Loading...
Loading...
A practitioner-focused tour of OECD transfer pricing methods—CUP, RPM, cost plus, TNMM, and profit split—with Amount B, functional analysis, and how to document method selection defensibly.
Borys Ulanenko
CEO, ArmsLength AI
Continue exploring
Browse the full resource library or contact us if you want recommendations for your specific use case.
Every intercompany transaction—whether a sale of goods, a service fee, a royalty, or a loan—must be priced as if it were agreed between independent parties at arm's length. This is the arm's length principle, the foundational standard reflected across OECD member countries under Article 9 of the OECD Model Tax Convention.
In practice, "arm's length" is not self-defining. You need a structured, defensible mechanism to translate the principle into a number. That is the purpose of transfer pricing methods.
Choosing the wrong method—or applying the right method incorrectly—is one of the most common causes of transfer pricing audits, penalty adjustments, and double taxation disputes.
The OECD Transfer Pricing Guidelines (Chapter II, ) frame the objective as finding the most appropriate transfer pricing method for the particular case. Importantly, the Guidelines are not method-neutral: where a traditional transaction method and a profit method can be applied with equal reliability, the traditional method is preferred; and where a reliable CUP exists, it takes precedence over all other methods.
If you are new to the vocabulary, start with our introduction to transfer pricing and the ultimate guide to the arm's length principle. This article focuses on how methods differ, when each tends to fit, and what practitioners get wrong in documentation.
The diagram below groups the five OECD standard methods by family—transactional (price or gross margin) versus profit (net margin or combined profit).
Click the diagram to open the full-resolution PNG in a new tab.
These methods analyze the price of the transaction itself (or a gross margin derived from it).
| Method | What it measures | Primary use case |
|---|---|---|
| Comparable Uncontrolled Price (CUP) | Transaction price | Commodities, raw materials, some IP licenses |
| Resale Price Method (RPM) | Gross margin of the reseller | Distributors with limited value-add |
| Cost Plus Method (cost plus) | Gross mark-up on costs | Contract manufacturers, many low-risk service providers |
These methods analyze net profit arising from the controlled transaction (or allocate combined profit).
| Method | What it measures | Primary use case |
|---|---|---|
| Transactional Net Margin Method (TNMM) | Net profit relative to an appropriate base | Most transaction types in practice |
| Transactional Profit Split Method (PSM) | Combined profit divided between parties | Unique intangibles, highly integrated operations |
Gross margin methods (RPM and cost plus) are typically more sensitive to accounting classification and often demand stricter functional comparability at the gross line. Net margin methods (TNMM) tolerate more functional variation—but that flexibility can reduce precision if PLI selection or comparables are weak.
The CUP method compares the price charged in a controlled transaction to a comparable uncontrolled transaction under comparable circumstances. If conditions are sufficiently comparable, the uncontrolled price provides a direct arm's length measure.
| Strengths | Limitations |
|---|---|
| Most direct method; closest to the arm's length ideal when CUP quality is high | Demands near-identical products, contracts, and economics |
| Often preferred when a credible CUP exists | Small differences in product or terms can swing price materially |
| Avoids profit-level comparables search when price comparables are sound | Often impractical for differentiated goods, bundled offerings, or novel intangibles |
describes CUP as the most direct and reliable way to apply the arm's length principle when a sufficiently comparable uncontrolled transaction exists—while acknowledging that achieving that level of comparability is often difficult in practice.
RPM starts from the resale price to an independent customer, then subtracts an appropriate gross margin that reflects the reseller's functions, assets, and risks.
Formula (conceptually): Transfer price ≈ Third-party resale price − Appropriate gross margin
| Strengths | Limitations |
|---|---|
| Well-suited to distribution-focused tested parties | Sensitive to how revenue, COGS, and discounts are classified |
| Less demanding on product comparability than CUP (in the right fact pattern) | Gross margins vary materially across business models even within industries |
| Widely accepted for distributor cases when functions align | Not a good fit if the reseller performs substantial manufacturing or holds unique intangibles |
For RPM, functional similarity is more important than product identity—but comparability of the property transferred still matters. The OECD allows broader product differences under RPM than under CUP, provided the reseller's functions, assets, risks, and economic circumstances are genuinely comparable. That said, comparing distributors across very different industries requires strong justification and careful adjustments; it is not a blanket license.
Amount B (in adopting jurisdictions) can provide a simplified and streamlined approach for certain baseline marketing and distribution activities, using a matrix-based return on sales outcome rather than a full RPM or TNMM study for eligible distributors. Scope is narrow: it is not universal, elective in some jurisdictions, and it explicitly excludes profiles with meaningful DEMPE functions (Development, Enhancement, Maintenance, Protection, and Exploitation of intangibles) or highly integrated structures. Treat adoption status, elections, and carve-outs as jurisdiction-specific.
