(C) Tax Analysts 2005. All rights reserved.

Tax Analysts does not claim copyright in any public domain or third party content.
Volume 40, Number 8 November 21, 2005

A Comprehensive Look at the

Berry Ratio in Transfer Pricing

by Martin Przysuski and Srini Lalapet

Reprinted from Tax Notes Int’l, November 21, 2005, p. 759

 Special Reports

A Comprehensive Look at the Berry Ratio

(C) Tax Analysts 2005. All rights reserved. Tax Analysts does not claim copyright in any public domain or third party content.
In Transfer Pricing
by Martin Przysuski and Srini Lalapet
in, that is, manufacturing, distribution, or service
Martin Przysuski is a partner with BDO provision, as well as the economic drivers that
Dunwoody’s Transfer Pricing and Competent govern its profitability. For instance, a manufactur-
Authority Services Team in Toronto, and a er’s profitability is governed by how efficiently it
part-time student in the Master of Taxation employs its operating assets, whereas a distributor’s
program at the University of Waterloo in Wa- profitability is governed by how successful its sales,
terloo, Canada. Srini Lalapet is a transfer marketing, and distribution efforts are in generat-
pricing economist with a diversified multina- ing sales to end-customers.
tional in the greater Toronto area.
The authors would like to thank Prof. Alan Indeed, choosing a wrong profit indicator ratio or
Macnaughton of the School of Accountancy at a misapplication of one would not only taint the
the University of Waterloo in Waterloo, entire transfer pricing analysis, but also mislead a
Canada, and Prof. (emeritus) Charles Berry at company into assuming that they are transacting at
Princeton University for their valuable time, arm’s length while in actuality they may be falling
comments, and guidance, which have greatly well short of the arm’s-length standard. Unfortu-
assisted in the writing of this article. nately, one ratio that has particularly suffered by
such misapplication is the Berry ratio. The purpose
of this article, therefore, is to reexamine the use of
profit indicators in transfer pricing analyses with a
T he theoretical basis for most transfer pricing
analyses performed in the world’s major econo-
mies is the arm’s-length principle, which stipulates
particular emphasis on the use and misuse of the
Berry ratio, one of the more interesting and unique
ratios used in transfer pricing analyses. However,
that related parties should transact with each other any such examination will not be complete without
as if they were dealing with independent third an understanding of the context underlying the use
parties. In theory, the principle is almost intuitive, of profit indicators in transfer pricing analyses in
although it is debatable whether it is reasonable to the first place. Therefore, we shall first briefly
apply the arm’s-length principle to multinational discuss the evolution of transfer pricing methods in
enterprises that often derive their competitive ad- the United States and in the OECD and in particu-
vantage based on the level of integration among lar investigate the use of profit-based methods,
their operating entities. Furthermore, in practice, whose application, as the name suggests, is based on
the application of the arm’s-length principle is the use of profit indicators. In addition, we shall
fraught with difficulty, particularly because of the discuss the various profit indicators that are typi-
enormous challenges associated with identifying cally used in a transfer pricing context, then exam-
arm’s-length transactions between independent ine in detail the Berry ratio to illustrate how it
third parties that might be comparable to the trans- should be ideally applied to yield the best possible
actions between two related parties within a multi- results and consequently also provide the best pos-
national enterprise. sible defense in the face of a transfer pricing audit by
In many transfer pricing analyses, practitioners the tax authorities.
have to resort to the so-called profit-based methods
to prove the arm’s-length principle due to the lack of The Evolution of Transfer Pricing
adequate comparable transactional data. In many Methods
such analyses, some type of ratio analysis is under-
taken to prove, albeit indirectly, that the entity in Given the difficulty of applying the arm’s-length
question has transacted at arm’s length with its principle, both the United States and the OECD
related companies. In this context, therefore, the have attempted, in radically different ways, to ad-
ratios that are used (or, in transfer pricing parlance, dress the issue and offer guidance on transfer pric-
profit indicators) are vital to an economically sound ing methods that can be used to prove that related
transfer pricing analysis. The choice of such ratios is parties have transacted at arm’s length. The United
usually governed by the activity an entity is engaged States was first off the mark with its 1968 transfer

Tax Notes International November 21, 2005 • 759

Special Reports

pricing regulations governing the application of In- contexts is that the United States accepts the best

