Tax Treaties Explained

Let's talk about the single most misunderstood document in cross-border tax. You have tax treaties explained to you in vague terms all the time (some accountant waves a hand and says "don't worry, there's a treaty"), but almost nobody tells you what one actually does. So today we're going to fix that. By the end of this you will understand how a double tax treaty allocates the right to tax your income, why residency is the whole ballgame, and how treaty benefits can quietly evaporate if you structure things the wrong way.

Here is the problem a treaty exists to solve. When you earn income across a border, two governments show up to the same dinner and both order off your menu. The country where the money is generated (the source state) wants a bite. The country where you live (the residence state) wants a bite too. Without any coordination, you get juridical double taxation, which is a fancy way of saying comparable taxes imposed by two states on the same taxpayer, on the same income, for the same period. A bilateral income tax treaty, often called a Double Tax Agreement, is the referee that decides who taxes what.

What a Double Tax Treaty Actually Is (and What It Doesn't Do)

Now I know what you're thinking. A treaty gives a country the right to tax me. It does not. This is the point everyone gets backwards, so read it twice: a tax treaty does not create a taxing right that did not already exist under domestic law. Treaties are restrictive instruments, not generative ones. They limit or modify the taxing rights a country already has under its own statutes. In practice, a treaty usually works by forcing the source country to give up some or all of the tax it would otherwise charge on income earned by a resident of the other country.

There are more than 3,000 active bilateral tax treaties in the world, and they all sit on top of domestic law rather than replacing it. That distinction matters because it tells you where to look when something goes wrong. If a treaty caps a tax at 10% but your home country's domestic relief already gets you to zero, the domestic rule can win. This is exactly why treaties interact so tightly with mechanisms like the Foreign Tax Credit. When a treaty's tie-breaker deems you a resident of the foreign country, it can force your home country to re-source the income as foreign, which unlocks the credit and stops the double tax. If you want the mechanics of that interaction, our breakdown of FEIE vs Foreign Tax Credit walks through how domestic relief and treaty sourcing fit together.

And here is the part that surprises people. A treaty protects you against a lot, but not everything. It generally offers no shield against your own country's anti-deferral rules. If you park mobile income in a low-tax subsidiary, your residence state can tax you on your share of that income before a single dividend is paid, and the treaty's business-profits and dividend articles will not save you. We cover this trap in detail in our guide to CFC rules, and it is the reason treaty planning and domestic anti-deferral planning have to be done together, never in isolation.

How Tax Treaties Work: The OECD Model Convention and the Shape of a Treaty

Countries do not draft each of these 3,000 treaties from a blank page. They start from a template, and the dominant template is the OECD Model Tax Convention, first published in 1963 and revised regularly ever since. Understanding how tax treaties work starts with understanding the bias baked into this model.

The OECD Model was designed for trade between developed economies where capital flows roughly evenly in both directions. Because money moves both ways, the model leans hard toward residence-based taxation. It asks the source country (the host, where the investment lands) to surrender some or all of its tax on categories like business profits, royalties, and certain gains, handing the primary taxing right to the country where the taxpayer lives.

The counterweight is the UN Model Convention, built for developing, capital-importing nations that would get stripped of revenue under a pure OECD approach. The UN Model preserves more taxing rights for the source country. Mechanically it does two things:

• Lower PE threshold: It makes it easier for a foreign business to trigger a taxable presence, including shorter time limits for construction sites and explicit rules for service-based presence.
• Higher source withholding: It lets the source country keep taxing outbound passive income like royalties and technical service fees, rather than dropping to zero.

The actual treaty text is organized article by article. Some articles define terms (who is a resident, what counts as a permanent establishment), some allocate taxing rights over specific income (dividends, interest, royalties, capital gains), and some handle disputes and information exchange. The OECD also publishes Commentaries, which are not legally binding but are the interpretive manual tax authorities and courts reach for when the words get ambiguous. When you read a treaty, you are really reading a negotiated compromise between the OECD residence bias and whatever source-state protections the two countries fought to keep.

Treaty Tie-Breaker Rules: Which Country Gets to Call You a Resident

Everything in a treaty hinges on residency, because a treaty only applies to residents of one or both of the contracting states. So what happens when both countries look at you and both say "you're ours"? That is dual residency, and it is common for anyone with a home in two places. This is where treaty tie-breaker rules do their work.

