What Triggers a Permanent Establishment for Tax

Summary

A permanent establishment (PE) - the threshold at which a foreign jurisdiction can tax an enterprise's business profits - arises under OECD Model Article 5 through a fixed place of business, a dependent agent who habitually concludes contracts, a construction project over 12 months, a service presence (commonly 183 days/12 months), or, since the November 2025 OECD update, remote work above a 50% working-time threshold combined with a genuine commercial reason. The 2025 Commentary (the first revision since 2017, carried into the revised 2026 Model edition) sets a sub-50% safe harbor and a two-part test for home-office PEs. PE rules are treaty-specific: India, Israel, Nigeria, and Malaysia have filed reservations, US treaties follow the US Model, 18+ countries levy unilateral Digital Services Taxes outside the framework, and Pillar One Amount A (EUR 20B revenue / 10% margin) remains unratified as of May 2026 because its Multilateral Convention has not opened for signature. Sub-national economic nexus ($100K-$1M+ receipts) applies independently of any treaty PE finding. [src1, src2, src5, src7]

Rule

A permanent establishment (PE) arises when an enterprise has a fixed place of business through which its business is wholly or partly carried on in a foreign jurisdiction, or when a dependent agent habitually exercises authority to conclude contracts on the enterprise's behalf. Under the November 2025 OECD Model Tax Convention update, remote workers who spend less than 50% of their working time in a foreign jurisdiction over any 12-month period generally do not create a PE (safe harbor), but those exceeding 50% trigger a commercial reason test that examines whether the employee's physical presence serves a genuine business purpose beyond personal convenience. [src1, src2]

Evidence

The November 2025 update is the first revision to the OECD Model Tax Convention and its Commentary since 2017, and its remote-work changes will be carried into the revised 2026 edition of the Model. [src7] The OECD's 2025 update to the Commentary on Article 5 introduced five illustrative examples: a three-month holiday rental arrangement lacks permanence (no PE); 30% home-based work falls below the 50% threshold (no PE); 80% home work with regular customer visits and commercial justification (PE exists); 60% home work with only quarterly customer visits (insufficient commercial reason, no PE); and near-exclusive home use facilitating real-time global services (PE exists). [src3] As of 2026, 18 countries have implemented unilateral Digital Services Taxes targeting digital businesses based on significant economic presence rather than physical PE. [src5] Construction or installation projects constitute a PE only if they last more than 12 months under the OECD Model, though bilateral treaties may specify shorter periods. [src1]

Key Properties

• OECD Article 5 three-part test: (1) a place of business exists, (2) it is fixed (established with a degree of permanence), and (3) the enterprise carries on business through it
• Remote work safe harbor: Less than 50% working time in a foreign jurisdiction over 12 months = generally no PE
• Commercial reason test: When the 50% threshold is exceeded, PE arises only if the employer has a genuine commercial reason for activities at that location (not merely employee convenience, talent retention, or cost savings)
• Construction PE threshold: 12 months under OECD Model; bilateral treaties may specify 6 months
• Service PE threshold: Typically 183 days within any 12-month period (treaty-dependent)
• Pillar One Amount A scope: Applies only to groups with revenue exceeding EUR 20 billion and pre-tax profit margins above 10%; its Multilateral Convention has not opened for signature and requires ratification by 30+ jurisdictions covering 60%+ of in-scope ultimate parent entities before entry into force - not yet ratified as of May 2026

Conditions

• Applies when: An enterprise has employees, agents, offices, construction projects, or significant digital presence in a foreign jurisdiction AND a tax treaty (or domestic law) with PE provisions governs the cross-border activity
• Does NOT apply when: Activities are purely preparatory or auxiliary (Article 5(4) exemptions - e.g., warehousing for delivery, purchasing goods, collecting information); the enterprise only uses an independent agent acting in the ordinary course of their business; or the individual works below the 50% remote work threshold with no commercial reason
• Confidence degrades when: Treaty text differs from OECD Model (many older treaties lack updated remote work commentary); country has filed reservations to the 2025 update (India, Israel, Nigeria, Malaysia); or digital PE / DST rules create separate unilateral obligations outside the treaty framework

Constraints

• PE definitions are treaty-specific - the OECD Model is a template, not law; each bilateral treaty may differ in thresholds, exceptions, and definitions [src1]
• The 2025 OECD Commentary update is interpretive guidance, not binding law - countries must adopt it into their treaty practice for it to apply [src1, src6]
• US tax treaties follow the US Model Treaty - OECD Commentary is non-binding and applies only when Treasury Technical Explanations are silent or both nations expressly adopt OECD interpretations [src6]
• India, Israel, Nigeria, and Malaysia have expressed reservations to the 2025 remote work PE guidance - local analysis required for these jurisdictions [src2]
• Pillar One Amount A remains unratified as of May 2026 - the Multilateral Convention has not opened for signature (US signature and ratification widely viewed as unlikely), so the EUR 20B / 10% profit margin threshold is not yet enforceable [src5, src7]
• US state/provincial economic nexus rules ($100K-$1M+ in receipts) operate independently of treaty PE - a finding of no treaty PE does not eliminate sub-national tax obligations [src5]

