Permanent Establishment: How Remote Workers Create Tax Risk

By Finance Fundamentals Staff | Reviewed for technical accuracy by Hannah Morgan, CPA, Finance Fundamentals editorial team.

This is an international topic. Permanent establishment rules differ by country and by the specific tax treaty, if any, between the two countries involved. Nothing here substitutes for advice from a tax professional licensed in the relevant jurisdictions.

A Remote Hire Can Quietly Turn Into a Foreign Tax Return

A mid-size software company in Ohio lets its best support engineer move to Lisbon. Nothing about the job changes. Same laptop, same Slack channels, same manager. Eighteen months later, a Portuguese tax inspector asks a question nobody had prepared for. Does this employee’s home office count as a place of business in Portugal, and if so, is the U.S. parent now liable for corporate income tax there on part of its profits?

That question is not hypothetical anywhere near as often as most finance teams assume. Permanent establishment, or PE, is the concept tax authorities use to decide whether a foreign company has enough of a presence in their country to owe corporate tax there. For decades PE was mostly a worry for companies that opened branch offices, ran factories, or sent sales teams to set up shop abroad. Remote work broke that assumption. An employee can create the same exposure from a spare bedroom.

Two Tests, One Country

The legal foundation is Article 5 of the OECD Model Tax Convention, the template that underpins most of the world’s roughly 3,000 bilateral tax treaties. It defines a permanent establishment as “a fixed place of business through which the business of an enterprise is wholly or partly carried on.” It separately treats certain dependent agents as creating a PE, even without any physical location at all. Two different tests, two different failure modes, and most companies are only watching for one of them.

The Fixed-Place-of-Business Test

The first and older PE test asks a simple question with a complicated answer: does the company have a fixed place of business in the other country, and is business actually carried on through it? Traditionally this meant offices, branches, factories, workshops, or mines. A home office did not fit neatly into that list, and for years most practitioners treated home-based work as low risk by default.

That assumption has been eroding for a decade. Tax authorities in several European countries have argued that a home office is “at the disposal of” the employer if the work performed there is core to the business. That argument holds even when the employer never signed a lease or reimbursed the internet bill.

A few things tend to matter most: how long the arrangement lasts, whether the work is central to the business rather than incidental, and whether the company effectively depends on that location. A separate factor carries weight too: how the employee’s actual conduct compares with the job description on file.

Denmark’s Home-Office Ruling

Denmark has been the most aggressive enforcer. In a 2024 ruling, SKM2024.432.SR, the Danish Tax Council found that a foreign company had a permanent establishment in Denmark simply because a senior employee performed core management functions from a home office there. There was no separate office, no local staff, and no client-facing activity in the country at all. The ruling did not require a lease, a sign on a door, or a single customer meeting on Danish soil.

The OECD’s New 50% Benchmark

The OECD moved to bring more structure to this in its November 19, 2025 update to the Model Tax Convention Commentary. The new guidance sets a threshold. A home office generally will not count as a fixed place of business if the employee works from it less than 50% of total working time, measured over any rolling 12-month period.

Cross that line, and a second question kicks in. Is there a genuine commercial reason for the employee working from that location, separate from convenience or cost savings?

The Commentary is explicit on this point. Letting someone work remotely purely to retain talent or cut office rent does not count as a commercial reason. Meeting local clients, coordinating with suppliers, or covering a time zone for live customer support does.

That 50% test is new, and it narrows the fixed-place test considerably compared with the vaguer facts-and-circumstances approach tax authorities used through the early 2020s. A revised full treaty text incorporating the update was expected to publish within a few months of the release. Several countries, including India, Israel, Nigeria, and Malaysia, have already flagged reservations about applying it uniformly. Treat the 50% line as the current OECD benchmark, not as a guarantee your local tax authority will apply it the same way.

The Dependent-Agent Test

The second test does not care where anyone sits. Under Article 5(5) and 5(6), a company can have a PE in a country if it has a person there, employee or otherwise, who habitually concludes contracts on the company’s behalf. The same is true of someone who habitually plays the principal role leading to contracts that the company then routinely approves without material modification. No office. No lease. No fixed location required at all.

This is the test that catches remote sales and business-development hires. A rep working from an apartment in Barcelona who negotiates pricing, gets verbal agreement from the customer, and sends the deal to headquarters for a rubber-stamp signature can trigger a dependent-agent PE. That is true even though headquarters technically signs every contract. Tax authorities increasingly look past the signature line to where the substance of the deal actually happened.

What the Data Says About Enforcement

A cross-border compliance survey of tax practitioners found that 84% of surveyed jurisdictions apply dependent-agent PE concepts drawn from the OECD framework in their domestic law. Most avoid rigid bright-line rules in favor of a qualitative test built around repetition, continuity, and how central the agent’s role is to closing the deal.

In that same survey, 78% of respondents said the location where a contract is formally signed still matters. But 73% said where the terms were actually negotiated carries more weight, and 63% flagged time spent in-country as relevant to whether the activity counts as habitual.

