CUP vs TNMM: When Direct Price Evidence Beats Margin Benchmarking

CUP and TNMM sit at opposite ends of the methods spectrum. CUP is the most direct evidence that exists — an actual price an unrelated party pays for the identical transaction. TNMM is the most usable evidence — a pool of comparable companies and a net profitability indicator that can be assembled for almost any routine fact pattern. Choosing between them is the first real decision in the study, and it drives everything downstream.

The fundamental difference

	CUP	TNMM
What is compared	the price itself	a PLI of the tested party vs a pool
Evidence type	a transaction	a distribution of companies
“Comparable” means	identical product/service, comparable circumstances	companies with comparable functions, assets, risks
Adjustment burden	high bar — only material, quantifiable differences	lower — differences that materially affect the PLI
Output	a price (or a narrow price band)	a range (IQR) and the tested party’s position in it
Data source	internal third-party sales, exchanges, unrelated contracts	comparable company financials
Where it is strongest	commodities, internal CUP, regulated prices	services, distribution, most Indian fact patterns
Where it fails	the moment nothing is “identical”	a weak pool, a wrong PLI, related-party leakage in the pool

The best method rule points at CUP first: where a genuine uncontrolled price exists, no method produces a more reliable result. TNMM is the fallback that becomes the default — because for most transactions a genuine CUP does not exist.

When CUP wins

• Commodity products with published or exchange-reported prices (metals, chemicals, energy). The uncontrolled price is public; the adjustment list is short (quality grade, delivery point, volume).
• Internal CUP — the same enterprise sells the identical product to unrelated third parties in the same market. The internal price is the strongest external evidence available, and the file’s work is showing the two transactions are comparable (terms, quantity, timing, credit).
• Regulated or contractually fixed prices where the “market” price is the published price.

When TNMM wins

• Services — bespoke services have no identical uncontrolled twin. The question is what a comparable service provider earns (OP/OC), not what an unrelated party paid for this service.
• No internal comparability — the group sells the product only to the affiliate, and the product is not traded independently.
• The price varies with negotiation — where every contract is different, a single uncontrolled price is noise; a pool of margins is signal.

Worked comparison on one fact pattern

An Indian manufacturer sells a standard industrial component to its overseas subsidiary. The same component is also sold to unrelated third-party buyers in India.

CUP route. Third-party sales of the identical grade: ₹410 per unit, ex-works, 60-day credit, 10,000-unit minimums. Controlled sale: ₹380, ex-works, 90-day credit, 50,000 units. Adjusting the third-party price for the extra credit period (financing cost on the additional 30 days) and the volume discount the affiliates do not get… the uncontrolled reference collapses toward the controlled price. The differences are real and quantifiable, but once adjusted the CUP says the controlled price is close to arm’s length — with the residual argument living entirely in the adjustment schedule.

TNMM route. The manufacturer is the tested party (the subsidiary is the limited-risk reseller). Comparable Indian manufacturers of comparable components return OP/OC of 9.1%–14.6% (IQR 10.2%–12.8%). The manufacturer’s OP/OC is 11.9% — inside the range. No price adjustment, no credit-period argument, no volume-discount argument: the margin is in the range and the study closes.

Same transaction, two files. The CUP file is more precise and more fragile — every adjustment is examinable. The TNMM file is less precise and more durable — the range absorbs the differences that the CUP route had to quantify.

The fallback rules

1. CUP exists and is clean → CUP. Document the source, the identity analysis and the adjustment list. Do not voluntarily step down to TNMM when a genuine CUP supports the price — the step-down becomes an unexplained choice.
2. CUP exists but is dirty (similar, not identical; different market; heavy adjustments) → say so, then move. The file should contain the CUP analysis and the TNMM result, with the documented reason the CUP evidence is unreliable for this transaction. A CUP that was considered and rejected with reasons is far stronger than one that was never looked for.
3. No CUP → TNMM with the PLI chosen for the tested party’s function, and the full screening discipline behind the pool.

What the TPO examines

• If CUP is declared: the identity analysis (is it really identical?), the source of the uncontrolled price, and each adjustment line by line. An internal CUP is examined on the other direction too — the third-party sale itself, if it is also a related-party-adjacent transaction.
• If TNMM is declared after a CUP existed: the documented reason for the step-down. “We used TNMM” with a CUP sitting unused in the file is a method-choice objection waiting to happen.

See also

• CUP method — complete guide
• TNMM — the complete guide
• How to choose a method
• Best method rule