an effective tax rate of 9.65%, computed to four decimal places.
5.35%Art. 5.2.1P2-TUTP-2024.1
Measure the gap.
The distance to the 15% minimum rate is the top-up tax percentage: 5.35%.
(33.35)Art. 5.3, 9.2P2-SBIE-2024.1
Substance stays out.
Payroll of 6.00 and tangible assets of 420.00, at the 2024 transitional rates of 9.8% and 7.8%, exclude 33.35. For an asset-heavy energy group, this line does a lot of work.
0.99Art. 5.2.3P2-JTUT-2024.1
Top-up tax for the jurisdiction: 0.99.
5.35% of 18.45. One jurisdiction of thirty-odd, all computed in the same run.
05 · The build
One system where there were many workbooks.
The solution on Oracle EPM Tax Reporting takes the data the rules need from where it already lives, and runs every jurisdiction through the same steps.
Simplified. Source systems are described by type.
06 · The result
Four days became three hours.
The same calculation, for every entity and every jurisdiction, run inside the group’s own systems instead of across workbooks.
to produce the Pillar Two calculation
entities in one computation, on Oracle EPM Tax Reporting
jurisdictions blended and tested in the same run
07 · Try the test
Would a jurisdiction have needed the full computation?
Move the figures and see the transitional simplified ETR test respond, per rule P2-TCBCR-2026.1. This is the first gate every jurisdiction passes through.
USD millions. Based on the figures you supplied, with the transition rate as published for rule P2-TCBCR-2026.1. The full safe harbour also has a de minimis test and a routine profits test, which this sketch leaves out.
Does not meet the transitional simplified ETR test as computed: 11.4% against a 15% threshold for FY2024. The full GloBE computation applies.
See a full computation, step by step