7 mistakes when setting up a foreign company that can cost you the business

Setting up a company abroad looks like an administrative exercise. Pick a country, sign the papers, wait for the certificate. In practice this is where the decisions get made that later determine whether the business can take payments, pay dividends, survive investor due diligence, or function at all.

The mistake is almost never visible at the time. The company exists, the documents are in hand, everything looks right. The problem surfaces six months later when a bank declines the account, or two years later when a tax authority somewhere asks where the company was actually being run from. By then the fix costs more than the correct structure would have cost at the start, and it takes months rather than days.

We cover the seven mistakes we see most often. They share one underlying logic: the founder treats incorporation as a problem separate from running the business. Incorporation is not the finish line, it is roughly step one of twelve, and every step after it depends on what was decided at the start.

The short version, if you are skimming. A jurisdiction is chosen on where your customers are, where the money to grow will come from and what infrastructure the business needs, not on the tax rate. A company’s tax residence is determined by where it is actually managed, not by where it is registered. The bank account is not part of the incorporation package and nobody can guarantee it. Treaty benefits are no longer automatic and, in some cases, no longer available at all. Notifying your own tax authority of an interest in a foreign company is a deadline, not an option. And annual upkeep almost always costs more than incorporation did.

What follows is the detail, with the consequences and how each one is fixed.

The most common sequence goes like this. The founder reads a list of the best countries to register a business, sees a low tax rate, incorporates. A few months later it emerges that the customers are in the EU and want a supplier with an EU VAT number, and that the payment provider does not work with the country chosen.
The problem is that the comparison starts at the wrong end. The tax rate is the last filter, not the first. The first three questions are different ones.

Where your customers are and what currency they pay in. A German corporate buyer purchasing from a Caribbean company deals with the reverse charge itself. Plenty of them would simply rather not, and will pick a supplier with a European entity instead. If your market is the EU, the jurisdiction has to give you a VAT number, and that consideration outweighs a few percentage points of corporation tax.

Where the money to grow will come from. If a venture round with US funds is planned, the corporate form has to be the one the standard financing documents are drafted for. A fund offered an unfamiliar structure will either require a reorganisation or price the extra legal risk into the valuation. Restructuring after the round costs more than the right structure at the start, and it pushes the timing of the deal.

What infrastructure the product needs. Software providers, payment processors and cloud platforms all maintain lists of supported countries. A company incorporated outside them cannot connect to the service it needs legally. A low tax rate is irrelevant in that case: there will be nothing to tax, because the product will not launch.

How to fix it. If the company is already in the wrong jurisdiction there are two routes. The first is to set up a second entity in the right country and move the operations across, leaving the first as a holding company or winding it up. The second is redomiciliation, moving the company itself, which only works where both jurisdictions permit it and almost always has tax consequences for shareholders. The first route is usually faster and cheaper.

This is the most expensive mistake of the seven, and also the least visible.

There is a persistent belief that a company registered in another country therefore pays tax in that country. In reality almost every tax system in the world determines corporate residence by where the company is actually managed, not by where it was incorporated. If the director makes the decisions, signs the contracts and handles the correspondence while physically sitting somewhere else, the tax authority of that country has grounds to treat the company as its own resident, with everything that follows: tax on worldwide profits, filing obligations and interest on prior periods.

The indicators are fairly mundane. Where the online banking is accessed from. Where the directors physically are when documents are signed. Where board meetings happen and where the minutes are drawn up. Whose phone number appears on the contracts. Where the seals and original documents are kept. There is no single formula; it is assessed on the balance of the facts.

The trend is running in one direction here. Cyprus extended its definition of a resident company from 1 January 2026: a company incorporated under Cypriot law is now automatically treated as a Cypriot tax resident unless a double tax treaty provides otherwise. Previously the company also had to not be resident anywhere else. It removes one ambiguity and creates another, because a company incorporated in Cyprus but managed from a third country now risks dual residence.

What to do. Either put real management in the country of incorporation, or accept residence where the management actually sits and build the structure around that fact. The third option, incorporate and hope nobody notices, does not exist in 2026: automatic exchange of financial information works, and details of the company’s accounts reach the tax authority where the beneficial owner is resident.

Incorporation and opening an account are two different procedures with different requirements, different timelines and very different odds. A company is registered by a state registry on formal grounds, and refusals are rare. An account is opened by a commercial bank on its own assessment of risk, and refusals are routine.
The timelines are not comparable either. Incorporation takes between two days and two weeks depending on the jurisdiction. Opening an account takes from one week to several months, and that is if the bank agrees at all. A founder who planned the launch around the incorporation timeline ends up with a company and no way to take money.

What the bank assesses: the nature of the business and how risky it is, the ownership structure and how clearly it traces to the ultimate beneficial owner, the source of the funds paying for the share capital and the first costs, whether the company has any real connection to the country or region, the nationality and residence of the beneficial owners, and where the future counterparties are.