If you need the parallel US terminology comparison, see CPM vs TNMM—US Comparable Profits Method aligns with OECD TNMM, not with RPM.
Cost plus applies an appropriate gross mark-up to the supplier's direct and indirect production costs (typically COGS). The mark-up should reflect functions, assets, and risks, and is tested using comparable independent companies (or transactions) with similar cost structures and routines.
Formula (conceptually): Transfer price ≈ Cost base + Appropriate gross mark-up
Gross vs net—do not conflate cost plus and TNMM. The OECD cost plus method applies a mark-up to production costs to arrive at a gross profit. By contrast, TNMM's net cost plus PLI applies operating profit to total costs (COGS + operating expenses) to measure operating profitability. Different cost bases, different profit lines, different comparables universe.
In OECD language, "cost plus" typically means mark-up on costs (gross return on costs). In US practice, "CPM" most commonly refers to the Comparable Profits Method—a net margin method aligned with TNMM. Do not conflate the two acronyms when talking to mixed US/OECD teams.
| Strengths | Limitations |
|---|---|
| Strong fit for many low-risk manufacturing or routine service entities | Sensitive to what is included in "cost" and how pass-throughs are treated |
| Economically intuitive where value is created through efficient execution | Mark-up benchmarking requires comparable cost accounting patterns |
| Supported by comparables in many routine services settings | Weak fit where intangibles, risks, or outbound marketing drive profit |
highlights typical strong use cases for cost plus—semi-finished goods sold between affiliates, long-term buy-and-supply arrangements, and many controlled services transactions.
For qualifying supportive services that are not part of the group's core business, the OECD offers a simplified approach under Chapter VII: a standard 5% mark-up on pooled costs without requiring a full benchmarking study to justify the mark-up. The service must genuinely be low value-adding (back-office, HR, accounting, IT support, etc.) and not involve significant intangibles or risk. Many jurisdictions have adopted this elective simplification—it is one of the most commonly used administrative shortcuts in day-to-day transfer pricing compliance.
LVAIGS eligibility is narrower than it looks: advisory, R&D support, procurement of raw materials for manufacturing, and treasury services are among the categories the OECD explicitly excludes. Document why a service qualifies before defaulting to the 5% mark-up.
TNMM examines the net profitability of the tested party relative to an appropriate base (commonly sales, total costs, or assets) and compares that outcome to independent comparable companies.
The profitability ratio is the profit level indicator (PLI). Common PLIs include:
| PLI | Formula (conceptual) | Typical use |
|---|---|---|
| Operating margin (OM) | Operating profit ÷ Revenue | Distribution, services |
| Net cost plus (NCP) | Operating profit ÷ Total costs | Contract manufacturers, shared services |
| Berry ratio | Gross profit ÷ Operating expenses | Very narrow: see caution below |
| Return on assets (ROA) | Operating profit ÷ Assets | Asset-heavy operations |
Berry ratio—use with extreme care. The OECD warns that Berry ratios are frequently applied in inappropriate circumstances. A Berry ratio is only suitable where the entity's value contribution is proportional to operating expenses and is not materially driven by the value or volume of products sold. Intermediaries that take inventory risk, perform significant sales or marketing functions, or contribute meaningful intangible value are generally not appropriate Berry ratio candidates. See our Berry ratio guide for the full conditions and worked examples.
For deeper PLI selection guidance, see our PLI selection guide and benchmarking study guide.
A TNMM analysis typically requires a structured comparables search, quantitative screens, adjustments where warranted, and a defensible arm's length range. The OECD does not mandate a specific statistical tool, but notes that methods such as the interquartile range may help narrow the range where residual comparability defects remain (see IQR calculation and IQR vs full range for jurisdictional practice).
Practice tip: insufficiently documented rejections are a common audit pain point. Authorities increasingly expect traceable, decision-grade rationale for why comparable candidates failed screening—not only why survivors passed.
AI-assisted benchmarking platforms can compress parts of this workflow while preserving an audit trail—particularly helpful when teams must defend accept/reject decisions at scale across many transaction types.
| Strengths | Limitations |
|---|---|
| Works across many business models when tested party selection is correct | Can mask economic differences if PLI/base mismatch |
| Large independent-company data availability for many jurisdictions | Requires disciplined tested-party and PLI narrative |
| Well-understood audit dialogue in many countries | Company-specific items (non-recurring costs, group synergies) can distort |
Many teams use TNMM as a secondary check even when another method is primary. If CUP or cost plus implies margins far outside a well-built TNMM range, expect auditors to ask for reconciliation—prepare that narrative early.
