Transfer pricing for intellectual property (IP) refers to the rules used to determine the price that one related company charges another for using IP. Such intellectual property may include the transfer of valuable intangible assets, including patents, trademarks, software, and technology. Intellectual property often accounts for a large portion of a multinational company’s profits, which can be shifted across countries through transfer pricing.
Here’s a simple example to begin with: Suppose that a U.S. parent company owns a patented manufacturing process. It permits its subsidiary in Germany to use that patent to manufacture products and, in return, pays the U.S. parent a royalty.
Since both companies are part of the same corporate group, they could agree on any royalty rate they wished, even an unfair one. For example, if the German subsidiary paid a particularly low royalty. This means more profit would remain in Germany rather than in the United States. This will reduce the United States parent company’s taxable income and, consequently, its U.S. tax liability.
To prevent companies from engaging in this practice, the IRS mandates that internal transactions between related parties be priced as they would be if the two companies were complete strangers negotiating in an open market. This approach is known as the arm’s length standard. In simple terms, it means “price it as you would if you didn’t know the other party or trust them.”
Before a company can apply these rules, it should first determine whether the transaction involves intellectual property, as special, stricter rules apply in that case. Let’s examine what counts as intellectual property and how such transactions usually operate.
In the United States, when it comes to transfer pricing, intellectual property typically refers to intangible assets that lack a physical form yet provide real economic value to a business.
Common examples of intellectual property include the following:
This intellectual property enables a business to earn more money than a competitor without these advantages. A company possessing a strong brand name can charge a higher price for the same product as a competitor with no brand name. Also, a company with a secret formula can produce something that no one else can copy.
Let’s understand this through a simple example.
For example, let us imagine that a company based in the United States develops a photo-editing application. It owns two intellectual properties in this situation, which are as follows:
The U.S. company enables its subsidiary in Kenya to use both the software and the brand name to sell the app to customers there.
The software code and the brand are valuable intangible assets that generate income and are therefore considered the intellectual property of the U.S. company. The Kenyan subsidiary pays the U.S. parent a royalty, which is a licensing fee, in return for using them.
The main point is that, as both firms belong to the same corporate group. So they might choose any royalty rate they wanted. That is precisely why transfer pricing rules have been introduced.
According to U.S. rules, the royalty in question should be the amount that two completely independent and unconnected companies would have agreed upon in similar circumstances. This ensures that profits are not artificially transferred between the United States and India simply by raising or lowering the internal royalty rate.
However, determining an arm’s-length price for intellectual property is difficult. Next, let’s examine why intellectual property is difficult to value for U.S. transfer pricing purposes.
One of the most difficult areas of U.S. transfer pricing is determining the value of intellectual property (IP). This is because IP is often unique, making it difficult to find a reliable market price.
For example, if you want to determine the price of steel or wheat, you can look at publicly available market prices. Thousands of independent companies buy and sell these products, creating an active market with observable prices.
The same is usually not true for intellectual property. Consider a company that owns a patented drug formula or proprietary software. There may be no other company with exactly the same IP. Also, there may be no publicly available transaction showing what that specific IP is worth.
Under IRC Section 482, transactions involving IP between related companies generally should be priced as if the companies were independent businesses dealing with each other at arm’s length.
Applying the arm’s-length standard can be particularly challenging for IP for three main reasons.
1. IP is often unique.
There may be no identical patent, software, trademark, or other IP available for comparison.
2. Comparable transactions may not be truly comparable.
Even when another company has entered into a similar IP transaction, the underlying technology, market, legal rights, expected profits, and business circumstances may be different.
3. IP may be bundled with other valuable items.
For example, a licensing arrangement may provide not only access to technology but also engineering support, brand recognition, technical know-how, or access to a distribution network. Separating the value of the IP from these other benefits can be difficult.
Since there is often no single, readily available “correct” price for IP, there is greater potential for disagreement between a taxpayer and the IRS. The IRS therefore closely examines related-party IP transactions to determine whether the pricing reflects what independent businesses would have agreed to under similar circumstances.
As a result, determining the value of IP generally requires a detailed economic analysis. Analysts may consider the following factors, such as:
In short, valuing IP for transfer pricing purposes is not simply a matter of finding a market price. It requires an analysis of the IP, the rights being transferred, the expected economic benefits, and the circumstances of the transaction.
Next, let’s take a look at some of the most common intellectual property transactions between related companies and how they affect transfer pricing in the United States.
Related companies have several methods for sharing or transferring intellectual property. The most common types are as follows:
By licensing intellectual property, the original owner retains legal ownership of the IP. However, it permits a related company to use it in return for royalty payments. You can think of it as renting an apartment, where the landlord still owns the apartment and the tenant simply pays to use it.
For example, a U.S. company owns a trademark. It licenses that trademark to its subsidiary in Japan. In return, the Japanese subsidiary pays an annual royalty for the right to sell products under that brand name.
Licensing is by far the most common type of IP transaction reviewed under U.S. transfer pricing rules. Here, the royalty charged should reflect an arm’s-length price.
