Case Study: The “Flip‑Flop” 50% Home‑Working Threshold

Facts

Co is a company resident in State R. Co provides professional services to customers in State S. Co employs E, a senior consultant, who is resident in State S. E performs Co’s service activities for customers located in State S. E works partly from Co’s office in State S and partly from E’s home in State S. Co does not own or lease E’s home.

For purposes of this case study, assume that all fixed‑place PE elements would be satisfied if (and to the extent that) E’s home is treated as a “place of business of the enterprise”, including that there is a commercial reason for Co to undertake its business activities in State S through E’s physical presence in State S, and that if E’s home were unavailable, Co would use other premises in State S.[2]

Co’s fiscal year is the calendar year ending 31 December 2026 (“FY2026”). Co tracks E’s working time by location. Over FY2026, the rolling twelve‑month percentage of E’s total working time performed from home “flip‑flops” around 50% as follows (illustrative):

Accordingly, within FY2026 there are multiple rolling twelve‑month periods that meet or exceed 50%, but the twelve‑month period ending on 31 December 2026 does not.

[2] OECD 2025 Update, Commentary on Article 5, paras 44.11–44.15

Question

How should the 50% screen in the OECD 2025 Commentary on Article 5 be applied where the threshold is crossed intermittently (in some rolling twelve‑month periods commencing or ending in FY2026, but not others), and what are the practical consequences for (i) recognising a fixed‑place PE and (ii) attributing profits to that PE?

Executive Summary

Based on the analysis set out below, we conclude as follows:

• The 2025 Article 5 home‑working language should be read as creating a safe harbour rather than a bright‑line rule. Where an employee works from home for less than 50% of their total working time over the course of any relevant twelve‑month period, the home “would generally not be considered” a place of business of the enterprise.
• The words “any twelve‑month period commencing or ending in the fiscal year concerned” import a rolling‑window approach. OECD Commentary on Article 15 uses the same drafting technique for the 183‑day test and confirms that all possible periods of twelve consecutive months must be considered, including overlapping periods.
• If the <50% safe harbour is treated as a relief, then it should be available only where the enterprise stays below 50% in all relevant rolling twelve‑month periods linked to the fiscal year. If any one such period meets or exceeds 50%, the safe harbour is not available for that fiscal year and the analysis moves to the “facts and circumstances” inquiry contemplated by para. 44.10.
• Intermittent crossing therefore should not be administered as a month‑by‑month (or day‑by‑day) PE “switch”. Once the safe harbour is unavailable, the PE question is addressed by reference to the underlying arrangement and whether, during the relevant period, the home is in substance at the disposal of the enterprise and used for carrying on the enterprise’s business with the necessary degree of permanence.
• In the present facts (where we assume all other fixed‑place PE elements would be satisfied if the home is treated as a place of business), Co should treat E’s home as constituting a fixed‑place PE in State S for entire FY2026 once the safe harbour is disapplied by any qualifying twelve‑month period crossing 50%. Tax exposure is then managed principally through profit attribution, which is required to reflect the functions performed, assets used and risks assumed through the PE. In a home‑working fact pattern, that typically turns on what E actually did in State S and for how long.

Analysis

We conclude so based on the analysis set forth below.

1. The 2025 Update frames home‑working through the continuous vs. incidental distinction. Where an individual only works from home intermittently and incidentally, the home “would not be considered” a place of business of the enterprise. Conversely, where the individual works from home on a more continuous basis over an extended period, whether the enterprise has a place of business at the home is determined on the basis of all relevant facts and circumstances.[1]

2. Within that framework, the Commentary inserts a quantitative screen. If the individual works from home (or other relevant place) for less than 50% of their total working time “over the course of any twelve‑month period commencing or ending in the fiscal year concerned”, the home/other place “would generally not be considered a place of business of the enterprise”.[2]

3. If the individual works from home (or other relevant place) for at least 50% of their total working time over the course of any such twelve‑month period, the Commentary does not say that a PE automatically exists. Instead, it states that whether the enterprise has a place of business at the home/other place “will be determined by the facts and circumstances”.[3]

4. This structure matters. The 50% threshold is not framed as a positive definition of a PE. It is a screen that (i) offers a general negative conclusion where <50% and (ii) pushes the analysis back to orthodox Article 5(1) factors where ≥50%.

