TL;DR
- The trap: The cheapest-looking first move is engaging them as a contractor, and it's also the move most likely to be wrong.
- When to do nothing: If the work is genuinely a defined deliverable, short, and outside your core business, you may not need to hire at all.
- What has to be true: Someone in the business can answer the question "is this person doing our work, our way, on our schedule" honestly. If yes, contractor is a stretch in most jurisdictions.
- How options split: Contractor, employer of record, local entity, or remote hire into an existing foreign entity. Each has a specific shape of failure.
- The decision rule: Pick the structure that matches the work and the jurisdiction, not the one that matches the candidate's preferences or your timeline.
- Outcome to expect: The right answer takes longer than the candidate wants, costs more than the contractor quote, and saves multiples of both by month twelve.
The File on Your Desk
A staff engineer in Lisbon has been your contract backend developer for seven months. She ships the work, knows your codebase, knows your customers, and is the person your CTO wants leading the new platform team. Yesterday she asked, on a call, what the plan is for the next year. You have a contract. You don't have an entity in Portugal. You don't have a Portuguese payroll. The candidates you ran against her were weaker in every dimension that matters, and you already know the answer. You're reading this because the paperwork is the part you don't know.
It's tempting to keep the contract going, add a small monthly retainer, and call it solved. The work is going fine. The invoices are landing. The candidate is happy. The shape of the problem is that the work isn't the problem. The paperwork is. And the paperwork is the part that turns a clean contractor file into a reclassification claim, a back-tax assessment, or a permanent establishment finding that the founders didn't see coming.
The real issue isn't whether to hire this person. You have already decided. The real issue is the order in which the legal, tax, and operational decisions get made, because the cheapest route at the start is the one that's hardest to undo, and the route you choose at hour one will quietly determine what you can and can't do at month twelve.
Best tools for Global HR
When You Genuinely Do Not Need to Act Yet
Not every relationship with a foreign person needs to become employment. Some genuinely should not. Four stages describe where the reader actually sits, and the first is more common than the content industry admits.
Your current setup is genuinely fine. You have a person in another country doing a defined piece of work: a market study, a translation, a security review, a quarterly design audit. There's a deliverable. There's a deadline. There's no recurring schedule. There's no tooling account with their name on it. There's no expectation of ongoing work past the next invoice. If that's the file, the contractor route is the right one, and the only thing to do is make the contract specific about what is being delivered, when, and what counts as done. The mistake here's to convert a clean file into an employment relationship because the person is good. Good isn't a legal test.
Friction is starting. The deliverable is done, the work is continuing. The same person is in your standups. The same person has a laptop you issued. You pay them monthly, on the same day, the same amount, and the work has the same scope it had three months ago. The contractor framing is getting harder to defend on paper, and you know it. At this stage, the choice is still yours. You can convert them to an employee, end the relationship, or renegotiate the contract so the work genuinely is project-based and time-bounded. What you can't do is nothing, because the file is starting to look like employment in everything but name.
Real risk is now visible. A competitor has just hired from the same talent pool, which means the candidate is now in two recruiters' inboxes, and one of those recruiters will sell them on an EOR. The candidate has asked about stock options. The contractor invoice is now the largest line on a cost line you don't own. You have a runway question that depends on this person. At this stage, the contractor route is still legal in places, but the cost of the question has changed. The decision you've not made is now blocking decisions you need to make.
The edge case. The person is so senior, or so rare, that you would accept almost any structure to keep them. This is where most readers actually sit, and it's the most dangerous stage, because it inverts the decision. You stop choosing the structure that fits the work and start choosing the structure that fits the urgency. The shape of every misclassification case in every jurisdiction looks like this. The candidate is good. The work is core. The paperwork was not done. The tax authority sends the letter three years later. The structure was chosen by calendar pressure, not by the file, and the calendar pressure isn't a defence.
The Five Questions You Ask Yourself at 11pm
Are they doing our work, our way, on our schedule, and using our tools? Yes in most files. That's the entire common-law inquiry, and the IRS weighs all factors together as a totality of the circumstances, with no single factor decisive and no presumption either way. If the honest answer is yes, contractor is a stretch, and the test you fail in California is even tighter than the federal one.