(C) Tax Analysts 2005. All rights reserved. Tax Analysts does not claim copyright in any public domain or third party content.
ternal Revenue Code section 482. The OECD fol- method rule, which allows the use of any method —
lowed in 1979 with its own report containing com- transactional or otherwise — as long as it produces
prehensive guidelines on transfer pricing.1 the best measure of arm’s-length results under the
At that early stage, the emphasis in both the U.S. circumstances. On the other hand, the OECD guide-
regulations and the OECD guidelines was on trans- lines emphasize the use of transactional methods
actional methods such as the comparable uncon- before the use of profit-based methods, imposing an
trolled price, the cost-plus, and the resale price implicit hierarchy of methods that has sometimes
methods, which required transactional comparabil- been made explicit by the administrative positions
ity. After a number of iterations, the final revised of certain tax administrators, most notably the
IRC section 482 transfer pricing regulations2 and Canada Revenue Agency (CRA).
the draft OECD guidelines,3 both issued in June On a pragmatic level, the CUP is the best possible
1994, acknowledged the use of profit-based methods, transfer pricing method, as it requires the highest
which — rather than using transactional data — degree of transactional and product comparability.
allow the comparison of profits between related In that context, one could conceive of two types of
entities and independent third parties to establish, CUPs: an internal CUP and an external CUP. An
albeit in a roundabout manner, whether related internal CUP exists when a related party is involved
parties had transacted at arm’s length. in a transaction with an independent third party
involving the same or very similar products under
the same terms and conditions as the transaction
On a pragmatic level, the CUP is between itself and another related party. An exter-
the best possible transfer pricing nal CUP exists when two independent third parties
method, as it requires the highest are involved in a transaction that mirrors the trans-
degree of transactional and action between two related parties. In the absence of
product comparability. such CUPs, which are invariably rather difficult to
find, taxpayers — whether in the United States or
any of the other OECD countries — often are forced
In the end, the United States settled on the use of to use profit-based methods such as the CPM or
transactional methods such as the CUP, cost-plus, TNMM when testing routine margins, and the
and resale price methods, as well as profit-based profit-split methods when testing margins derived
methods such as the comparable profit method and from the use of shared intangibles.
the profit-split method. The 1995 OECD guidelines4
enshrined the transactional CUP, cost-plus, and
resale price methods, as well as the profit-based The Use of Profit-Based Transfer
transactional net margin method (TNMM) and Pricing Methods
profit-split methods. Among the profit-based meth- The use of profit-based transfer pricing methods
ods, the CPM in the United States and the TNMM is driven mainly by the lack of adequate transac-
elsewhere in the OECD generally are used to test tional data to apply any of the preferred transac-
the routine returns of a related entity, whereas the tional transfer pricing methods. Because profit-
profit-split methods are used to test nonroutine based methods can be applied on aggregate-level
contributions in which intangibles are present and data using third-party comparables, and that com-
are used by both parties to the transaction. parable company data can be acquired from publicly
However, the crucial difference between the ap- available company databases, the availability of
plication of those methods in the U.S. and OECD data generally is not a serious concern. Nonetheless,
in order to select the best possible comparables (that
is, those that have the highest degree of functional
comparability), it is imperative that a systematic
1
Organization for Economic Cooperation and Develop- comparable search process be undertaken. In many
ment, ‘‘Transfer Pricing and Multinational Enterprises — instances, it is the economic integrity of the compa-
Report of the OECD Committee on Fiscal Affairs,’’ 1979 rable screening process that determines the quality
(Paris: OECD). of the third-party comparables in the final sample.
2
U.S. Treas. reg. sections 1.482-1 through -8. Equally important is the choice of profit indicators
3
Organization for Economic Cooperation and Develop- (also called profit-level indicators in U.S. parlance)
ment, ‘‘Transfer Pricing Guidelines for Multinational Enter- used to derive the arm’s-length range of profits
prises and Tax Administrations,’’ (Discussion Draft) June against which the profits of the related party are
1994 (Paris: OECD).
4
compared.
Organization for Economic Cooperation and Development
‘‘Transfer Pricing Guidelines for Multinational Enterprises This section focuses mainly on the CPM and
and Tax Administrations,’’ 1995 (Paris: OECD), hereinafter TNMM, as they often are the methods of choice, in
referred to as the OECD guidelines. the United States and by the OECD, respectively, for

760 • November 21, 2005 Tax Notes International

 Special Reports

determining routine returns of entities engaged in word ‘‘transactional,’’ it is a profit-based method, the