Rationale

Permanent establishment rules exist to prevent double taxation while ensuring that countries can tax business profits genuinely connected to economic activities within their borders. The 2025 OECD update responded to the post-COVID remote work explosion, which created widespread uncertainty about whether home offices could constitute PEs - the 50% safe harbor and commercial reason test balance the need for tax certainty with the principle that taxing rights follow genuine economic activity, not mere employee presence. [src1, src4]

Framework Selection Decision Tree

START - User needs cross-border tax presence guidance
+-- What type of foreign activity?
|   +-- Fixed office, branch, factory, warehouse
|   |   +-- Fixed Place PE (Article 5(1)-(3)) - standard analysis
|   +-- Employee or agent signing contracts abroad
|   |   +-- Dependent Agent PE (Article 5(5)-(6))
|   +-- Construction or installation project
|   |   +-- Construction PE (Article 5(3)) - 12-month threshold
|   +-- Remote workers / home office abroad
|   |   +-- Remote Work PE (2025 Commentary) <-- YOU ARE HERE
|   |       +-- Under 50% working time? -> SAFE HARBOR - no PE
|   |       +-- 50%+ working time?
|   |           +-- Commercial reason exists? -> PE likely
|   |           +-- No commercial reason? -> No PE
|   +-- Digital services, no physical presence
|   |   +-- Digital PE / DST (unilateral - 18+ countries)
|   +-- Service delivery over extended period
|       +-- Service PE - 183-day / 12-month threshold
+-- Is there a tax treaty?
|   +-- YES (OECD-based) -> Apply treaty PE definition
|   +-- YES (with reservations) -> Check country-specific rules
|   +-- NO -> Apply domestic PE / nexus rules
+-- Is sub-national nexus also relevant?
    +-- YES (US states, Canadian provinces) -> Separate analysis required
    +-- NO -> Treaty PE analysis is sufficient

Application Checklist

Step 1: Identify the activity type and jurisdiction pair

• Inputs needed: Home country of the enterprise, foreign jurisdiction where activity occurs, nature of activity (office, agent, construction, remote work, digital services)
• Output: Classification into one of six PE types (fixed place, dependent agent, construction, remote work, service, digital)
• Constraint: If no tax treaty exists between the two countries, skip to domestic PE rules - OECD Model analysis does not apply [src1]

Step 2: Check the applicable treaty and any reservations

• Inputs needed: Specific bilateral tax treaty text, any country reservations to the 2025 OECD Commentary
• Output: Applicable PE definition and thresholds (which may differ from the OECD Model defaults)
• Constraint: Do not assume OECD Model thresholds apply - India, Israel, Nigeria, Malaysia, and others have filed reservations; US treaties follow the US Model [src2, src6]

Step 3: Apply the relevant PE threshold test

• Inputs needed: Duration of activity (days/months), working time percentage (for remote work PE), contract authority scope (for agent PE), project timeline (for construction PE)
• Output: Preliminary PE / no-PE determination based on quantitative thresholds
• Constraint: For remote work PE, the 50% threshold is measured over actual conduct across any 12-month period, not contractual arrangements - tracking data required [src1, src3]

Step 4: Apply the commercial reason test (remote work PE only)

• Inputs needed: Business justification for employee's foreign presence, nature of local interactions (customer meetings, supplier management, time-zone coverage)
• Output: PE / no-PE determination for cases exceeding 50% threshold
• Constraint: Employee convenience, talent retention, and office cost savings are explicitly excluded as commercial reasons - only genuine business-operational needs qualify [src2, src3]

Step 5: Assess compliance obligations and escalation needs

• Inputs needed: PE determination from Steps 3-4, enterprise size, profit allocation method, local registration requirements
• Output: List of compliance obligations (tax registration, filing, profit allocation, VAT/GST) or confirmation of no PE
• Constraint: If PE is found, engage local tax counsel - profit allocation under transfer pricing rules and potential withholding obligations require jurisdiction-specific advice [src4, src5]

Decision Logic

If a remote employee spends under 50% of working time in the foreign jurisdiction over any 12-month period

--> Treat as a safe harbor: no PE under the 2025 OECD Commentary, provided the relevant treaty has adopted the update and the jurisdiction has filed no reservation. Still track actual time so the percentage is defensible. [src1, src2]

If a remote employee spends 50% or more of working time abroad but only for personal convenience, talent retention, or office cost savings

--> No PE on the commercial-reason limb: those motivations are explicitly excluded as commercial reasons. Document the absence of local client/supplier interaction. [src2, src3]

If a remote employee spends 50%+ of working time abroad AND regularly meets local clients, manages suppliers, or delivers real-time services there

--> PE likely: both limbs of the two-part test are met. Engage local tax counsel on registration and profit attribution. [src2, src3]

If the activity is a construction or installation project

--> Apply the 12-month OECD Model threshold (a PE arises only past 12 months), but check the specific treaty - many specify a shorter 6-month period. [src1, src5]

If an agent in the foreign country habitually concludes contracts in the enterprise's name, or acts exclusively/almost exclusively for it

