CPM vs TNMM: US and OECD Methods Compared

CPM and TNMM are the same idea — one-sided profitability testing — in two regulatory dialects. CPM is the US method (Section 1.482-5); TNMM is the OECD method that replaced CPM in the 2010 OECD Guidelines and is the workhorse in India (Rule 10B). Practitioners who move between US and Indian files stumble over the difference more than the label suggests, because the two methods treat the comparables differently.

Regulatory lineage

	CPM	TNMM
Home	US Treas. Reg. Section 1.482-5	OECD Guidelines (2010 revision, Ch. 3.3 / Ch. 5.4), India Rule 10B
Origin	The one-sided method that survived the 1994 US final regulations	Built from CPM by replacing transaction comparables with a company pool + IQR
Status in India	Not listed in Rule 10B	Listed (Rule 10B(1)(e)); the de facto default
Status in the US	A listed Section 482 method	Recognized in practice via the same mechanics (the IRS’s CPM examination is effectively pool-based)

The 2010 OECD revision is the hinge: the OECD kept CPM’s one-sided logic and its PLI menu, but dropped the requirement that comparables be independent transactions of the same kind. Comparables became companies whose profitability, on the same PLI, forms a distribution — and the arm’s length range became the IQR of that distribution.

The PLI difference

CPM’s regulation lists six indicators (gross margin, net profit margin, cost-based mark-up, selling price-to-value-added, inventory mark-up, other reasonable indicator) — a menu chosen case by case. TNMM practice narrows the menu:

• OP/OC (operating profit on operating cost) — the Indian default for service and R&D entities.
• OP/Sales — distribution and trading.
• NCP / cost-based mark-up — where the cost base is the natural base.
• ROCE / ROA — capital-intensive fact patterns.

The difference in practice: a CPM file must defend the PLI choice against the regulatory list; a TNMM file must defend the PLI choice against the pool’s behaviour (variance across the comparable companies, stability across periods). The second test is more mechanical — which is one reason TNMM won the practitioner default.

What changes when you run the benchmark

Step	CPM	TNMM
Comparable selection	independent parties doing the same kind of transaction	companies with comparable FAR, filtered to a pool
Data grain	transaction-level where possible	company financial statements
Adjustment	for differences affecting the PLI	for differences that materially affect the PLI (WC adjustment is the standard tool)
Output	the controlled party’s PLI vs the comparable’s PLI	the tested party’s PLI vs the pool’s IQR
Failure mode	no true transaction comparables exist	pool too thin, too mixed, or PLI not isolating

Two practical consequences:

1. Sourcing. CPM comparables are scarce — independent parties running the same contract with disclosed, same-year, same-PLI data is a thin set. TNMM pools are assembled from public filings and commercial databases; the constraint is screening quality, not existence.
2. The range. CPM produces a comparison against one or a few comparables — a point or a narrow band, sensitive to each comparable’s quality. TNMM produces a distribution — the IQR absorbs single-company noise, and the tested party’s position in the range becomes the argument, not the exact margin.

Where the TPO and the IRS diverge

• US (CPM examination). The IRS examines the tested-party selection, the PLI, and the comparability adjustments — and pushes hard on the cost base and on whether the “least complex” selection is right. The method label (CPM vs TNMM) matters less than the one-sided discipline.
• India (TNMM examination). The TPO examines the pool (the Accept-Reject matrix), the PLI, the period, and the working capital adjustment — and method substitution (the TPO’s method replacing the declared one) is the standard escalation. A file that is really a CPM analysis wearing a TNMM label — transaction comparables, no pool, no IQR — will be examined as flawed TNMM, not forgiven as clever CPM.

The rule of thumb

• US file: use the method the fact pattern supports under Section 1.482; the CPM label is fine when the transaction comparables are real.
• Indian file: run TNMM with a real pool. If the data you actually have is transaction-level (a handful of true transaction comparables), document them inside the TNMM analysis — as the source of the mark-up reference — rather than relabelling the method.
• Cross-border group file: pick one method and one cost definition across jurisdictions, and let the rule numbers differ (Rule 10B in India, Section 482 in the US) while the economics stay identical.

See also

• CPM guide
• TNMM — the complete guide
• How to choose a method