Under the sale of intellectual property, one company in the related group can sell full ownership of the IP to another company in the same group.
For example, under a sale agreement, a U.S. company transfers full ownership of its software to its Irish subsidiary.
Since ownership is transferred permanently, it is particularly important to set the purchase price correctly. Once the intellectual property together with all the profit it will produce in the future passes to Ireland, that value is permanently removed from the U.S. tax base. For this reason, the IRS keeps a close eye on how these sale prices are determined.
Sometimes related companies don’t merely license or sell existing IP. However, under a cost-sharing arrangement, these related companies collaborate to create new IP. Here, each company funds research and development in proportion to the benefit it expects to gain upon completion of the IP.
For example, a parent company in the United States and its subsidiary in Canada jointly finance the development of new software. Each company pays part of the development costs and, in return, is granted the right to use the finished software in its own market.
To ensure that each participant makes a fair, arm’s-length contribution toward the cost of developing the IP, the IRS has established specific rules. For a detailed discussion of cost-sharing arrangements, see the following article.
When related companies share or use intellectual property (IP), an important question is:
This question matters because legal ownership of an IP asset does not automatically mean the legal owner is entitled to all profits generated by that IP.
For example, suppose a U.S. parent company develops a valuable software platform but registers the patent or other IP in the name of its foreign subsidiary. The foreign subsidiary may be the legal owner on paper, but that does not necessarily mean it should receive all of the profits from the IP.
For transfer pricing purposes, the IRS looks at what each related company actually does, what it contributes, and what risks it takes. This is known as a functional analysis.
A functional analysis generally looks at the following activities, often referred to as DEMPE:
The analysis may also consider questions such as:
Functional analysis is important because it helps determine how IP profits should be allocated among related companies.
For example, suppose a foreign subsidiary legally owns a patent, but the U.S. parent:
In that situation, simply pointing to the foreign subsidiary’s legal ownership of the patent may not be enough to justify giving the subsidiary all of the profits from the IP.
The IRS may look at the actual economic contributions of both companies when determining an arm’s-length result.
In simple terms: transfer pricing looks beyond “Who owns the IP on paper?” and asks “Who actually created, developed, protected, and used the IP to generate value?”
This analysis is important because the answers help determine which company should receive the income and how much one related company should pay another for the use of the IP.
Since most IP is unique and has no direct market price, companies must estimate an arm’s-length value using one of several IRS-recognized methods. Here’s each one explained simply:
This method looks at royalty rates or license agreements between unrelated companies that deal with similar IP and uses them as a benchmark.
Analogy: If you want to know a fair rent for your apartment, you check what similar apartments nearby are renting for.
This is the preferred method when good comparables genuinely exist — but for unique IP, finding a truly similar deal is often difficult, which limits how often this method can be used in practice.
Instead of comparing the IP itself, this method compares the profits of the company using the IP to the profits of similar companies that don’t have such valuable IP.
How it works: Find a group of similar companies and figure out their typical, “routine” profit margin. If the company using the IP is earning noticeably more than that routine range, the extra profit is assumed to be coming from the value of the IP — and that excess becomes the basis for the royalty.
Used when both related companies contribute something genuinely valuable to developing or exploiting the IP — for example, one company invents the technology while the other builds the market and customer relationships for it.
Here, instead of trying to price the IP in isolation, the combined profit from the whole venture is split between the companies based on each one’s relative contribution.
The method involves working out the current value of the income the intellectual property is expected to produce—that is, forecasting future profits and then calculating the present value of those future amounts of money.
This method is used mainly when there simply aren’t any reliable comparable deals to look at. However, U.S. courts have pushed back hard against this method when it relies too heavily on speculative, optimistic future projections rather than real, grounded transaction data (see the court cases below).
The IRS asks that companies choose the method which is most appropriate in view of the facts and the data available — there is no universal solution.
When a patent is licensed, the royalty rate imposed should be the amount that two completely independent and unconnected businesses would actually have agreed upon.
In order to find this rate, the IRS considers factors such as:
In plain terms, setting a defensible royalty means answering a checklist of questions:
Several major court cases have shaped how IP transfer pricing disputes are decided in the U.S. — and they share a common theme.
The big takeaway: across all three cases, U.S. courts have consistently favored valuation methods based on real, comparable transactions over the IRS’s preference for speculative future-income projections. This has made comparables-based approaches (like the CUT method) the safer, more defensible path for companies to follow.
Getting the price wrong can be expensive. If the IRS makes a transfer pricing adjustment of $5 million or more — or determines that a company’s pricing was off by more than double (or less than half) the correct arm’s-length price — it can impose a 20% penalty on top of the additional tax owed.
The only real protection against this penalty is having proper contemporaneous documentation — meaning documentation prepared and filed by the tax return due date, not created after the fact once an audit begins. This documentation should include:
Companies that want even more certainty can also apply for an Advance Pricing Agreement (APA) with the IRS — essentially pre-negotiating and locking in an agreed royalty rate for several years in advance, which greatly reduces the risk of a costly dispute down the road.