Key Takeaways
- The 183-day threshold is not a universal safe harbour - permanent establishment can arise in 30-120 days under many bilateral treaties.
- The OECD's November 2025 update introduced a 50% working time benchmark and commercial reason test for fixed place of business PE - but left dependent agent PE unaddressed.
- CFOs need to assess who works abroad, what they do, and which treaty applies - not just count days.
The 183-Day Myth: Why Day-Counting Alone Fails
Here's a scenario that plays out more often than you'd think. Your VP of Sales spends two weeks working from a rented apartment in Lisbon. She takes a few client calls, signs off on a partnership agreement, and flies home. No big deal, right? She was there for 10 days - well under the magic 183-day number. This type of cross-border setup often appears alongside broader global HR structuring decisions.
Except there's no magic number. And your company may have just created a taxable presence in Portugal.
Permanent establishment - or PE - is the concept in international tax law that determines whether a foreign country can tax your company's profits. If your business triggers a PE in another jurisdiction, you're looking at corporate income tax, profit attribution obligations under Article 7 of the applicable tax treaty, potential payroll withholding requirements, and penalties for non-compliance. It's the single biggest tax risk hiding inside most work-from-anywhere policies.
The widespread belief that employees can work abroad for up to six months without creating PE exposure comes from a misunderstanding. The 183-day threshold that appears in many tax treaties relates to personal income tax residency, not corporate permanent establishment. These are fundamentally different concepts, and conflating them is where companies get into trouble.
Under the UN Model Tax Convention, a services PE can be triggered when employees furnish services in a foreign country for more than 183 days within any 12-month period. But many bilateral treaties set that bar even lower - some at 90 to 120 days. And for certain types of PE, there's no day threshold at all. It's purely about what the employee is doing.
That last point is worth sitting with. A study by Grant Thornton across 21 countries found that 85% of digital nomad visas provide no corporate tax exemption whatsoever. The visa lets your employee in the door - it doesn't protect your company from PE exposure. Mobility permission and employer compliance are separate tracks, as we also discuss in our Europe work visa guide.
Fixed Place of Business PE vs. Dependent Agent PE: The Distinction That Matters Most
This is where the conversation gets nuanced - and where most summaries of the 2025 OECD update fall short. There are two distinct ways your company can trigger PE, and they operate under completely different logic.
Fixed Place of Business PE (Article 5(1))
This is the traditional form of PE. It requires a physical location with sufficient permanence through which your business is carried on - think offices, branches, factories, or, increasingly, an employee's home.
The OECD's November 2025 update to the Model Tax Convention introduced a much-needed framework for assessing when a home office constitutes a fixed place of business. It works in two stages.
First, a time-based indicator. If an employee works less than 50% of their total working time from a foreign location over any 12-month period, that location generally won't be treated as a fixed place of business. This is effectively a safe harbour for short-term or occasional remote stints abroad.
Second, a commercial reason test. If the employee exceeds the 50% threshold, the OECD asks whether there's a genuine business reason for their presence in that country. Serving local clients, accessing regional markets, or providing on-the-ground services counts as a commercial reason. Working from the south of France because the employee prefers the weather does not - and neither does enabling remote work purely to retain talent or reduce office costs.
There's an important caveat here. If the employee is effectively the business - a founder, sole consultant, or primary operator - their home office is likely to be treated as the enterprise's place of business regardless of these tests. The OECD commentary makes this quite clear: the more central the individual is to the enterprise, the higher the scrutiny.
Dependent Agent PE (Article 5(5))
This is the one that catches people off guard. A dependent agent PE arises when someone acting on behalf of your enterprise habitually concludes contracts - or plays the principal role in getting contracts to the finish line - in a foreign jurisdiction. It doesn't matter whether your company has an office there. It doesn't matter how many days the person has been in the country.