Is their work outside the usual course of our business? This is the prong of the ABC test that founders forget exists, and it's the one that breaks their case most often. AB5 applies the ABC test in California and presumes a worker is an employee unless the hiring entity proves all three prongs: freedom from control, work outside the usual course of the hiring entity's business, and the worker being engaged in an independently established trade. A company can satisfy the IRS test and still fail the ABC test. If your backend contractor is doing the work your engineering team does, prong B is gone before you start.
Do I actually want them on payroll, or do I want this problem to go away? Those are different decisions. The EOR route makes the problem go away on a Wednesday. The right route makes the problem go away and the file defensible in three years. The first question isn't whether the candidate is good. The first question is whether the relationship should be employment at all, and if so, where.
What does this look like in their country, not mine? The classification rules you read about are mostly the rules where your company is. The payroll rules, the social contributions, the notice periods, the termination rules, the works council triggers, the data protection obligations, the equity tax treatment, are all where the worker sits. Your home-country intuition is the wrong one to reach for, and it's the one most founders reach for, because it's the only one they have.
What happens if the relationship ends badly? Every relationship ends eventually. Most end amicably. Some don't. The question to ask before signing isn't what happens if it works. The question is what the off-ramp looks like, what the notice period is, what severance is owed, and whether you've any of that in writing. If the answer is "we'll figure it out", you're betting your company on a guess about a foreign legal system you don't work in.
The Three Honest Categories the Approaches Split Into
Contractor. You engage the worker through a service agreement, they invoice you, they pay their own taxes, they have their own equipment, they set their own hours, they have other clients. The cost is low, the setup is fast, the paperwork is minimal, and you can walk away on thirty days. It's the right answer when the work is genuinely a deliverable, genuinely bounded, and genuinely outside your core. It's the wrong answer the moment the work looks like employment, because in most of the jurisdictions a founder cares about, the worker is treated as your employee regardless of what the contract says. The IRS weighs all factors together as a totality of the circumstances, with no single factor decisive and no presumption either way. California applies the ABC test, which presumes employee and forces you to prove all three prongs. A company can satisfy the IRS test and still fail the ABC test. The same misclassification risk is the failure mode in the UK, where the off-payroll rules determine who pays the employment taxes, and in the EU, where the Platform Work Directive shifts the burden of proof in defined platform cases. Contractor is the cleanest file when the file is genuinely contracting, and the riskiest file when it isn't.
Employer of record. A third party employs the worker through its own local entity, runs the local payroll, handles the statutory filings, and the worker is on their books while doing your work. The cost is a percentage of payroll, the setup is weeks rather than months, you don't incorporate, and the worker gets local employment protections on day one. It's the right answer when the relationship is genuinely employment, the country is one where you don't yet have presence, and the headcount in that country is one. It's the wrong answer when the worker is one of many, when the role is senior enough that you want direct control, when the activity in country goes beyond employing staff, or when the cost curve of the EOR fee starts to rival the cost of doing it properly. The EOR substantially reduces permanent establishment exposure but doesn't eliminate it. Activity beyond employing staff, such as concluding contracts locally, maintaining a fixed office, or holding inventory, can still create one, and an EOR doesn't answer corporate tax. The honest example: a single senior product manager in Germany on a multi-year assignment, with no other German headcount, no local office, and no locally-concluded customer contracts.
Local entity. You incorporate a company in the worker's country, you employ them through it, you run your own payroll, you file your own returns, and you accept the overhead of a local presence. The cost is high, the setup is months, the ongoing burden is real, and the file is fully defensible in any later review. It's the right answer when the headcount in country is growing past one or two, when the activity in country includes sales or customer-facing operations, when the role is senior enough to justify the structure, and when you've the operational maturity to run a second payroll. It's the wrong answer when you've one person in one country and no plan to add a second, because the overhead of a local entity is the same whether you've one employee or ten, and you'll pay for the entity for as long as it exists.