Jurisdictional risk matters here in a specific and mechanical way. The EU maintains a list of third countries with strategic deficiencies in their anti-money-laundering regimes, and financial institutions in the EU are required to apply enhanced due diligence to any relationship connected to a listed country. Russia was added to that list with effect from 29 January 2026. This is not a prohibition, but it makes the process noticeably longer and more demanding for structures with Russian beneficial owners. Note that this is a separate list from the EU list of non-cooperative jurisdictions for tax purposes, which has different consequences; the two are frequently confused.

The right way round. Check the banking route before incorporating, not after. In practice that means agreeing the company profile with a bank or payment institution while the jurisdiction is still being chosen. We run that check before filing the incorporation documents, so that the company is not built for a route that is closed to it. The final decision always rests with the bank, and nobody can guarantee an account. Anyone offering that guarantee either does not understand the process or is being dishonest about it.

Classic international structuring was built on reduced withholding tax rates under double tax treaties. Dividends flowing to a European holding company were taxed at 5% or 10% instead of the domestic rate; interest and royalties often were not taxed at source at all.

Three things have changed, and any structure designed before them needs recalculating.

Treaty benefits are no longer automatic. The Multilateral Instrument introduced a principal purpose test into most of the world’s treaty network: if obtaining the benefit was one of the principal purposes of an arrangement, the benefit can be denied. A holding company inserted purely to reduce withholding tax now has to justify its commercial reason for existing.

Beneficial ownership is tested, not assumed. Tax authorities look at whether the recipient of the income has genuine control over it or is simply passing it through. A conduit entity with no functions, no staff and no risk does not qualify as a beneficial owner in most jurisdictions, and the treaty rate is refused.

Treaties can be suspended. Russia suspended key provisions of its treaties with 38 states from 8 August 2023 by presidential decree, confirmed by federal law in December 2023. The suspended articles are precisely the ones that gave reduced rates on dividends, interest and royalties. Any structure with Russian operations or Russian-resident shareholders that was designed around a treaty route stopped working at that point.
The practical problem is that structures built before these changes carry on by inertia. The company was created for a specific tax route, the route closed, and the entity remains, generating costs and returning nothing. A separate category is structures created after the change on the basis of advice that was never updated.

What to do. If you have a foreign structure that was designed around a treaty position, recalculate it on current rules. It may be cheaper to wind it up. The optimal country may have changed. There may still be a saving, but a smaller one than the cost of maintaining the structure. This is arithmetic, and it is better done before the next filing deadline than after it.

The company is registered, the address is rented from a service provider, a director is named, and the business is run from somewhere else entirely. Five years ago that worked. It no longer does.

Economic substance requirements have been introduced across essentially every low-tax jurisdiction under pressure from the OECD and the EU. The principle is the same everywhere: a company benefiting from a preferential regime must have real activity in the country, meaning premises, people, expenditure and decision-making. The mechanics differ; the direction does not.

What happens when substance is absent. The jurisdiction of incorporation can refuse the preferential regime, impose penalties or strike the company off. The country where the business is actually carried on can find a permanent establishment and tax the profits there. Counterparties and banks reviewing the company see an empty shell and decline to work with it. And the tax authority where the beneficial owner is resident treats the structure as artificial and disallows the deductions.

One widespread misunderstanding is worth naming. Director and shareholder services, offered by most incorporation providers, do not substitute for economic substance. They solve specific problems, such as privacy in an open register, but they create neither activity nor management. A structure whose entire connection to a country consists of a rented address and an appointed person does not survive bank review or tax review in 2026.

A useful test you can apply to yourself. If the company has no expenditure in its country of incorporation other than the mandatory fees paid to the registered agent, there is no substance. If not a single decision during the year was physically taken on that country’s territory, there is no management there either.

What to do. Either build substance proportionate to the size of the business, meaning premises, a member of staff, local costs, or choose a jurisdiction where your type of activity genuinely takes place and stop claiming a relief you are not entitled to. The second route is cheaper and safer than imitating the first.

Registering a company abroad is not an offence. The offence is failing the obligations that arise for a tax resident once the company exists. The distinction is fundamental, and it is where most people come unstuck.

Most developed countries operate controlled foreign company rules. Under them, the profits of a foreign company can be attributed to its controlling shareholder and taxed at home, even where nothing has been distributed. Alongside that sit notification duties: many jurisdictions require you to inform the tax authority of an interest in a foreign entity within a fixed period, with penalties for missing it.

The Russian rules are a concrete example of how tight the deadlines are. An interest above 10% triggers a notification within three months of the interest arising, with a penalty of RUB 50,000 for failing to file. A separate controlled foreign company notification is due by 30 April of the year following the reporting year for individuals, with a substantially higher penalty. Thresholds, deadlines and calculation methods differ significantly between countries, so this has to be checked against your own residence rather than assumed.

There is usually a currency or exchange control layer as well, with its own duties to report foreign accounts and the movements through them, on separate forms and separate deadlines.

The assumption many people rely on, that nobody will find out, stopped working some time ago. Automatic exchange of financial account information covers most jurisdictions. Beneficial ownership registers are accessible to competent authorities. Banks ask for the tax number and country of residence of the beneficial owner at account opening, and that information travels through the exchange channels.