For individuals, the OECD Model runs a strict cascade under Article 4. You go down the list in order, and the moment one test gives a clear answer, you stop:

1. Permanent home. Where do you have a home available for continuous use, not just a hotel or a short-term rental? If that points to one country, you are done.
2. Centre of vital interests. If you have a permanent home in both countries (or neither), the treaty asks where your personal and economic life is centered. Authorities weigh family, social ties, your main occupation, and where you manage your wealth.
3. Habitual abode. If vital interests are a genuine toss-up, and for hyper-mobile executives they often are, the test moves to where you actually spend your days.
4. Nationality. Still tied? You are treated as a resident of the country you are a citizen of.
5. Mutual agreement. A citizen of both or neither goes to the two governments' competent authorities to settle it by negotiation.

The winner gets the primary right to tax your worldwide income. The loser can still tax income sourced inside its borders, but it has to give relief, usually an exemption or a credit. If you are planning a move and want to control which way that cascade breaks, the sequencing matters enormously. Our piece on UAE tax residency shows how the permanent-home and habitual-abode tests play out when you relocate to a low-tax hub.

Companies get their own tie-breaker. Historically, a dual-resident company was assigned to the country where its "place of effective management" sat, meaning where the board actually makes the real decisions, not just where the certificate of incorporation lives. After the OECD's BEPS work, the standard rule changed. Under the Multilateral Instrument and the 2017 model, a dual-resident company now goes to the competent authorities, who weigh effective management, place of incorporation, and other factors. The risk sits in that last clause: if the two authorities cannot agree, the company can be denied treaty benefits entirely.

Permanent Establishment: When Your Business Gets Taxed Somewhere Else

For businesses, the whole question of "can this other country tax my profits" comes down to one concept: the permanent establishment, or PE. Under Article 7, an enterprise's profits are taxable only in its home country unless it operates in the other country through a PE there. Cross that line and the source country gets to tax the profits attributable to that presence.

So what creates a PE? Traditionally, a fixed place of business through which the company operates. The model spells out obvious examples:

• A place of management, branch, office, or factory.
• A workshop, mine, oil or gas well, quarry, or other extraction site.
• A construction or installation project, but only if it lasts more than twelve months.

There is also the agency PE. Even with no office at all, a company creates a PE if a dependent agent habitually concludes contracts in its name in the other country. Work through a genuinely independent broker acting in the ordinary course of their own business and you are fine. And there are carve-outs: a facility used purely for storage, display, delivery, or collecting information is preparatory or auxiliary and does not create a PE on its own.

One recent shift every remote-first company should have on its radar. The 2025 OECD Model Update, the first comprehensive revision since 2017, tackled the home-office problem head on. The new commentary sets a benchmark: if an employee works from a home office abroad for less than 50% of their working time over any 12-month period, that location generally does not become a PE for the employer. Cross that 50% line and authorities look at whether there is a real commercial reason for the arrangement (a time-zone advantage for serving customers, say) rather than the employee's personal preference. When the continuous use and a commercial rationale line up, that spare bedroom can crystallize into a fixed-place-of-business PE, pulling the company's profits into the host country and putting it on the hook for tax it never intended to owe there.

How Withholding Tax Rates Get Determined (Dividends, Interest, Royalties)

Active business profits run through the PE rules. Passive income (dividends, interest, royalties) runs through a completely different mechanism: withholding tax. When a company sends these payments across a border, the source country grabs its cut at the point of payment, because the recipient is beyond its reach. Left to domestic law, those statutory rates are steep, often 20% to 30% on this kind of income. Treaties exist to knock them down and keep capital moving.

Here is roughly how the numbers shake out under the OECD default and modern treaties between developed economies:

• Dividends (Article 10): A two-tier structure. Portfolio investors and small shareholders historically pay a reduced 15%. Corporate shareholders holding a substantial stake (typically 10% to 25% of capital or votes) drop to 5%, and sometimes all the way to 0%. This lower rate is what makes tax-neutral repatriation of profits from a foreign subsidiary up to the parent possible.
• Interest (Article 11): The OECD caps source-country tax at 10% of the gross amount, but modern treaties between developed economies frequently take it to 0%, which is vital for intra-group financing.
• Royalties (Article 12): The OECD Model advocates a clean 0% at source, with the residence country taxing instead. The UN Model breaks from this and lets the source country keep withholding, which is why royalty taxation stays a genuinely contested area.