 “Any twelve‑month period”: importing the Article 15 interpretive instruction

5. The interpretive difficulty arises from the phrase “any twelve‑month period commencing or ending in the fiscal year concerned”. Read literally, that phrase requires testing more than one reference period: multiple twelve‑month windows can commence or end in the same fiscal year, and such windows overlap.

6. OECD Commentary on Article 15 uses the same phrase in relation to the 183‑day condition in Article 15(2)(a). It explains that this requires considering “all possible periods of twelve consecutive months commencing or ending in the fiscal year concerned”, and provides an example where a twelve‑month period overlaps two fiscal years.[4] The Commentary further clarifies that the twelve‑month period used for the 183‑day test “may overlap two fiscal years”.[5]

7. The relevance for Article 5 is not that the legal consequences are identical:

• Article 15 is a source‑taxation condition in the employment‑income rule;
• Article 5 is an enterprise‑level nexus concept.

The relevance is methodological: the OECD has elsewhere instructed that the “any twelve‑month period” wording is a rolling‑window test that requires checking overlapping periods, not a single fiscal‑year average.

The “one‑time is enough” reading: why intermittent crossing disapplies the safe harbour

8. The 2025 Article 5 wording is asymmetrical. It gives a general negative conclusion only in the <50% case (“would generally not be considered a place of business”) but offers no comparable general positive conclusion in the ≥50% case (which merely sends the reader to “facts and circumstances”).[6]

9. That asymmetry is naturally read as a relief:

• If the enterprise stays below the 50% threshold, it may generally treat the home as not constituting a place of business;
• If the threshold is met, the enterprise loses the comfort of that relief and must undertake the ordinary Article 5(1) analysis.

10. Once the <50% rule is characterised as a relief, the “any twelve‑month period” drafting does the same work it does in Article 15: it defines the conditions for accessing the relief. In practical terms, the enterprise can rely on the safe harbour only if E works from home for less than 50% of total working time in every twelve‑month period that commences or ends in the fiscal year concerned.

11. Where the enterprise’s data show that at least one qualifying rolling twelve‑month period in FY2026 meets or exceeds 50% (e.g. the periods ending 31 March 2026 and 30 September 2026 in the table above), the safe harbour is not available for FY2026. The analysis therefore moves to the “facts and circumstances” inquiry contemplated by para. 44.10.

12. Importantly, this reading does not treat one day above 50% as “creating” a PE. It only removes the general presumption that the home is not a place of business, requiring a full analysis of the place‑of‑business and disposal criteria.

Does the PE “switch on and off”?

13. A literal daily application of rolling twelve‑month windows could produce a conceptual “flip‑flop”: today’s rolling period is ≥50% (safe harbour unavailable), tomorrow’s rolling period is <50% (safe harbour available), and so on. That is the administrability concern behind this case study.

14. Two features of the Commentary indicate that this is not how the rules are meant to operate in practice.

• First, the 2025 text emphasises continuous vs. intermittent or incidental working from home. The home is not a place of business where work from home is “intermittent and incidental”. The question becomes live where the employee works from home “on a more continuous basis over an extended period”.[7] This is language of overall arrangement, not daily toggles.
• Second, the Commentary notes that whether an enterprise has a place of business is determined by the facts and circumstances “applicable during a given period”.[8] That invites segmentation where facts materially change (e.g. the home ceases to be used, the arrangement ends, or the enterprise no longer has the place at its disposal), not where only the arithmetic of a rolling percentage changes.

15. On the assumed facts, the commercial and functional reason for Co to have E operating in State S is stable across FY2026, and (by assumption) Co would replace the home with other premises in State S if the home were unavailable.[9] That is the type of arrangement that, once the safe harbour is unavailable, supports treating the home as at the enterprise’s disposal and as a place through which Co’s business is carried on.