What to do if the deadlines have already passed. It is unpleasant but not terminal. The notifications are filed late, the penalty is paid, and the company then runs normally. That is substantially better than continuing to stay silent: the exposure grows with every reporting period, and a voluntary correction made before an audit begins is treated differently from a failure discovered during one. The right course depends on how many periods were missed and whether the company had profits, so this needs a review of the actual situation rather than a general rule.

A founder compares incorporation quotes, picks the one that is two hundred dollars cheaper, and considers the matter closed. A year later the maintenance invoice arrives and the saving turns out to be negative.

Incorporation is a one-off payment. Upkeep is annual, and it is almost always the larger number. A typical annual budget includes the renewal or annual return fee, registered agent and registered office services, bookkeeping to local standards, preparation and filing of the tax return, and an audit where one is required. Separate items are translation and apostille of documents, international transfer fees, and the cost of re-verification when changing banks.

Several specific things tend to fall into the blind spot.

A dormant company keeps costing money. The project is on hold and there is no activity, but the obligations remain. A Delaware company with no revenue still pays franchise tax, files an annual report by 1 March and a federal return by 15 April. A foreign-owned company additionally files an information return, and the penalty for failing to file it is $25,000. A UK company with no activity still pays £50 a year for the confirmation statement, and late filing of accounts carries penalties from £150 to £1,500.

The authorised shares trap. A standard US startup charter authorises ten million shares. Under the Authorized Shares Method the franchise tax on that number runs into five figures, and the founder loses sleep over the invoice. It does not have to be paid: the law permits a different calculation, and with the par value set correctly the figure comes to around $400. But this is worth knowing before incorporation rather than on receipt of the notice.

Closing costs money too. A company cannot simply be abandoned. An abandoned company accumulates penalties and interest, and a strike-off initiated by the registrar leaves a mark on the director’s record. Voluntary dissolution is cheaper: in the UK, for instance, a digital strike-off application has cost £13 since February 2026, while restoring a company that was struck off in error costs £500.

What to do. Cost the structure over three years, not on the entry price. The gap between jurisdictions on upkeep is usually wider than the gap on incorporation, and it is what actually determines the economics of the decision.

Most of the situations above are fixable, and fixing them is almost always cheaper than doing nothing. The route depends on the type of mistake.
The wrong jurisdiction is resolved by setting up a second company and moving the operations, or less often by redomiciliation. The first company then becomes a holding entity or is wound up voluntarily. Timeline: one to three months.
A residence problem based on place of management is resolved either by moving management to the country of incorporation, or by accepting the real residence and rebuilding the structure around it. Acting before an audit matters here: a voluntary correction is assessed differently from a discovered breach.
A bank refusal is not a dead end. It usually means the company profile did not suit that particular bank. The workable approach is to establish why, adjust the documents and the structure, and apply to an institution with a different risk appetite. Sometimes the answer lies in changing the company’s jurisdiction, sometimes in adding real substance.
Missed notifications are filed late and the penalties are paid. That is unpleasant but finite. Continuing to stay silent is not finite.
Insufficient substance is closed gradually: a lease, a local employee, real expenditure, minutes of decisions taken on the country’s territory. A history of substance cannot be created quickly, but it can be started at any point.
The general principle: the earlier the work on a mistake begins, the less it costs. The price of a fix rises with every closed reporting period, because the accumulated obligations get added to the original problem.

Reduce all seven mistakes to a positive programme and you get the following order. It takes longer at the start and considerably less time afterwards.
First, describe the business task: who the customers are, where they are, what currency they pay in, where investment will come from, what services the business needs to operate. This happens before any conversation about countries.
Then draw up a shortlist of the jurisdictions that meet the task. It usually comes to two or three, not thirty.
Next, check the banking and payment route for each option on the shortlist. Some options drop out at this stage, and that is fine: better to lose an option now than a company in six months.
After that, establish where actual management will sit and what substance the company can realistically maintain. The answer affects both the choice of jurisdiction and the tax model.
Only now compare tax regimes within what is left, and cost the structure over three years rather than pricing the incorporation.
Then prepare the documents and run incorporation, tax number registration and account opening in parallel rather than in sequence. That saves weeks.
Finally, set up the compliance calendar: filing dates in the country of incorporation, notification dates in the beneficial owner’s country of residence, and a named person responsible for each deadline. A calendar drawn up in the first month costs less than a penalty in the third year.

All seven mistakes come down to the same thing: the incorporation decision is taken in isolation, without regard to how the company will actually live afterwards. The jurisdiction is picked from somebody else’s list, the bank is treated as a formality, the filings are left for later, and the budget is calculated for the day of incorporation instead of three years of ownership.
The right order is the reverse. Task first, then infrastructure, then tax. It takes a week longer at the start and saves months later.

Finextwin is an international corporate services firm with offices in Hong Kong and Tbilisi. Over 6+ years we have registered companies in 30+ jurisdictions for more than 1,250 clients and we work with 65+ banking partners. We work through the structure for a specific task before any documents are filed, including agreeing the banking route in advance. If the company already exists and one of the problems above has surfaced, we handle that too: the fix is almost always cheaper than it looks at the start of the conversation. A manager responds within 30 minutes.