Cross-Border Hiring: Why Competent Teams Still Keep It In House

A startup's first cross-border hire usually gets budgeted like a vendor contract: pick an Employer of Record, sign, and the payroll problem is considered solved. What the EOR actually stands between the company and is three separate exposures, whether the work sits inside a compliant employment relationship, whether the arrangement creates a tax obligation in the worker's country, and whether a labor authority would call this person an employee no matter what the paperwork says. Get the order wrong and the company finds out after the bill for it arrives, not before.

What changes the answer is not a new rule

The decision usually flips for one of two reasons, and neither is a regulatory change.

This is not an argument for outsourcing. It is the set of things that determine what your own decision will cost you if it turns out to be wrong.

The test that ignores what the contract is titled

Every guide to cross-border hiring is a list of risks. A competent HR leader reads that same list, understands every item, and still decides to keep the work in house.

Three answers come up more than any other, and none of them is ignorance of the rules.

This is not a uniquely Taiwanese habit. Most jurisdictions that regulate employment run some version of the same substance-over-form test, they differ in name and in how aggressively regulators enforce it, but the underlying question, who actually controls the work, is close to universal. A company using a contractor structure to avoid the cost of employment is making a bet on enforcement risk in a country it may not have visited.

The cost floor that shows up after the pilot, not during it

Two numbers in Taiwan law are worth knowing before quoting a hiring cost to a founder. Under Article 14 of the Labor Pension Act, an employer must contribute at least 6% of an employee's monthly wage into that employee's individual pension account, on top of gross salary. Under Article 27 of the National Health Insurance Act, for a standard salaried employee the employer, as the insured unit, bears 70% of the National Health Insurance premium, with the employee paying the remaining 30%. Neither number is a service fee an EOR charges. Both are a floor set by statute, and they apply whether the hire is made through a local entity, an EOR, or gets misclassified as a contractor and later reclassified.

The reason this matters for sequencing is that a startup pricing its first hire off the gross salary line will be short by a predictable, statutory percentage the moment payroll actually runs, and short again if the arrangement is later reclassified from contractor to employee, because the pension and insurance obligations apply retroactively to the period the relationship existed, not from the date it was corrected.

The tax question that has a name and a definition

The third layer has a formal name in international tax law: permanent establishment. Article 5 of the OECD Model Tax Convention on Income and on Capital sets out when a foreign company's presence in a country becomes a taxable one, and one of the routes it lists, the agency permanent establishment under Article 5, paragraphs 5 and 6, is triggered when a person habitually concludes contracts on the company's behalf, regardless of whether that company ever rented an office. A remote employee closing deals, signing local agreements, or acting with authority on the company's behalf can create exactly this kind of taxable presence in the country where they sit.

An EOR is often marketed as removing this risk entirely. What it actually does is put a local legal employer between the company and the tax authority, it does not change what the worker does day to day, and agency permanent establishment turns on function and authority, not on who signs the paycheck. A company whose remote hire is closing contracts, not just doing back-office work, should treat the tax question and the labor law question as two separate diligence items, because a structure that solves one does not automatically solve the other.

Lay the three common hiring structures side by side and the sequencing problem becomes visible. A local entity puts every layer of risk on the company but also gives it full control. An EOR is supposed to absorb the labor law and payroll layer by standing in as the legal employer, but it does not erase a badly structured relationship, it only relabels who is holding it. An independent contractor arrangement looks like it avoids all three risks, that is exactly why founders reach for it first, and it is also the structure regulators are trained to test first.

Read the misclassification row again. It is the only row where the answer does not depend on which vendor a company picks, it depends on what the work actually looks like day to day. An EOR contract does not fix a relationship where the company is directing daily tasks, setting hours, and integrating the worker into internal systems while calling them a contractor on paper. It just moves that same fact pattern onto a different legal employer.

Where EvoScale reads this

Operator-led investing means looking at a hiring plan the way someone who has sat on both sides of the employment relationship would, having built an HR function inside a company and having sold HR technology to companies building one. From that seat, the recurring pattern is not that founders skip diligence, it is that they run it in the wrong order. Legal classification and statutory cost get checked after a hire is already live, usually once someone asks why the payroll run does not match the offer letter. The fix is not a longer compliance checklist. It is moving the labor law question, and the tax question, ahead of the payroll question, before the first cross-border offer goes out, not after the first invoice does.

For a HRTech company selling into this problem, the product worth underwriting is not the payroll dashboard, dashboards are commoditized and every competitor has one. The defensible layer is the sequencing logic itself: knowing which jurisdiction's subordination test applies, which permanent establishment trigger the role in question could hit, and which statutory floor changes the moment a headcount plan crosses a border. That knowledge does not show up on a feature comparison chart, and it is exactly the kind of operating detail that is hard to fake and hard to copy quickly.

None of this argues against hiring across borders, it argues for pricing the whole decision correctly the first time. A founder who treats an EOR as a payroll vendor will still get a functioning payroll run. What they will not get, unless they ask for it specifically, is a clear answer on which labor law applies, whether the role creates a taxable presence somewhere else, and whether the underlying relationship would survive a subordination test if anyone ever looked closely. Those three questions cost nothing to ask before the offer letter goes out.

The category is not glamorous, and the invoice for getting it wrong rarely comes due on the day of the mistake. It comes due on the day someone finally reads the contract closely, and that is usually much later, and much more expensive, than the day it was signed.

Sources: Labor Standards Act (Taiwan), Article 2, Clause 6, Laws & Regulations Database of the Republic of China (law.moj.gov.tw). Labor Pension Act (Taiwan), Article 14, Laws & Regulations Database of the Republic of China (law.moj.gov.tw). National Health Insurance Act (Taiwan), Article 27, Laws & Regulations Database of the Republic of China (law.moj.gov.tw). OECD Model Tax Convention on Income and on Capital, Article 5 (Permanent Establishment), paragraphs 5 to 6 (oecd.org).