16. Therefore, once FY2026 includes at least one qualifying twelve‑month period in which E’s home‑working reaches 50% (and the safe harbour is thus unavailable), the better administrable approach is:

• assess PE status for FY2026 by reference to the underlying arrangement and the degree of permanence and business use, rather than recomputing PE status month‑by‑month; and
• let the profit attribution step reflect the varying intensity of E’s activities in State S during the year.

17. That approach avoids two unattractive outcomes:

• an on/off PE status that changes without any real change in business reality, and
• the view that the enterprise can “re‑enter” the safe harbour later in the same fiscal year after it has been disapplied by a qualifying twelve‑month period, despite the “any twelve‑month period” drafting designed to capture overlapping periods.

Practical consequences: profit attribution is the main ‘fairness valve’

18. Treating the PE analysis as stable across the fiscal year does not imply that State S may tax a fixed percentage of Co’s global profits. Article 7 limits State S’s taxing rights to profits that are attributable to the PE.

19. OECD Commentary on Article 7 states that the profits attributable to a PE are the profits the PE “might be expected to make if it were a separate and independent enterprise…”, taking into account the functions performed, assets used and risks assumed through the PE and through other parts of the enterprise.[10]

20. For a home‑working PE, the functional analysis will typically focus on the employee’s activities carried on through the home and the risks (assets) connected to those activities. Where the employee spends relatively little time working from home in State S, the functions performed through the home will typically be correspondingly limited, and profit attribution should reflect that limitation.

21. Practically, this means that a ‘one‑time is enough’ reading of the 50% screen is not necessarily harsh. It may widen the set of situations that require a PE analysis and compliance, but the amount of profit ultimately attributable to the PE can still be closely aligned to the employee’s actual work performed in State S (often proxied, in service‑business cases, by time sheets, chargeable hours, and location‑based allocation of relevant revenue and costs).

22. The administrable compliance message is therefore: once the safe harbour is unavailable for FY2026, Co should

• assume PE registration/reporting may be required in State S (subject to local law and the treaty),
• build a defensible functional record of what E did in State S and when, and
• attribute only the profits connected to those activities through the PE in accordance with Article 7 principles.

Conclusion

23. In an intermittent ‘flip‑flop’ fact pattern, the key interpretive move is to recognise the 2025 <50% rule as a safe harbour. Consistent with the OECD’s treatment of the same “any twelve‑month period” language in Article 15, the safe harbour should be tested across all overlapping twelve‑month periods commencing or ending in the fiscal year concerned. If any such period meets or exceeds 50%, the safe harbour is unavailable and the analysis turns on facts and circumstances.

24. Where (as assumed here) the enterprise’s commercial rationale and the other fixed‑place PE elements would be satisfied if the home is treated as a place of business, the enterprise should not administer PE status as switching on and off within the year simply because a rolling percentage crosses 50% in some windows but not others. Instead, the principal practical discipline should occur at the profit‑attribution stage: profits taxable in the source State should track the functions actually performed through the home‑working presence.

[1] OECD 2025 Update, Commentary on Article 5, para. 44.7.

[2] OECD 2025 Update, Commentary on Article 5, para. 44.8.

[3] OECD 2025 Update, Commentary on Article 5, para. 44.10.

[4] OECD Model 2017, Commentary on Article 15, para. 4.

[5] OECD Model 2017, Commentary on Article 15, para. 4.1.

[6] OECD 2025 Update, Commentary on Article 5, paras 44.8 and 44.10.

[7] OECD 2025 Update, Commentary on Article 5, para. 44.7.

[8] OECD 2025 Update, Commentary on Article 5, para. 44.3.

[9] Compare OECD 2025 Update, Commentary on Article 5, paras 44.11–44.15.

[10] OECD Model 2017, Commentary on Article 7, para. 15.

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 we have tried my best to avoid. If you find any inaccuracies or errors, please let us know so that we can make corrections.