Skip to main content
Quartyl
Methods & PLIsprofessional

Profit Split vs TNMM: One-Sided vs Two-Sided Approaches

Profit Split vs TNMM: when both parties are non-routine, why one-sided margin testing fails, the comparability burden, and how Indian TPOs respond to a split argument.

Quartyl Team

The choice between TNMM and Profit Split is not a method preference — it is a statement about the transaction. TNMM says: one party’s contribution is routine, so price it against the market and let the rest fall where it falls. Profit Split says: no party’s contribution is routine, so no one-sided reference prices the transaction — allocate the combined profit.

Getting that statement wrong is the most expensive method error in transfer pricing, in both directions.

The decision test

Ask one question: can one party’s contribution be priced against an independent market reference, such that the transaction is resolved?

Answer Method Why
Yes — one party is genuinely routine (limited-risk distributor, contract manufacturer, routine service provider) TNMM (or CUP/Cost Plus where the price evidence is better) The routine side has a market price/margin; the residual value belongs to the other side by construction
No — both parties contribute unique intangibles, entrepreneurial functions or risk Profit Split Pricing only the routine-looking side against comparables misallocates: the “routine” side is not routine, and the comparables do not capture its value
Transactions are integrated and cannot be priced separately Profit Split (or a documented single-entity view) There is no separable controlled price to test

The trap in direction one: declaring TNMM on a transaction where the Indian entity performs key development or holds valuable intangibles. The TPO accepts the one-sided framing, benchmarks the Indian entity against comparable routine providers, and the entire entrepreneurial profit stays abroad — at the cost of a file that cannot be defended if the characterization is challenged later. The trap in direction two: declaring a split where one side is genuinely routine — the TPO rejects the split, substitutes TNMM, and the taxpayer has bought complexity for nothing.

The comparability burden

TNMM Profit Split
Sides examined one (the tested party) both
Reference a pool of comparable companies none — the parties’ own contributions
Key evidence the Accept-Reject matrix, the PLI the DEMPE allocation, risk control, the integration analysis
Data tested party financials + pool both parties’ financials + contribution analysis
Sensitivity pool quality, PLI choice the allocation key — the single most examined number in the file

A TNMM file lives or dies on the pool. A split file lives or dies on the allocation key — and unlike a pool, the key has no market to point to. It is constructed from the functional record, and it is where the examination goes.

Worked contrast: the captive that is more than a captive

Party A (India): develops and maintains a software platform for Party B (US), which owns the IP and sells subscriptions. Booked: A earns 18% OP/OC; B earns a 28% net margin on subscriptions.

TNMM framing. A is the tested party (the “less complex” entity — it owns no IP, earns fees). Comparable Indian software providers: OP/OC IQR 16%–24%. A’s 18% is inside the range. Done. The 28% with B is B’s — B owns the intangible, so the residual belongs to B. Correct if A’s work is truly contract development under B’s direction. Wrong if A engineers the product roadmap, holds the key patents, and B is primarily a commercial wrapper.

Split framing. Both contributions are non-routine: A’s development is a key value driver (the product’s market position is built on it); B’s commercialization and IP ownership are entrepreneurial. Routine returns paid first (A’s development at its benchmarked cost-plus; B’s distribution at its limited-risk return), then the residual allocated on the DEMPE contribution analysis — say 40:60. A’s total return rises materially above the TNMM position; the residual stays partly in India.

The same books, two economics. The TPO will default to the TNMM framing unless the functional record forces the split — which is why the DEMPE and risk-control evidence (who decided, who funded, who bore the loss years, who owns the IP and the decision rights) is the whole game.

How the TPO responds

  • Split declared, routine characterization wins. The TPO rejects the split, substitutes TNMM on the Indian party, and the file reverts to the one-sided answer — usually at the lower return. This is the standard outcome where the Indian entity’s contracts say “under our direction”.
  • Split declared, the functional record is strong. The TPO examines the DEMPE allocation line by line, often accepting a different key rather than the method — the adjustment is the split, not the substitution. Appeals (AAR/ITAT) are where the key is argued properly.
  • TNMM declared on a non-routine party. The TPO may accept the TNMM result in the year (it usually gives the Indian party less), but the characterization sits exposed — a later year, a different officer, or a connected CbCR/DEMPE review can reopen the method choice.

The documentation that decides it

  • The characterization memo: functions, assets, risks of both parties, with the conclusion on whether either contribution is routine — written before the method is chosen, not after.
  • For a split: the DEMPE allocation with primary evidence (contracts, board minutes, IP registry, funding records, decision rights), the routine-return benchmarks for the first step, and the allocation key with sensitivity — what the result is if the key moves 10 points either way.
  • For TNMM on a party the group calls routine: the record showing it is — contracts, direction, risk bearing — because the TPO’s default is to believe the characterization only as far as the evidence supports it.

See also

Run the screens as a study, not a spreadsheet

Quartyl applies the method, PLI and screening steps above as a pipeline — and keeps a documented reason for every exclusion.

Related docs

Book a Demo

Tell us what you'd like benchmarked

We'll confirm a 30-minute screen-share slot within one business day.

We reply within one business day. Your details are used only to arrange the demo — never shared or sold.