Five Diagnostic Questions You Can Self-Assess Against
Is this person's work part of how we make money? If yes, and you're operating in California, the ABC test will treat the worker as your employee unless you can prove all three prongs. Be honest about what you can prove. If the work is in your usual course of business, the second prong is already gone.
Could we defend the contractor relationship in writing? Pull the contract. Read it. Ask whether the document actually describes what is happening. If the contract says deliverables and milestones and the file shows monthly retainers and weekly standups, you already know the answer, and so will anyone who reads the file later.
Do we've a real reason to want this person on payroll, or a real reason to want this problem gone? "On payroll" is a structure. "Problem gone" is a feeling. They're not the same decision, and conflating them is the move that creates the misclassification file.
What is our twelve-month plan for the country? One person is a hire. Two is a pattern. Three is a footprint. If you don't know, the EOR is the right starting point, because it lets you convert to an entity later when the pattern is real, rather than incorporating on the back of a guess.
Who in this business is going to own this file in eighteen months? Not who today. Who in eighteen months. Because the question you're answering now isn't "is this person employed today". The question is "is this file defensible when the next people director opens it".
Four Routes for a First Hire, Reviewed
Engaging Them as a Contractor
You sign a service agreement. They invoice you. They use their own equipment. They set their own schedule. The file is light, the cost is low, the off-ramp is clean. The IRS weighs all factors together as a totality of the circumstances, with no single factor decisive. California applies the ABC test, which presumes employee and forces the hiring entity to prove all three prongs. A company can satisfy the IRS test and still fail the ABC test. The UK off-payroll rules determine who carries the employment tax liability, with the small end-user threshold rising from 10.2 million pounds to 15 million pounds at the start of the financial year, and a separate change making agencies and end clients potentially liable for PAYE underpayments supplied through umbrella companies. The same misclassification risk exists in most of the jurisdictions a founder cares about. Contractor earns its place when the work is a defined deliverable, short, and outside the hiring entity's core. It fails when the work is ongoing, integrated, and on the same schedule as your team, which is the file most readers actually have.
An Employer of Record
An EOR employs the worker through its own local entity, running local payroll, statutory filings, and employee-level obligations, which removes the need to incorporate. Setup is weeks. The fee is a percentage of payroll. The worker is locally employed on day one with the statutory protections that brings. The EOR substantially reduces permanent establishment exposure but doesn't eliminate it. Activity beyond employing staff, such as concluding contracts locally, maintaining a fixed office, or holding inventory, can still create one, and an EOR doesn't answer corporate tax. It earns its place when the relationship is genuinely employment, the headcount is one, and you don't yet have a footprint in the country. It fails when the role is senior enough that you want direct employment terms, when the cost curve of the fee outpaces the cost of a local entity, and when the activity in country expands past employing staff.
Incorporating a Local Entity
You set up a company in the worker's country, register with the local authorities, open a local bank account, and employ the worker through it. You run the payroll. You file the returns. You accept the ongoing cost of a second legal entity. The file is fully defensible. The structure scales. The cost is months of setup, real legal fees, and ongoing accounting overhead. It earns its place when you've more than one or two people in the country, when the role is senior, when the activity in country goes beyond employing staff, or when you've a real twelve-month plan for the market. It fails when you've a single hire and no plan to add a second, because you'll pay for the entity for as long as it exists, and entities don't close quietly.
Hiring Them Into an Existing Entity Elsewhere and Having Them Work Remotely
You keep them on the payroll of an entity you already have, in a country you already operate in, and they work from their home country. No new entity. No new payroll. The structure is already in place. The cost is low and the setup is fast. The OECD permanent establishment framework released in November 2025 introduced a two-part test for when a remote employee's home office may be a fixed place of business giving rise to a permanent establishment. Part one is a time threshold: a home office is generally not a fixed place of business where the employee works there for less than half of their total working time over any rolling twelve-month period. Part two is a commercial reason test: remote arrangements driven solely by employee preference, talent retention, or internal cost efficiency such as reducing office space aren't commercial reasons that give rise to a permanent establishment. Where the time threshold is crossed and the arrangement has a commercial reason, a permanent establishment can arise, with corporate tax consequences. It earns its place when the existing entity is in a sensible tax jurisdiction, the role is genuinely remote, and the time and commercial reason test is on your side. It fails when the home country activity grows past employing staff, when the worker starts handling local customers, or when the role is so senior that the home office starts to look like a branch.