--> Dependent-agent PE under Article 5(5)-(6): the independent-agent exemption does not apply. Exclusivity is the key trigger. [src1, src5]

If the foreign jurisdiction is India, Israel, Nigeria, or Malaysia (reservations to the 2025 guidance), or the home country is the US (US Model treaties)

--> Do not apply the OECD 50% safe harbor; run a jurisdiction-specific analysis under local PE rules or the US Model Treaty / Treasury Technical Explanations. [src2, src6]

If the business is digital with no physical presence but significant revenue/users in a market jurisdiction

--> Treaty PE likely does not arise, but screen for unilateral Digital Services Taxes (18+ countries) and significant-economic-presence rules, which sit outside the OECD Model. [src5]

If no treaty PE is found but the activity is in a US state, Canadian province, or Swiss canton

--> Run a separate sub-national economic-nexus analysis ($100K-$1M+ receipts thresholds) - it operates independently of treaty PE and often has lower thresholds. [src5]

Anti-Patterns

Wrong: Assuming the OECD 50% safe harbor applies globally

Companies rely on the 50% working time threshold as a universal rule, failing to check whether the specific bilateral treaty between the two countries has adopted the 2025 Commentary or whether the foreign jurisdiction has filed reservations. India, for example, does not accept the new tests. [src2]

Correct: Check the specific treaty and country reservations first

Before applying the 50% safe harbor, verify that both treaty partners recognize the 2025 OECD Commentary. For treaties with India, Israel, Nigeria, or Malaysia, conduct a jurisdiction-specific analysis using local PE rules. [src2, src6]

Wrong: Treating all remote work as PE-safe because it is "employee-driven"

Employers classify all work-from-abroad arrangements as employee convenience without examining whether activities serve a commercial purpose. An employee working 80% from a foreign country while regularly meeting local clients creates a PE regardless of who initiated the arrangement. [src3]

Correct: Apply the commercial reason test to activities, not motivations

Examine what the employee actually does at the foreign location. Regular customer engagement, supplier management, or real-time service delivery indicates commercial reason - even if the employee chose to relocate for personal reasons. [src3, src4]

Wrong: Ignoring sub-national nexus after confirming no treaty PE

A US-focused enterprise determines no PE exists under the US-UK treaty and stops the analysis. However, the employee's activities in New York trigger state economic nexus at the $1M receipts threshold, creating state-level filing obligations. [src5]

Correct: Always layer sub-national nexus analysis on top of treaty PE analysis

After completing the treaty PE assessment, separately evaluate state, provincial, or cantonal nexus rules - these operate independently and often have lower thresholds than treaty PE. [src5]

Counter-Arguments

• Some jurisdictions argue the 50% threshold is too generous and allows large enterprises to structure remote work arrangements just below the threshold to avoid PE - India's refusal to accept the new tests reflects this concern. [src2]
• Digital PE advocates argue that traditional physical-presence PE rules are obsolete for digital business models generating billions in revenue without any physical presence - the 18 countries with DSTs have already moved beyond the OECD framework. [src5]
• Small-country economists contend that PE rules systematically favor headquarter jurisdictions, as the "commercial reason" test and preparatory/auxiliary exemptions allow MNEs to concentrate taxable profits in low-tax home countries while extracting value from market jurisdictions. [src5]

Common Misconceptions

Misconception: A home office always creates a PE if the employee works there regularly.
Reality: Under the 2025 OECD update, working from home below 50% of total working time over 12 months generally does not create a PE. Even above 50%, a PE requires a genuine commercial reason - employee convenience alone is insufficient. [src1, src2]

Misconception: The 183-day rule is a universal PE threshold that applies to all activity types.
Reality: The 183-day threshold applies specifically to service PEs in certain treaties. Fixed-place PEs have no specific day count (permanence is assessed on facts and circumstances); construction PEs use a 12-month threshold; and remote work PEs use the 50% working-time test. Different PE types have different thresholds. [src1, src5]

Misconception: Using an independent agent in a foreign country eliminates PE risk entirely.
Reality: Independent agents acting in the ordinary course of their business generally do not create PE, but agents who act exclusively or almost exclusively for one enterprise may be reclassified as dependent agents under Article 5(6), creating PE risk. Exclusivity arrangements are a key trigger. [src1, src5]

Misconception: Pillar One / Amount A has replaced traditional PE rules for digital companies.
Reality: As of May 2026, the Multilateral Convention to implement Amount A has not even opened for signature, let alone been ratified - entry into force requires 30+ jurisdictions covering 60%+ of in-scope parent entities, and US ratification (a two-thirds Senate vote) is widely viewed as unlikely. Traditional PE rules remain the primary framework, supplemented by unilateral DSTs in 18+ countries. Amount A applies only to groups exceeding EUR 20B revenue and 10% profit margins. [src5, src7]

Comparison with Similar Rules

When This Matters

Fetch this rule when a user asks whether their business activities, employees, agents, construction projects, or digital services in a foreign country create a taxable permanent establishment - including questions about remote worker PE risk, the OECD 50% working time threshold, dependent agent PE, construction PE timelines, or digital PE / DST obligations.