For in-scope baseline distributors in adopting jurisdictions, Amount B can replace a full TNMM comparable exercise with a matrix outcome—without eliminating TNMM for transactions outside scope, for entities with unique intangibles or DEMPE-heavy profiles, or in non-adopting jurisdictions.
PSM starts from combined profit and asks how independent parties would split it. Common variants include:
| Strengths | Limitations |
|---|---|
| Can be the only coherent approach when reliable one-sided comparables do not exist | Needs robust financial data for both parties (and often careful segmentation) |
| Reflects integrated value creation | Split keys and profit definitions are inherently judgment-heavy |
| OECD post-BEPS guidance emphasizes PSM for certain unique-contribution cases | Documentation intensity and controversy risk are typically higher |
discusses profit split in contexts involving unique and valuable contributions—not only as a method of last resort when everything else fails.
When PSM (or any method) is applied to transactions involving unique intangibles that are difficult to value at the time of transfer, the OECD's HTVI framework (Chapter VI) allows tax authorities to use ex-post financial outcomes to test whether ex-ante pricing assumptions were reasonable. This is an increasingly active audit tool: if actual profits from an intangible significantly exceed what was projected at transfer, authorities may adjust using the benefit of hindsight—unless the taxpayer can show the original valuation was based on reliable projections and that the deviation results from unforeseeable developments. Document valuation assumptions, sensitivity analyses, and contingent pricing mechanisms upfront.
The OECD also recognizes that when reliable comparables for intangibles do not exist, valuation techniques—particularly income-based or DCF-style approaches—may be useful as part of the arm's length analysis (Chapter VI). This is especially relevant for IP transfers, cost-sharing buy-ins, and restructuring compensation where PSM or CUP cannot produce a credible benchmark on their own.
Amount B carve-outs can be read as a scoping diagnostic: where simplified treatment is unavailable for a distributor because DEMPE-like functions or integration exist, expect authorities to scrutinize whether TNMM alone tells the full story—and whether PSM elements deserve consideration.
The OECD recognizes five standard methods but does not restrict analysis to them. explicitly permits the use of other methods when none of the five standard methods can be reliably applied—provided the alternative produces an arm's length outcome consistent with the Guidelines' principles.
In practice, the most common "other method" is some form of income-based valuation (often a discounted cash flow model). This is particularly relevant for:
Using an alternative method does not remove the documentation burden—in many ways it increases it. Taxpayers must demonstrate why the five standard methods are inadequate and why the chosen technique produces a more reliable arm's length result.
The OECD frames this as accurate delineation of the actual transaction: before selecting a method, you must document functions, assets, and risks and test whether the contractual allocation of those elements is consistent with actual conduct:
The party with more complex operations, more valuable intangibles, and more significant economically substantiated risk generally earns a higher return—and is often not the right tested party for one-sided methods.
Method selection should read like a conclusion from FAR and accurate delineation, not a template picked before the facts are nailed down. Weak FAR narratives—especially those that skip the risk-control and financial-capacity tests—are an easy opening for audit challenges.
For methodology on tested party choice, see tested party selection guide.
OECD guidance ( to ) requires selecting the method that produces the most reliable arm's length outcome given the facts, comparability, and data quality. The hierarchy matters: traditional transaction methods are preferred over profit methods when both are equally reliable, and CUP is preferred over all other methods when a reliable CUP exists. Practical selection factors include:
OECD vs US §482 terminology at a glance. The OECD's "most appropriate method" rule and the US "best method rule" (IRC §482) are functionally similar—both ask which method is most reliable for the facts. Key naming differences:
When working across OECD and US frameworks, map the terminology before comparing conclusions. See CPM vs TNMM for a full comparison.
| Transaction type | Often-primary methods | Common alternatives |
|---|---|---|
| Commodity purchase/sale | CUP | TNMM |
| Baseline distribution (Amount B eligible, adopting jurisdiction) | Amount B (SSA) | TNMM / RPM |
| Finished goods distribution | RPM or TNMM | CUP (rare, if resales are comparable) |
| Contract manufacturing | Cost plus or TNMM | — |
| Toll manufacturing | Cost plus | TNMM |
| Routine intra-group services | Cost plus, TNMM, or CUP (direct-charge where comparable third-party fees exist) | LVAIGS 5% simplified approach for qualifying supportive services |
| IP license (good external comparables) | CUP | TNMM sanity check |
| IP / intangibles (unique, integrated) | PSM | Residual structures with TNMM for routines |
| Highly integrated / digital operating models | PSM | Transactional methods only for clearly separable pieces |
| Financial transactions (Chapter X) | CUP-style pricing for loans (yield approach, credit analysis); specialized methods for guarantees, cash pools, hedging, captive insurance | Threshold question: is the purported debt respected as debt or re-characterized as equity? Risk-free / risk-adjusted return analysis for certain arrangements |
US practitioners should also reference the best method rule under IRC §482 and the specific regulations (§1.482-3 through §1.482-6) when documenting method selection for US-filed returns.
