RoyaltyStat Blog

The Berry Ratio is Illegitimate Under the TNMM

Posted by Ednaldo Silva

Le secret d’ennuyer est celui de tout dire. Voltaire (1694-1778)

The Berry ratio is vulnerable to the flexible accounting allocation of costs and expenses among the tested party and its comparables.

The Berry ratio was introduced in DuPont de Nemours & Co. v. United States, 608 F.2d 445 (Ct. Cl. 1979) when Charles Berry, then professor of economics at Princeton University, proposed a substitute to the resale price and the cost-plus (gross profit) methods under the 1968 US transfer pricing regulations. Berry was consulted in the IRS drafting of Rev. Proc. 63-10, 1963-10 C.B. 490, a progenitor of the 1968 regulations.

The Berry ratio is specified as a linear relation between gross profits and operating expenses.

Like the profit split method, the Berry ratio was considered an unspecified “fourth method” until it was incorporated under the CPM in the 1994 US transfer pricing regulations.  The Berry ratio is a nullius filius because under the CPM (or under the TNMM born in 1995), the applicable numerator is operating profits, and not gross profits. Thus, under the CPM or TNMM, the Berry ratio has an inconsistent dependent variable (numerator).

In addition to assuming that COGS and XSGA (i.e., operating expenses using Standard & Poor’s Compustat mnemonics) are reported (allocated) on the income statements of the tested party and the comparables in consistent fashion (as stated below, this strong assumption cannot be verified), two other key assumptions are made: first, the purchasing of good for resale (but not for further processing) is reflected in COGS; second, the marketing and selling (distribution or retail) functions are reflected in operating expenses:

     (1)     G(i) = β X(i) + U(i)

measured from i = 1 to N comparables to the tested party.

The variable G denote gross profits, X denotes operating expenses (XSGA), and U represents random uncertainty. The parameter β > 1 is the Berry ratio.

To make the Berry ratio a consistent (legitimate) CPM or TNMM profit indicator, equation (1) must be transformed such that operating profits become the prescribed dependent variable:

     (2)     P(i) = γ X(i) + V(i)

where the slope coefficient γ = (β – 1) > 0 is the operating profit markup on operating expenses alone, and the variable V represents random uncertainty.

The dependent variable P(i) = [G(i) – X(i)] denotes operating profits; and equation (2) is obtained by subtracting X(i) from (1).

Berry (1999) was cautious about certain applications of his novel gross profits ratio. According to Berry, the tested party and the proposed comparables must have similar COGS composition; and COGS must exclude direct labor (which is indicative of other value-added functions, such as assembly, blending, packaging or manufacturing, separate from selling). Also, operating expenses must exclude depreciation because depreciation is “related primarily to [asset purchase] timing.” Berry (1999), Footnote 6.

Accounting flexibility of cost and expense allocations between COGS and XSGA (operating expenses) is a major obstacle to using the Berry ratio. To make matters worse, SEC filings do not report adequate information (even considering the disclosed accounting footnotes) to ascertain the similarity of allocations between cost and expense among the selected comparables. Thus, the similarity of cost and expense allocations cannot be subject to empirical verification.

More diffused operating profit indictors, such as the operating profit markup (or the derived operating profit margin) under the CPM or TNMM, are exempt from the flexible accounting allocation problems affecting the Berry ratio, because COGS and operating expenses (XSGA) are combined (summed together).

Using equation (2), the reliability of the Berry ratio can be tested against the rival operating profit markup on total cost, including the sum of COGS and XSGA as the independent variable; and accepted measures of statistical reliability (such as the t-statistics of the slope coefficient (γ) compared to the coefficient of the profit markup equation on total cost, and the regression R-squared) can demarcate the most appropriate profit indicator calculated with comparable data to the tested party.


Charles Berry, “Berry Ratios: Their Use and Misuse,” Journal of Global Transfer Pricing, April-May 1999, reprinted by CCH Inc.

Martin Przysuski & Srini Lalapet, “A Comprehensive Look at the Berry Ratio in Transfer Pricing,” Tax Notes International, Vol. 40, No. 8, November 21, 2005.

This article by Przysuski & Lalapet is interesting and benefits from the joint authors’ discussion with Berry. However, irrelevant to the authors’ central message, it is misconceived (p. 761) that the CPM or TNMM depends on the assumption of equal rates of return. The ideal equal rates of return assumption underlined the defeated BALRM, but not the victorious CPM. The CPM is based on the realist assumption that operating profits depend on market structure (memorialized in the US transfer pricing regulations and the OECD guidelines that operating profits are dependent on functions performed, assets employed, risks assumed, and geographic market).

Also, the article is wrong claiming (p. 762) that return on assets is “firmly grounded in economic theory.” In economics, return on assets (called “return on capital”) is defiled by long-standing controversies.

See Joan Robinson, “The Measure of Capital: The End of the Controversy,” Economic Journal, Vol. 81, No. 323, September 1971. Stable URL:

Unlike Jim Morrison’s song “The End” (1967) (“When the music’s over, turn out the lights”), the capital controversy is unresolved. See the survey by Geoff Harcourt, “Some Cambridge Controversies in the Theory of Capital,” Journal of Economic Literature, Vol. 7, No. 2, June 1969. Stable URL:

See also Avi Cohen & Geoff Harcourt, “Whatever Happened to the Cambridge Capital Theory Controversies?” Journal of Economic Perspectives, Vol. 17, No. Winter 2003. Stable URL:

The 1968 US Treasury Regulation section 1.482-2(e)(1)(ii) described three methods to determine arm’s length prices, listed by priority: (a) comparable uncontrolled price (CUP) method, (b) resale price method, and (c) cost-plus method. The courts found their own “fourth methods” in the absence of comparables.

A profit split was used by the court in Eli Lilly & Co. v. IRS Commissioner, 84 TC (1985); hence, the Lilly method was another name for the profit split. As a drafter of the US regulations, I heard the backstage dictum: “We must adopt rules to prevent the Lilly [aka judicial profit split] method from becoming a legal precedent.”

Avi-Yonah contains 31 citations to the Lilly litigation. See Reuven Avi-Yonah, “Transfer Pricing Disputes in the United States” (2012), available at:

As stated, the Berry ratio was another “fourth method” based on comparables.

Published on Dec 24, 2020 3:03:29 PM

Ednaldo Silva (Ph.D.) is founder and managing director of RoyaltyStat. He helped draft the US transfer pricing regulations and developed the comparable profits method called TNNM by the OECD. He can be contacted at:

RoyaltyStat provides premier online databases of royalty rates extracted from unredacted license agreements
and normalized company financials (income statement, balance sheet, cash flow). We provide high-quality data, built-in analytical tools, customer training and attentive technical support.

Schedule an Online Demo 

Topics: Net Profit Indicator, CPM/TNMM, Berry Ratio