Transfer Pricing for a Foreign Subsidiary: Cost-Plus R&D Agreements Explained

Short answer: pay your foreign R&D subsidiary its costs plus a benchmarked markup, under a written intercompany service agreement, and keep a file showing how you arrived at the markup. That is the cost-plus model, and it is the default structure for a startup subsidiary that works only for its parent. This article is the detailed companion to our complete guide to transfer pricing for startups, and it stays on one question: how a cost-plus arrangement for a foreign subsidiary is built and defended.

Why Does a Foreign Subsidiary Create a Transfer Pricing Issue?

The standard comes from IRC Section 482 in the U.S. and the OECD Transfer Pricing Guidelines elsewhere, which the OECD encourages member and non-member countries to follow. If the fee is too high, the U.S. loses tax base; if it is too low, the foreign country does. The issue appears the day the subsidiary has employees, regardless of size. Founders often set up the entity, wire it money to cover payroll, and never put a price on the relationship, which works until a diligence request, a local filing, or an audit asks what the arrangement is.

What Is Cost-Plus Transfer Pricing?

Under a cost-plus (or cost-of-services-plus) arrangement, the subsidiary performs development work for the parent and is paid its operating costs plus a markup. The parent owns the resulting IP and bears the commercial risk that the product succeeds or fails. The subsidiary earns a modest, predictable return, which is what an independent contract R&D firm would accept.

This is sometimes called a contract R&D or routine service provider model. In transfer pricing terms, the subsidiary is the tested party: because it performs simpler, more routine functions than the parent, its margin is the one benchmarked against independent companies.

What Should the Intercompany Agreement Say?

A usable agreement is specific enough that someone outside the company could recompute the fee from it. At a minimum it should state:

• Services. What the subsidiary does, for whom, and under whose direction.
• IP ownership. That the parent owns all work product, with a present assignment of rights.
• Cost base. Which costs are included (salaries, benefits, rent, equipment, local professional fees) and which are excluded or passed through at cost.
• Markup. The percentage, or the formula and review cadence used to set it.
• Billing and payment. Invoice frequency, currency, and payment terms.
• True-up. A mechanism for year-end adjustment so the subsidiary lands at the agreed margin rather than whatever the budget produced.
• Taxes. Responsibility for value-added tax, withholding, and similar charges.

A Worked Example of Cost-Plus Pricing

Suppose a Delaware parent funds a wholly owned Israeli subsidiary with 15 engineers. For the year, the subsidiary's operating costs (salaries, benefits, rent, equipment, and local professional fees) are $2,000,000. The benchmarking study supports a cost-plus markup of 10% for comparable contract R&D providers.

The parent treats the $2.2 million according to its own tax rules for research costs, and the subsidiary reports $200,000 of profit locally. The 10% is illustrative only; the supportable markup depends on the functions performed and the comparable companies selected. If costs come in at $2,300,000, the true-up raises the fee to $2,530,000 so that the subsidiary still earns 10% on actual cost.

How Is the Markup Determined?

The markup is normally supported by a benchmarking study. An analyst screens databases of independent companies that perform similar development or technical-services functions, narrows the set to those that are genuinely comparable, and computes a profit level indicator such as net cost plus (operating profit divided by total costs). The result is an interquartile range, and a markup inside the range is generally treated as arm's length. Under Treas. Reg. §1.482-1, if a result falls outside the arm's-length range, the adjustment may be to any point in the range, and when the interquartile range is used it will ordinarily be to the median.

A services subsidiary is typically tested under the comparable profits method or the cost of services plus method in Treas. Reg. §1.482-9. The services cost method, which allows certain covered services to be charged at cost with no markup, lists research, development, or experimentation (and engineering or scientific activities) among its excluded activities, so an R&D subsidiary's fee is not set under that method.

How Does Cost-Plus Compare With the Alternatives?

How Do You Put a Cost-Plus Arrangement in Place?