The Decision Table
| Situation | Scale | Setup | Primary Pain | Recommended Starting Point |
|---|---|---|---|---|
| Defined deliverable, short, outside core | One person, project | Days | Cost of getting the contract wrong | Contractor with a specific, written scope |
| Recurring work, integrated, single country | One person, ongoing | Weeks | Classification if left informal | Employer of record |
| Multiple hires planned, country growth real | Two or more, building | Months | Setup cost and ongoing overhead | Local entity from the start |
| Senior hire, existing entity, genuinely remote | One senior person, established | Weeks to a quarter | Permanent establishment if it grows | Remote into existing entity, monitor the OECD test |
| Stock grant material to the offer | One or more, equity-led | Varies by structure | Equity tax treatment in the worker's country | Decide structure before the grant, not after |
| Local customer or sales activity in country | Activity beyond employment | Months | Permanent establishment, corporate tax | Local entity |
| UK small company, end of off-payroll regime | Single contractor, UK | Days to weeks | Losing the small exemption as the threshold rises | Reassess the file before relying on the small end-user exemption |
| Platform-style work in EU member state | One or more, defined activity | Varies | Burden of proof shifts in transposition | Confirm local transposition, do not assume the directive does not apply |
The Decisions, In Order
The sequence below is the order in which the founder actually makes the calls. The earlier ones block the later ones. Leave any of them open and the next one is made on a guess.
| Decision | Who Owns It | What It Blocks If Left Open |
|---|---|---|
| Confirm the relationship is employment, not project work | Founder, with HR or legal review | The choice between contractor and every employment route |
| Choose the country of employment for tax and entity purposes | Founder, with finance and legal | Whether to set up a new entity, use an EOR, or hire into an existing one |
| Pick the employment structure for that country | Founder, with HR, finance, and local advice | Drafting the contract, running payroll, equity grant treatment |
| Set the employment terms: notice, probation, termination, IP | HR or people lead, with local advice | The first draft of the employment contract |
| Decide equity treatment at grant and at vest | Founder, with finance and tax advice | The grant agreement, the cap table entry, the individual's tax position |
| Confirm whether the role creates permanent establishment exposure | Finance, with local advice | Whether the role is sustainable in this structure past the OECD time threshold |
| Run first payroll and file locally | Operations or external payroll | The first invoice from the worker, the first statutory deadline |
Once the structure is set, the next set of decisions is what to put in writing before day one, because these are the documents that are harder to add later.
What to Get in Writing Before Day One
Employment terms in most jurisdictions are harder to change after the fact than to set correctly at the start. Probation, notice, and termination terms are the ones people wish they had thought about, because the off-ramp gets written when the relationship is friendly, and it's read for the first time when it isn't.
| Document | Why It Matters Now | What It Hardens If You Wait |
|---|---|---|
| Employment contract with locally-compliant terms | Sets notice, probation, termination, IP, confidentiality | The statutory defaults in country, which may not match your intent |
| IP assignment, broad enough to cover the work | Confirms the company owns what is built | Local default rules, which in many countries do not auto-assign |
| Confidentiality and non-solicit, if enforceable locally | Protects customers and other staff | Inability to enforce later in the local courts |
| Equity grant agreement, structured for the country | Aligns the grant with the tax treatment | A grant done on the wrong structure, with a tax bill the worker did not expect |
| Probation terms, with explicit review milestones | Sets the off-ramp if the role is not working | The statutory notice period, which can be long and expensive |
| Data protection and equipment policy | Covers handling of personal and company data | The local data protection regime, which is read literally |
| Termination terms, including severance framework | Sets what ending the relationship costs | The statutory minimum, which is often higher than what you would have agreed |
The Cost of Getting This Wrong
The bill that arrives after a misclassification finding isn't the bill you expected. There's the back tax, of course, and the employer social contributions that were never paid, and the interest, and the penalty. There's also the time. The finance lead is on the file for six months. The founder is on the file for three. The candidate is distracted for a year. The audit opens a door to a wider review of every other classification in the business, including the home country contractors you forgot you had, and the conversation with the board that follows is one you can't have in a single page.