And here's the critical gap: the 2025 OECD update did not revise dependent agent PE guidance at all. The new 50% benchmark and commercial reason test apply only to fixed place of business PE. For anyone with contract-signing authority - sales directors, regional managers, business development leads - the existing, stricter principles still govern. In many remote work scenarios involving revenue-generating roles, dependent agent PE actually presents the greater risk. This is one reason many teams evaluate EOR operating models without local entities before approving long-term remote arrangements.
The 2025 OECD update clarified when a home office creates a fixed place of business. But it left the dependent agent question - often the bigger risk for sales and leadership roles working abroad - completely untouched.
| Factor | Fixed Place of Business PE | Dependent Agent PE |
|---|---|---|
| Trigger | Location + permanence + business activity | Person + contract authority + habitual pattern |
| Day threshold | ~50% working time (2025 OECD) | None - entirely activity-based |
| 2025 OECD update | New two-part framework introduced | No changes - existing rules apply |
| Highest risk roles | Key executives, founders, sole operators | Sales directors, BD leads, anyone signing contracts |
| Mitigation | Limit time abroad, document personal reasons | Restrict contract authority, centralise signing |
The PE Risk Decision Tree: A Practical Framework for CFOs
Theory is useful. But when an employee requests to work from Bali for three months, you need a practical way to assess the risk quickly. We built this decision tree around the questions that actually determine PE exposure.
- Is the employee a director, officer, or founder? If yes, elevated risk under both PE types. Their activities receive heightened scrutiny from tax authorities. Seek specialist advice before approving.
- Does the employee have authority to conclude contracts on behalf of the company? If yes, dependent agent PE risk is high regardless of how many days they spend abroad. This is the most commonly overlooked trigger.
- Will they spend >=50% of working time in the foreign location over 12 months? If no, low fixed place of business PE risk under the 2025 OECD safe harbour. If yes, continue to the next question.
- Is there a commercial reason for their presence? Serving local clients, accessing local resources, providing on-site services? If no, likely no fixed place of business PE. If yes, PE risk is materially elevated.
- Do they have a local mailing address, co-working space, or long-term rental? Physical presence indicators strengthen "disposal" arguments by tax authorities. Document everything.
- Which bilateral tax treaty applies? Check whether it follows the OECD Model (higher PE threshold) or the UN Model (lower threshold, services PE at 183 days or less). Some treaties have unique provisions at 90-120 days.
- Is the employee creating intellectual property in the foreign jurisdiction? Some countries - Germany, notably - treat local IP creation as a PE trigger even when other criteria aren't met.
Download the PE Risk Decision Tree
Get the full printable framework - including country-specific treaty variations and a risk-scoring matrix for your HR and finance teams.
What CFOs Should Do Now
If your company has a work-from-anywhere policy - or even an informal culture of approving remote work requests on a case-by-case basis - here's what needs to happen.
Audit your policy against the 2025 OECD framework. Most existing policies were drafted before the November 2025 update. If yours still relies on a blanket day-count rule without distinguishing between fixed place of business and dependent agent risk, it needs updating.
Implement day-count tracking across all jurisdictions. You can't manage what you don't measure. This means monitoring not just the countries employees travel to, but the cumulative time spent there - including business trips, "workations," and informal relocations.
Restrict contract-signing authority for employees working abroad. This is the simplest, highest-impact mitigation for dependent agent PE. Centralise contract execution in your home jurisdiction and make it part of your travel approval process.
Map your bilateral treaty network. Identify which treaties have lower PE thresholds. Countries that have negotiated treaties based on the UN Model - often developing nations - tend to have more aggressive PE provisions. Treaties with services PE clauses at 90 or 120 days require extra vigilance.
Brief your board. If directors or senior executives regularly work from foreign jurisdictions, the PE exposure isn't theoretical. It's the kind of risk that deserves a line item in your compliance reporting.
Consider Employer of Record structures for high-risk jurisdictions where you have employees working on an ongoing basis. An EOR creates legal separation between your company and the local activities, reducing - though not eliminating - PE exposure.