The order matters, because documentation has to exist by the time the return is filed and the agreement should precede the period it governs:

1. Describe the functions and risks. Write down who performs the development work, who directs it, who funds it, and who bears the risk that it fails.
2. Define the cost base. List the cost categories included in the markup base and reconcile them to the subsidiary's ledger.
3. Select the method and tested party. For a captive R&D subsidiary this is usually a cost-based or profit-based method tested on the subsidiary.
4. Benchmark the markup. Obtain a comparables search that supports an arm's-length range for the subsidiary's profit on costs.
5. Sign the agreement. Execute it before, or as early as possible in, the period it covers, and invoice according to its terms.
6. Track and true up. Compare actual costs with budget during the year and settle the difference at year-end so the subsidiary lands at the agreed margin.
7. Assemble the file. Keep the agreement, the functional analysis, the benchmarking, and evidence that the parties followed the agreement.

Why Isn't a Contract R&D Label Enough on Its Own?

Tax authorities look at what the parties actually do, not only at what the agreement says. The OECD Transfer Pricing Guidelines set out a six-step process for analyzing risk in a controlled transaction. Two of the steps matter most here: the analysis checks whether the contractual assumption of risk is consistent with the parties' conduct, including whether the party assuming a risk actually controls it and has the financial capacity to bear it; and where it does not, the risk is allocated to the party that does.

For a cost-plus subsidiary, that means the facts should match the label. If the subsidiary's engineers set the product roadmap, decide what to build, and fund their own work, a routine-service return may be hard to defend, and the agreement should be revisited. If the parent's management makes those decisions and funds the work, the cost-plus label and the conduct line up.

What Mistakes Do Startups Commonly Make?

• No agreement. Money moves as informal funding and the service relationship is never written down.
• Reimbursing costs with no markup. An independent service provider would expect a profit, so cost recovery alone is hard to defend as arm's length.
• An undefined cost base. Without a clear list of included and excluded costs, the base drifts and cannot be reconciled to the ledger.
• Ignoring equity compensation. Options granted to subsidiary staff are a real cost; set the treatment deliberately, using a 409A valuation for the fair market value input.
• Never updating. A subsidiary that began as a contract developer and now makes product decisions may no longer fit a cost-plus label.
• Inconsistency with other documents. Pricing that allocates most profit abroad while the 409A treats the U.S. entity as earning everything invites questions; see international startups and 409A.

What Are the Penalties and Filing Requirements?

In the U.S., IRC Section 6662(e) applies a 20% penalty to the underpayment attributable to a net Section 482 adjustment above the lesser of $5 million or 10% of gross receipts, and 40% above the lesser of $20 million or 20% of gross receipts. The net-adjustment penalty can be avoided if the taxpayer applied a specified Section 482 method reasonably, had documentation in existence when the return was filed, and provided it to the IRS within 30 days of a request.

Separately, a U.S. person who is a Form 5471 filer, for example one that controls (owns more than 50% of the vote or value of) a foreign corporation, reports on Form 5471, with transactions between the foreign corporation and related persons reported on Schedule M by controlling persons; a U.S. corporation that is at least 25% foreign-owned, or a foreign corporation engaged in a U.S. trade or business, reports on Form 5472. Failures carry their own penalties even when the pricing is correct: $10,000 for each annual accounting period of each foreign corporation under Section 6038(b) for Form 5471 (with continuation penalties of up to $50,000), and $25,000 for each taxable year under Section 6038A(d) for Form 5472 (plus $25,000 for each 30-day period that the failure continues more than 90 days after IRS notice). Local rules are independent: Israel (Section 85A of the Income Tax Ordinance), India (Chapter X of the Income-tax Act), and Canada (Section 247 of the Income Tax Act) each have arm's-length and documentation requirements with different deadlines and thresholds.

How Does Cost-Plus Pricing Relate to a 409A Valuation?

They answer different questions: a 409A valuation sets the fair market value of common stock for option pricing, while transfer pricing sets the intercompany fee. They intersect where options are granted to foreign-subsidiary employees and in how the company's financial profile is described to different authorities. If the intercompany arrangement changes materially, review whether the change is a material event; see when to update your 409A. For the broader transfer pricing picture, start from our transfer pricing guide, or return to the 409A Valuations home page.

Note: this article is educational, not legal or tax advice. Transfer pricing depends on facts and on the law of each country involved.