The cost that doesn't appear on any invoice is the option cost. The structures you could have used but didn't, because the file is now locked into a remediation path. The markets you could have entered but didn't, because the country risk is now visible. The senior hire you could have made but didn't, because the equity grant you did get right on the first person would have been the template, and the template is now the wrong one. None of this shows up on a P&L. All of it's real.
So the question to ask before you sign isn't "what is the cheapest way to make this work". The question is "what is the file I want to hand to the next people director in three years". If the answer is a file you can defend without flinching, the cheapest route is rarely the one that gets you there.
When You Are Ready to Go Further
If the file in front of you is more than a single hire, if you're mapping a country strategy, or if you want a second opinion on the structure before you commit, the HROpsLab team works on independent comparison research on the providers in this space, including EORs, local payroll, and global mobility. We don't sell software, we don't provide consulting, and we don't provide legal advice. We are a review publication, and our funding model is reader-supported and clearly disclosed. The independent comparison work is the part of what we do that's most useful at the stage the reader is at, because by then the questions have stopped being definitional and started being structural.
The fastest way to use it's to start with our comparison of the operational models in your target country, which sets out the trade-offs by structure and by headcount, and to come back with a specific file rather than a general question. The team is small, the queue moves on a real schedule, and the answers are written by people who have run the file.
Frequently Asked Questions
Should I start them as a contractor and convert later?
The IRS weighs all factors together as a totality of the circumstances, with no single factor decisive, and California applies the ABC test, which presumes employee. A company can satisfy the IRS test and still fail the ABC test. Starting as a contractor and converting later is the pattern that creates the worst misclassification files, because the file at conversion shows the integrated, ongoing relationship the law treats as employment. If the work is employment, the right move is to employ from day one, even if the structure is an EOR.
How long does setting this up actually take?
The contractor route takes days. The EOR route takes weeks. A local entity takes months, and the timeline is set by the country, not by you. The honest answer is to confirm the timeline locally, because country-level incorporation and payroll registration speeds vary, and the only number that matters is the one in your worker's country.
Can they just work on my home-country contract?
In some cases yes, in many no. The OECD permanent establishment framework released in November 2025 sets out a two-part test for when a remote employee's home office may be a permanent establishment, with a time threshold and a commercial reason test. Where the time threshold is crossed and the arrangement has a commercial reason, a permanent establishment can arise, with corporate tax consequences. The answer depends on where the worker sits, how much of their time is in the home office, and what the activity in country is. Confirm locally.
What happens if they move country again?
The structure you set up for the first country may not work for the second. The employment relationship is governed by the law of the country where the worker is, and a move changes the country, the tax residence, the social security position, and the equity treatment. The right move is to plan for the move as a possibility from day one, and to confirm locally what the new country requires before the move happens.
Who handles their tax?
If they're on an EOR, the EOR does, as the local employer. If they're on your local entity, your local entity does. If they're a contractor, they do, in theory, and that's the file that creates the exposure when the tax authority disagrees. The answer depends on the structure, and the structure has to match the file, not the other way around.
Does one person justify a local entity?
Sometimes. The honest answer is that the cost of a local entity is the same whether you've one employee or ten, and the decision is about your twelve-month plan for the country, not the one person in front of you. If the plan is real, the entity is justified. If the plan is a guess, the EOR is the right starting point.
What do I do about equity?
Decide the structure before the grant, not after. The equity treatment in the worker's country is set by the structure and the grant type, and changing it after grant is harder than getting it right at grant. Confirm the local tax treatment before the grant, because the candidate's net share count is set by it.
We give founders and people teams the operational truth on hiring, paying, and retaining people across borders, so the next decision is made on the file, not the pitch.
