For years, the Profit Split Method (PSM) carried a reputation as the method practitioners reached for only when nothing else worked, when comparables were unavailable, when both parties to a transaction made unique and valuable contributions, or when transactions were too integrated to test one side in isolation. That reputation is shifting. As global value chains grow more integrated and intangible-driven, PSM is increasingly the method that best reflects economic reality, particularly in digital, platform, and R&D-intensive businesses where no single entity can meaningfully be treated as the routine, tested party.
When PSM Is the Right Fit
The OECD identifies PSM as most appropriate where both parties to a transaction make unique and valuable contributions, typically unique intangibles, such that a one-sided method testing only one party would fail to capture the value the other party contributes; where the transaction is so highly integrated that it cannot reliably be evaluated on a separate, transaction-by-transaction basis; and where the parties share the assumption of economically significant risks, or separately assume closely related risks, in relation to the same transaction.
Contribution Analysis vs Residual Analysis
PSM can be applied through a contribution analysis, which divides combined profits based on the relative value of each party's functions, assets, and risks, typically requiring some form of external valuation benchmark to anchor the split. More commonly, groups apply a residual analysis in two stages: first, each party is allocated a routine return for its non-unique functions using a one-sided method like TNMM or cost-plus; second, any residual profit remaining after those routine returns is split between the parties based on the relative value of their unique contributions, most often their respective intangibles.
The Splitting Factor Problem
The most contested element of any PSM application is the splitting factor, the metric used to divide the residual profit. Common factors include relative R&D spend, relative marketing spend, headcount of key value-creating personnel, or capitalised intangible development costs. None of these factors is a perfect proxy for value creation, and tax authorities frequently challenge the chosen factor as understating their own jurisdiction's contribution, particularly where a market jurisdiction argues that local sales, marketing, or customer relationships deserve greater weight than a pure R&D-cost-based split would assign them.
PSM and Digital Business Models
The rise of digital and platform business models, where user data, network effects, and platform-side intangibles are jointly created across multiple jurisdictions, has pushed PSM from an exceptional method toward a mainstream one for these sectors. Amount A of Pillar One, though a separate mechanism from traditional bilateral transfer pricing, itself borrows the profit-split logic in reallocating a portion of residual profit to market jurisdictions, reflecting a broader policy convergence toward profit-split thinking wherever integrated value creation resists one-sided testing.
Testing the Result for Reasonableness
Because PSM outcomes depend so heavily on the splitting factor chosen, a prudent practitioner tests the result against alternative splitting factors and against a sanity check of what each party's implied return looks like on a standalone basis. A split that leaves one party earning materially less than a routine return for the functions it performs, despite ostensibly holding unique intangibles, usually signals that the splitting factor has been misapplied rather than that the underlying value creation genuinely favours the other party so heavily.
Conclusion
Under Chapter II, Part III, PSM is appropriate only where unique and valuable contributions exist on both sides of a highly integrated transaction, and its defensibility rests almost entirely on the splitting factor chosen for the residual profit stage. A PSM position should be tested against at least one alternative splitting factor before filing, since the factor, not the method itself, is what a tax authority will contest first.
