Woman in suit interviews man across table

You usually meet this problem sideways. A candidate you want lives in Lisbon. Or the engineer who has been contracting for most of a year says she would like a real contract. Or a founder meets someone at a conference in Toronto and cannot let it go. Nobody sat down and drew an expansion map. The hire came first and the structure question arrived afterwards, which is the order it happens in for most companies of ten or thirty people.

That sequence shapes the whole decision. You are not weighing market entry in the abstract, you are trying to get one specific person employed, paid on schedule, with whatever the local system says they are owed, before they accept something else.

And you are doing it without an HR function. Maybe there is a head of operations who also owns recruiting and the payroll spreadsheet. Maybe there is nobody in particular, and it is a founder reading a government website at eleven at night, trying to work out whether a thirteenth-month payment is a local custom or a statutory entitlement.

What a local HR team is actually for

It helps to separate the work from the job title. When people say they need local HR, they are usually pointing at a bundle of fairly different things that happen to sit together in larger companies.

  • Drafting an employment contract that holds up under the country’s own labour law, in the required language, with the clauses that country expects to see
  • Registering the employer and the employee with the tax authority and the social insurance system
  • Running payroll on the local calendar, with local withholding, filings and payment deadlines
  • Administering statutory leave, sick pay, parental leave and public holidays as that country defines them
  • Enrolling people in mandatory and customary benefits, including occupational pension arrangements where they exist
  • Knowing what a lawful exit looks like, because termination rules are where the gap between countries is widest

Almost none of that is judgement work. It is procedural knowledge that changes by jurisdiction and gets amended without announcing itself. A startup does not need a person sitting in the country to hold that knowledge. It needs the knowledge to exist somewhere accountable, with someone answering for it when a filing date moves. Accounts of how startups expand globally with an EOR are worth reading mainly because they describe that handover procedurally: who signs the contract, who holds the payroll registration, who carries the filing obligation, and what stays with you regardless.

Because some of it does stay with you. Whether the role is scoped properly. Whether the person has anyone to talk to. Whether their work is visible to the team that hired them. No structure fixes those.

The three routes, and what each one costs you in attention

Broadly there are three ways to employ someone in a country where you have no presence.

You can incorporate. A local subsidiary gives you a permanent structure, direct control over employment terms, and a sensible base for a long-term country plan. It also gives you statutory accounts, a local director requirement in some places, a corporate tax filing obligation, and a payroll registration process that can run for weeks before anyone gets paid. For a first hire it is usually out of proportion to the problem, and unwinding it later is slower than setting it up was.

You can engage the person as a contractor. Sometimes that genuinely describes the relationship and sometimes it is wishful thinking, and the difference is worth examining rather than deciding by convenience.

Or a third party becomes the legal employer while you keep directing the work. That arrangement is usually called employer of record, though it is not a single statutory category that exists in the same shape everywhere. In much of Europe it runs through licensed staffing or temporary work agency frameworks. In Germany it sits under the AüG licensed temporary work agency regime, which brings its own conditions about how long an assignment can run and how the worker is treated relative to direct staff. Those constraints are real and they shape what the arrangement can do for you in a given country.

Contractors are a legitimate route when the work is genuinely independent

There is a version of this conversation where contracting gets treated as the shortcut and employment as the grown-up option. That framing causes problems in both directions.

If someone runs their own practice, serves other clients, sets their own hours, uses their own equipment and carries commercial risk on the outcome, they are a contractor. Agent of record arrangements exist for precisely that situation: agreements, classification review, payments and offboarding handled properly for people who are independent in substance as well as on paper.

The trouble starts when a contractor agreement gets written for a full-time, exclusive, managed role because employment looked complicated. Most countries assess the substance of the relationship rather than the label on the document, and the tests they apply differ. Control over how the work is done. Integration into the business. Exclusivity. Who supplies the tools. Whether the person can send a substitute. A reclassification finding tends to land as unpaid social contributions, back taxes, penalties, and sometimes employment rights that accrue retroactively.

You can usually sense which way it would go by asking a plain question. If this person stopped working tomorrow, would you look for another contractor, or would you post a job?

What actually changes in the first month

The practical difference between the routes shows up early, in things that feel small until they are not.

Payroll cutoff dates you did not know existed. A tax registration number that has to appear on the payslip. A probation period capped by statute rather than by your handbook. A notice period that scales with seniority. Mandatory health insurance enrolment with a window on it. Employee representation obligations that switch on once headcount in the country passes a certain point, which nobody notices passing unless somebody is watching for it.

Each of those needs an owner. That is the real decision underneath the structure question.

The costs that stay out of the model

Finance usually models employer contributions and a provider fee, then stops. The lines that get missed are the ones that never arrive as an invoice.

Time is the main one. Entity setup consumes a founder’s weeks rather than an analyst’s hours, and it does it at the point when the company can least spare either. Then there is the tail: annual filings, a local accountant, a registered address, bank account maintenance, and the eventual cost of closing the entity down if that country turns out not to be where your market was.

There are also currency and banking mechanics. Paying someone in their own currency from an account held in another one means spreads, cutoffs and the occasional failed transfer. Nobody forgets the month their salary arrived late.

Permanent establishment sits in this category too, as a live question rather than a settled one. Employing through a third party does not by itself determine whether your activity in a country creates a taxable presence, particularly where the person there is selling, negotiating or concluding contracts on your behalf. It turns on the facts and on the relevant treaty, and it belongs with your own tax and legal advisors rather than with a comparison table.

The thin internal layer you still need

Even with the employment mechanics handled outside the company, a few things have to live inside it or the arrangement drifts.

Keep one person as the owner of the employment relationship, whoever that is. Someone has to approve changes, answer the provider’s questions quickly, and notice when a contract amendment is needed. Compliance work stalls in inboxes more often than it fails on the merits. That division of labour is the useful part of any honest account of an EOR versus your in-house HR, because the question is rarely about capability and usually about where the marginal hour of attention goes when you have very few of them to spend.

Write down your compensation logic before the second country, too. Nothing formal, just a defensible reason why roles pay what they pay, because the third hire will compare notes with the first and you want an answer that holds.

Then decide what is global and what is local in your benefits, and say it plainly. Statutory entitlements vary and you cannot flatten them. What you can keep consistent is the discretionary layer: leave above the minimum, equipment budget, learning allowance, how flexible the hours really are in practice rather than in the policy.

The second country tells you what you built

One international hire can be absorbed by improvisation. Two in different countries is where the pattern surfaces. You find out whether your onboarding works when nobody shares an office, whether your policies say anything useful once the statutory floor moves, whether a manager knows what leave their report is entitled to without having to ask.

Teams that come through it cleanly tend to have made one decision early. They treated the employment structure as something to get boring quickly, so the interesting work could stay on the person they hired. The paperwork was never the point. It just has to be quiet enough that you can stop thinking about it.