(C) Tax Analysts 2005. All rights reserved. Tax Analysts does not claim copyright in any public domain or third party content.
related-party transactions. Routine returns gener- application of which is broadly (but not exactly)
ally refer to returns earned by a related party solely similar to the application of the CPM. As pointed out
by virtue of the functions it performs, the risks it by many eminent commentators, the degree of dif-
assumes, and the assets it employs, assuming that ference between the two methods often can be traced
those assets do not include any valuable intangibles. to certain nuances in their practical application,
In that context, it may be worthwhile to point out rather than to dramatic differences in the underly-
that both the CPM and the TNMM are one-sided ing theory of application, as some would have us
methods, meaning that they are applied to only one believe.
side of a transaction. Therefore, practitioners gener-
ally apply the CPM or TNMM to the least complex
entity (that is, the entity that does not employ any The residual profit-split often is the
valuable intangibles and whose results can be veri- method of choice when intangibles
fied using the most reliable data and the fewest
adjustments), which is also called the ‘‘tested party.’’
are involved in the course of a
related-party transaction.
When intangibles are shared between two related
entities engaged in a transaction, profit-split meth-
ods are more appropriate to determine the arm’s- Indeed, there are possibly more similarities be-
length returns attributable to the related entities. In tween the two methods than differences, and some
that event, the application of the profit-split meth- of those similarities are worth pointing out here. For
ods is done in two primary ways: first, by reference
example, both methods use aggregate-level data
to comparable profits earned by third parties that
from third-party comparable companies selected on
perform broadly similar functions (and incur associ-
ated risks) and employ similar types of intangibles the basis of their functional similarity to the tested
(the comparable profit-split method); and second, by party to derive the final arm’s-length range of profits
determining the routine returns attributable to the against which the results of the tested party are
related entities through the application of either the evaluated. To select the closest functionally similar
CPM or the TNMM and then splitting the residual third-party comparables, a variety of quantitative
profit between the related entities based on the and qualitative screens are employed in both in-
extent of their contribution to earning the residual stances. An appropriate profit indicator — usually a
nonroutine returns (the residual profit-split profitability ratio of some sort, using either income
method). or balance sheet account items — is selected under
The comparable profit-split method is consider- both methods. At the end of the comparables screen-
ably more difficult to apply than the residual profit- ing process, an arm’s-length range of the selected
split method, mainly because of the difficulty of profit indicator is derived from the ratios of the
finding comparable third parties, both of which comparable companies, and that range is then com-
engage in transactions similar to the related parties pared to the results of the tested party, after making
and employ intangibles similar to those of the re- the necessary comparability adjustments to level
lated parties. Therefore, the residual profit-split the playing field between the comparable companies
often is the method of choice when intangibles are and the tested party. Some of those comparability
involved in the course of a related-party transaction. adjustments relate to differences in working capital
Nonetheless, the ratio of the split of the residual (adjustments for accounts receivable, accounts pay-
profit between the related entities must be deter- able, and inventory) as well as differences in inven-
mined very carefully by taking into account the tory valuation methods (LIFO, or last-in, first-out, to
relative contributions of the parties to the develop- FIFO, or first-in, first-out), and a whole array of
ment of whatever intangible is shared between other differences.
them.
The differences between the two methods stem
However, in almost all cases in which intangibles mainly in how the application of the TNMM is
are not shared, the method of choice is the CPM or interpreted vis-à-vis the CPM, mainly because the
TNMM, which is applied to determine the routine OECD wants taxpayers to preserve the spirit of the
returns attributable to the least complex entity in a transactional methods even when applying the
related-party transaction. The CPM and TNMM are TNMM, although that is not always possible. It may
based on the economic concept that returns earned be argued, for example, that the TNMM emphasizes
by firms operating in the same or a similar industry, the identification of tested-party profits related
and under similar economic conditions, tend toward solely to the relevant related-party transactions,
equality over a reasonably long period of time. rather than the whole entity. However, that argu-
At this juncture, it may be worthwhile to reiterate ment can hold only if such detailed data is available,
that despite the fact that the TNMM contains the and the profits associated with the related- and

Tax Notes International November 21, 2005 • 761

Special Reports

unrelated-party transactions within the tested en- dard ratios in most, if not all, OECD countries. The