Two things make this test harder to manage than the fixed-place test. There is no working-time percentage to check against, and job titles do not settle the question. A “customer success manager” who spends calls talking a client into an upsell and getting a verbal yes is functionally negotiating, regardless of what the offer letter says.

Treaty Exceptions and the Thresholds That Actually Apply

Not every foreign presence creates a PE, even under the broader dependent-agent rule. Article 5(4) of the Model Convention carves out activities that are “preparatory or auxiliary,” such as storage, display, purchasing, or pure information-gathering, as long as that is genuinely all the person or location does. A remote employee who only drafts internal reports, without any client contact or deal involvement, sits closer to this exception. That line between “auxiliary” and “core” gets blurry fast, though, once that employee talks to customers at all.

An independent-agent exception also exists. The person representing the company abroad might instead be a genuinely independent contractor or agency, acting in the ordinary course of its own business rather than economically dependent on one principal. In that case, no PE is generally created for the company being represented.

This is why some businesses route early-stage foreign hiring through an employer-of-record or a genuinely independent local distributor instead of a direct employee. The independence has to be real, though. A contractor who works exclusively for one company and takes instructions like an employee will not hold up under scrutiny.

Treaties Are Not Uniform

Specific bilateral treaties can also shift the numbers. Some treaties set a minimum duration, often around six months for a construction or installation PE under Article 5(3)-style clauses, though that provision is narrower and mostly irrelevant to office-based remote work.

What matters more for a typical remote hire is that treaty language, and the domestic law behind it, is not identical from one treaty to the next. The OECD Model is a template, not the law itself. Individual treaties, and countries without a treaty at all, can and do deviate from it.

One judgment call worth stating plainly: companies that treat the OECD’s 50% figure as a hard legal ceiling everywhere are taking on more risk than they realize. It is the OECD’s own benchmark for its own Commentary, adopted by member countries at different speeds and, in a handful of cases, rejected outright. Treat it as a strong signal, not a universal shield.

How Tax Authorities Are Actually Enforcing This

Enforcement has not caught up with remote work’s growth, but it is moving. A 2026 practitioner survey found that 30% of respondents saw tax authorities paying increased attention specifically to cross-border remote work arrangements, up from a much smaller share a few years earlier. Most respondents, even so, still said PE risk from remote workers was not yet a top enforcement priority. Every European jurisdiction surveyed reported that PE non-compliance carries administrative penalties, and several allow for criminal liability in cases of intentional evasion.

Employers are also watching themselves more closely than they used to, which is its own kind of evidence that the risk is real. EY’s global mobility research found that 98% of employers now track domestic and international employee movement in some form, compared with 49% in 2023.

That jump reflects payroll and immigration concerns as much as PE risk. But PE exposure sits right next to both on the same compliance checklist, and the same tracking gaps that create payroll tax problems create PE problems.

Headcount Changes the Calculus

Our companion piece on payroll tax for distributed teams covers the withholding side of that same employee. That is usually the first compliance gap a company discovers, often well before anyone asks the PE question.

Scale matters too. Research from the UK’s Office of Tax Simplification found that most multinationals have 2% to 5% of eligible employees working temporarily in a country other than their contracted location. Some large employers reported figures closer to 10%.

A separate industry guide points to 15 to 20 employees in a single foreign country as the rough threshold where companies typically stop relying on an employer-of-record and set up a local entity instead. PE risk becomes harder to argue away once headcount and local activity both grow.

A Worked Scenario: One Sales Rep, One Tax Bill

Year One: Low Risk

Consider a 40-person U.S. software company selling into Europe. It hires a single account executive who relocates to Munich for family reasons and keeps her U.S. job. For the first year she works from her apartment, answers email, and joins video calls with prospects. No German clients sign anything without a lawyer on the U.S. side reviewing terms first. Risk is low.

Year Two: Everything Changes

In year two, growth pressure changes her role. She starts running the full sales cycle for German-speaking prospects: discovery calls, pricing negotiation, verbal agreement, and a contract that headquarters signs within a day of her sending it over, almost never with changes. She works from that apartment nearly full time, well past the OECD’s 50% threshold, and the reason she is there is now squarely commercial: covering the German market in German business hours, not personal convenience.

At that point the company likely has exposure under both tests at once. The apartment may qualify as a fixed place of business given her time there and the commercial rationale for it. She personally may also qualify as a dependent agent, given how routinely her negotiations turn directly into signed deals.

A German tax authority reviewing the arrangement would have two independent arguments to work with, not one. The company’s U.S. tax and legal teams had no reason to flag any of this. Nothing about her employment contract, payroll setup, or job title changed when her role evolved.

Risk Factors at a Glance

The table below maps common remote-work arrangements against rough PE exposure levels. These are general patterns drawn from OECD guidance and practitioner experience, not a substitute for a fact-specific review.