Amount B (incorporated into the OECD Guidelines as an Annex to Chapter IV, February 2024; effective for fiscal years beginning on or after 1 January 2025 in adopting jurisdictions) provides a Simplified and Streamlined Approach (SSA) for baseline marketing and distribution activities. It is technically a standardized TNMM application within a tightly defined scope—not a new "seventh method."
Scope is broader than just buy-sell distributors. Amount B also covers certain sales agents and commissionaires performing qualifying baseline activities. However, significant exclusions apply: commodities, digital goods, services, and non-distribution activities are out of scope, as are distributors performing DEMPE functions, holding unique intangibles, or operating in highly integrated arrangements.
The pricing framework has several moving parts:
Adoption is jurisdiction-specific: some countries have adopted Amount B on a mandatory basis, others on an elective basis (e.g., the US via IRS Notice 2025-04), and others have explicitly declined or expressed material reservations (e.g., Australia, New Zealand). There is also an Inclusive Framework political-commitment layer relevant to tax certainty for covered jurisdictions.
If Amount B is available in your jurisdiction, the compliance win is real for genuinely routine distributors—but you still need documentation proving in/out of scope (functional analysis remains load-bearing). Do not assume Amount B applies simply because an entity is labelled a "distributor."
Mutual agreement procedure (MAP) practice has been professionalizing steadily through BEPS Action 14 peer reviews and updated OECD procedural guidance—including the revised Manual on Effective Mutual Agreement Procedures (MEMAP, last updated in 2026). The trend is consistent: competent authorities increasingly expect clearer, earlier factual presentations, pre-MAP engagement, standardized position papers, and more structured workflows.
For transfer pricing teams, the operational takeaway for method selection is straightforward: the method narrative, comparables screening rationale, and supporting functional analysis must be in the file before the first audit interview—not reconstructed after a proposed adjustment lands. Well-documented method selection reduces MAP exposure; poorly documented positions create unnecessary controversy risk regardless of technical merit.
Transfer pricing methods are only as strong as the analysis underneath them. The most defensible files tend to share three traits:
If you want to tighten benchmarking execution specifically, start with the benchmarking study guide and the quantitative screening filters playbook.
There is no default "best" method. The OECD framework asks for the most appropriate method for the particular case, given reliability of comparables and data. However, the OECD does maintain a hierarchy: traditional transaction methods (CUP, RPM, cost plus) are preferred over profit methods when both can be applied with equal reliability, and a reliable CUP takes priority over all other methods. In practice, TNMM is common because net margin data is frequently available—but prevalence does not mean automatic correctness.
Conceptually, yes in many cases: both are one-sided net margin methods applied to a tested party. Legal citations, procedural rules, and exam narratives differ. See CPM vs TNMM.
When you can demonstrate high-quality comparability on price for the controlled transaction—often in commodities, certain raw materials, or some licensable IP with close uncontrolled evidence.
Because unique intangibles and integrated development may make one-sided benchmarks unreliable for either party. PSM is designed for consolidated profit allocation problems that TNMM cannot slice cleanly.
No. Amount B targets in-scope baseline distributors in adopting jurisdictions. Everyone else still needs a conventional method analysis (commonly RPM or TNMM) unless a different simplification applies.
Increasingly, teams lose credibility on how comparables were eliminated—not only the survivors. Thin or inconsistent rejection rationale is an avoidable unforced error.
Many groups use rolling three-year data sets for TNMM, but refresh timing should follow facts: material business changes, M&A integration, new intangibles, or shifted FAR can force an earlier update regardless of calendar rhythm.
For qualifying supportive intra-group services that are not part of the core business, OECD Chapter VII offers an elective simplified approach: a 5% mark-up on pooled costs with no requirement for a full benchmarking study. The service must genuinely be low value-adding—advisory, R&D support, procurement of raw materials, and treasury services are among the excluded categories.
HTVI refers to intangibles that are difficult to value reliably at the time of transfer—often because projections of future income are speculative. The OECD's HTVI framework (Chapter VI) allows tax authorities to use ex-post financial outcomes to challenge ex-ante pricing. Taxpayers can rebut by showing the original valuation was based on reliable projections and that deviations resulted from unforeseeable events.
Yes. permits "other methods" when none of the five standard methods can be reliably applied—provided the alternative technique (commonly DCF or income-based valuation) produces an arm's length result consistent with OECD principles. The documentation burden is higher: you must explain why the standard methods are inadequate.