There is one condition that overrides all of this, and it catches people who try to be clever. To claim any reduced treaty rate, the recipient has to be the beneficial owner of the income. A conduit company that collects a payment and immediately passes it to someone in a third country gets nothing. Tax authorities read "immediately" generously, so a holding company that exists only to catch and forward a payment does not become a genuine owner just because it sits on the cash for a quarter. That is the treaty's built-in defense against dressing up a payment and slipping it through a favorable jurisdiction.

The same source-versus-residence logic governs capital gains under Article 13. The default is that gains are taxed only where the seller resides, but there are big exceptions. Real estate is always taxable where it sits. And the anti-avoidance rule everyone should know: if you sell shares in a company that derives more than 50% of its value from real estate in a country, that country can tax the gain, using a 365-day look-back so you cannot dilute the ratio with a last-minute cash injection. Put the property inside a corporate shell and it still does not hide from the source state.

Tax Treaty Benefits and Their Limits: Treaty Shopping, LOB, and the MLI

Tax treaty benefits are meant for genuine residents of the two contracting countries. For years, people abused that by treaty shopping, meaning a resident of a third country sets up a hollow holding company in a treaty jurisdiction purely to grab withholding reductions and gains exemptions they were never entitled to. The way tax people describe it, any one treaty should be treated as a potential treaty with the whole world, because payments get routed through whichever low-substance hub offers the best rate.

The response was BEPS Action 6 and the Multilateral Instrument (MLI), and the MLI is a well-designed piece of legal engineering. Rather than making countries renegotiate thousands of treaties one at a time (a project that would take decades), the MLI sits on top of existing treaties and modifies them all at once when a country ratifies it. It sets minimum anti-abuse standards, and countries pick their weapon:

• The Principal Purpose Test (PPT): A broad, subjective rule. If one of the principal purposes of an arrangement was to obtain a treaty benefit, the benefit is denied. This has become the global default, adopted by the UK, Spain, and Canada, among many others.
• The Limitation on Benefits (LOB) clause: An objective, mechanical set of corporate tests. This is the US approach, and notably the US did not sign the MLI at all, preferring to negotiate detailed LOB provisions bilaterally so it keeps unilateral control over the definitions.

Under an LOB, simply being a resident under Article 4 is not enough. You have to separately qualify as a "qualified person" by passing at least one structural test: a publicly-traded test, an ownership and base-erosion test, an active-trade-or-business test, or a derivative-benefits look-through. This is not academic: the US-UK treaty offers a coveted 0% withholding on subsidiary dividends, but to get it the beneficial owner must hold 80% or more of the voting power for a 12-month period and clear the full Article 23 LOB. After Brexit, plenty of UK holding companies that relied on the derivative-benefits test (which counted EU and EEA residents as equivalent beneficiaries) found the ground had shifted, and the "US LOB problem" is still a live structural headache for European groups holding US investments through the UK.

Layer on top of all this the OECD's Pillar Two rules, a 15% global minimum tax for large multinationals that treats a PE as its own constituent entity, and you can see why real structuring is never a matter of reading one article in one treaty. It is reading the treaty network, the domestic anti-deferral rules, the LOB or PPT overlay, and the minimum-tax regime together. This is the kind of multi-layered analysis our cross-border structuring services are built around, because a plan that is clean under one treaty can fail under the anti-abuse rule sitting right beside it.

So What Does All of This Mean for You

If you take nothing else from this, take the order of operations. First, establish which country the treaty makes your residence, because the tie-breaker cascade decides who taxes your worldwide income and everything downstream depends on it. Second, figure out whether any of your business activity crosses the PE line, because that is the switch that turns on source-country taxation of your profits, and the 2025 remote-work rules moved that switch closer than it used to be. Third, run every dividend, interest, and royalty flow through the withholding article and confirm you are the genuine beneficial owner, not a conduit. Fourth, before you rely on any of it, check that you actually qualify under the LOB or survive the PPT, because being a resident is no longer enough to claim the benefit.

The reason a treaty feels intimidating is that it looks like a wall of numbered articles, but underneath it is doing one simple job: deciding, for each type of income, whether the source country or the residence country gets to tax it, and by how much. Once you can read a treaty as that allocation exercise rather than a magic shield, you can plan around it with real precision. And as you can see, it is a lot more complicated than "don't worry, there's a treaty" ever suggested.