Comparing Exposure by Scenario

The chart below lines up four common arrangements against the OECD’s 50% working-time benchmark, shown as the dashed line. Bars past that line do not automatically mean a PE exists, but they mean the fixed-place test is in play and the commercial-reason question needs an answer.

International tax advisors have been saying versions of this for years: the OECD’s newer thresholds add welcome precision, but they resolve the fixed-place question far better than the dependent-agent question. A company can pass the 50% test entirely and still trip the second one on a single employee’s sales calls.

Common Mistakes Companies Make

Most PE exposure is not the result of aggressive tax planning gone wrong. It is the result of nobody in HR, payroll, or tax connecting a routine staffing decision to a tax consequence three departments away. A few patterns show up repeatedly.

• Treating “remote” as a single category. A support engineer answering tickets and a business-development lead closing deals carry very different PE profiles, even if both report to the same manager and sit in the same time zone. Lumping them into one policy misses the distinction that matters to a tax authority.
• Assuming the payroll fix covers the tax exposure. Registering for local payroll withholding, often required well before PE risk arises, feels like compliance. It addresses individual income tax and social security. It does nothing for corporate tax nexus. Our companion guide on payroll tax for distributed teams goes into that gap in more detail.
• Letting job titles drift without updating the risk review. An employee hired into a “customer success” role often ends up negotiating renewal pricing within a year. Nobody revisits the PE analysis when that happens, because nothing in the org chart signals that the job changed.
• Assuming a written policy protects the company. A contract that says the employee has no authority to negotiate helps, but tax authorities look at what actually happens on sales calls, not what the offer letter says. If conduct and paper disagree, conduct tends to win in an audit.

A Practical Risk-Mitigation Checklist

None of this requires a legal department the size of a mid-tier accounting firm. It requires a short list of habits, done consistently.

• Track where each remote employee actually works, not just where payroll thinks they live, and revisit that list at least twice a year.
• For any employee working from a foreign location more than occasionally, log the rough percentage of time spent there against the 50% benchmark.
• Separate “prepares and drafts” roles from “negotiates and closes” roles explicitly, in writing, and confirm the distinction matches real behavior on calls.
• Route contract signature and final approval through a country without PE exposure, and make sure that is genuinely where the substantive negotiation happens too.
• Review independent-contractor arrangements for real independence: multiple clients, their own equipment, and control over their own working methods.
• Loop in a local tax advisor before a remote hire’s role changes materially, not after a tax authority asks the first question.

Key Takeaways

FAQ

Can one remote employee really create a permanent establishment?

Yes. Neither PE test in the OECD Model Tax Convention requires a minimum number of employees. A single person working from a fixed home office past the relevant threshold, or a single person habitually negotiating deals on the company’s behalf, can each independently create PE exposure.

What is the OECD’s 50% rule for home offices?

It is part of the November 2025 update to the OECD Model Tax Convention Commentary. A home office generally does not count as a fixed place of business if the employee works from it less than 50% of total working time over any rolling 12-month period. Crossing that line does not automatically create a PE; it triggers a further question about whether there is a genuine commercial reason for the arrangement.

Does a written policy against negotiating contracts protect the company?

It helps, but it is not decisive on its own. Tax authorities generally look at what the employee actually does on calls and in email, not just what the job description says. A mismatch between stated authority and actual conduct tends to be resolved in the tax authority’s favor.

Is PE risk the same as payroll tax risk?

No. Payroll tax and social security obligations typically arise sooner and depend on where the employee physically works, regardless of PE status. PE is about whether the company itself owes corporate income tax in that country. A business can have payroll obligations without PE exposure, and in rarer cases the reverse. See our companion article on payroll tax for distributed teams for the payroll side.

Does an employer-of-record arrangement eliminate PE risk?

It reduces some forms of risk but does not eliminate it. An EOR can handle local employment and payroll compliance. That does not automatically clear the client company, though. If the underlying worker still habitually negotiates contracts or works from a fixed, commercially justified location on the client’s behalf, dependent-agent or fixed-place exposure can attach to the client company, not just the EOR.

How is this different from tax residency?

Tax residency usually concerns where an individual or a company itself is considered resident for tax purposes. Permanent establishment is narrower. It asks whether a company resident in one country has enough presence in a second country to owe tax there, on income connected to that presence, without becoming resident there itself.

References

• OECD, Model Tax Convention on Income and on Capital, Article 5 and Commentary, update released November 19, 2025.
• Danish Tax Council, binding ruling SKM2024.432.SR (2024).
• EY, global mobility and immigration tracking research, 2026, cited alongside 2023 baseline figures.
• UK Office of Tax Simplification, research on cross-border and hybrid working arrangements.
• Cross-border tax compliance practitioner survey on dependent-agent PE application, 2025-2026.
• Thomson Reuters, guidance on employer-of-record to local-entity transition thresholds, 2026.
• Deloitte and EY technical alerts summarizing the OECD’s 2025 remote-work Commentary update.