(C) Tax Analysts 2005. All rights reserved. Tax Analysts does not claim copyright in any public domain or third party content.
tity are clearly demarcated in the company’s finan- U.S. regulations outline three primary PLIs:6
cial statements. Often, taxpayers have to settle for • return on capital employed (ROCE);
testing the profits of the entire entity rather than
just the transactional profits associated with par- • return on sales (ROS); and
ticular related-party transactions. Other differences • return on operating costs (the Berry ratio).
relate to how comparables are selected, how many
are selected, and how the arm’s-length range is Other PLIs, such as return on total costs (ROTC)
derived. and other asset-based PLIs, also can be used to test
entities. It is important to keep in mind that the
In Canada, the CRA strictly opposes the use of profit margins are always determined relative to a
statistical methods to narrow the range, as is done base that can be either a balance sheet item, such as
in the United States with the interquartile range, assets, or an income statement item, such as costs or
opting instead for an approach that emphasizes sales. Generally, the choice of the base depends on
more functional screening that is also inevitably how an entity earns its returns. For example, for an
more subjective than standard statistical measures. entity that employs significant assets in its opera-
Advocates of that approach point out that the use of tions, measuring operating profit relative to assets
statistics does not inherently add any value when may be most prudent. For entities that do not
small sample sizes of somewhat functionally compa- employ significant assets in their operations, a base
rable companies are considered, especially when such as costs or sales may be more appropriate.
those companies have already been selected using
In that context, it may be useful to examine each
qualitative measures. At the other extreme, some
of these PLIs in turn to determine their usefulness
commentators would like to dispense with the use of
in specific situations.
qualitative screens altogether and rely on statistical
methods to narrow the ranges of a large number of Return on Assets (ROA) and ROCE
comparable companies. Notwithstanding the idio-
The ROA is the ratio of operating profit to oper-
syncrasies associated with their application in dif-
ating assets (OP/OA),7 while the ROCE is a special-
ferent jurisdictions, the critical factors that deter-
ized application of the ROA whereby the operating
mine the integrity of such profit-based methods is
assets are defined as capital employed, which usu-
undoubtedly the quality of comparables in the final
ally computes as the total assets minus cash and
sample, and the choice of the profit indicator.
investments.
Choosing an Appropriate Profit Level Of the various net profit margins that can be
Indicator examined, the ROA and ROCE (which is defined in
the U.S. Treasury regulations) are most firmly
Profit indicators or profit level indicators (PLIs) grounded in economic theory. That is because com-
are ratios that measure relationships between the petitive firms in a perfectly competitive marketplace
profits earned by a tested party and the costs in- grow or decline by gaining or losing capital invest-
curred or resources employed. Generally speaking, ment. Therefore, the return on assets for firms
the choice of a PLI should be determined by the type operating in a perfectly competitive market should
of activity performed by the tested party and the be equal.
economic circumstances of the related-party trans- However, the use of an ROA measure may not
action, as well as the reliability of the available data always yield reliable results in certain situations,
for the third-party comparables. The U.S. Treasury including:
regulations offer some guidance on the type of PLIs
that should be used in various circumstances when • when substantially fixed assets are not used to
applying the CPM, while the OECD guidelines are generate operating profit;
vague, except to indicate that net margins should be • when there are significant differences in the
used in the application of the TNMM. Specifically, age and condition of assets;
the OECD guidelines recommend that the ‘‘net
profit margin relative to an appropriate base (e.g.
costs, sales, assets) that a taxpayer realizes from a
controlled transaction’’ be examined.5 6
U.S. Treas. reg. section 1.482-5.
7
The PLIs outlined in the more explicit guidance in Operating profit is defined as profit from the relevant
the U.S. Treasury regulations have become the stan- operating activities before interest, tax, and extraordinary
items. Operating assets can be defined in a variety of ways,
including but not limited to: capital employed, which is
defined as total assets less excess cash and investments in
subsidiaries; gross assets; gross assets minus current liabili-
5
OECD guidelines, section B, para. 3.26. ties; assets minus liabilities; and so forth.

762 • November 21, 2005 Tax Notes International

 Special Reports

• when not all assets can be accounted for accu- operating expenses than gross margins, even if they

(C) Tax Analysts 2005. All rights reserved. Tax Analysts does not claim copyright in any public domain or third party content.
rately on the balance sheet; and are measured relative to sales or the cost of goods
• when asset usage is significantly different for sold.10
companies operating within the same industry. Return on Sales (Operating Margin)
Therefore, the ROA and ROCE generally are The ROS is the ratio of operating profit to sales
considered the preferred profit indicators for the (OP/sales).11 It measures profit after the cost of sales
application of the CPM/TNMM only when the tested and operating expenses, and is a well-understood
party has substantial fixed assets that can be accu- measure of profitability. It typically is used in the
rately accounted in terms of age, condition, and case of distributors and, occasionally, as a corrobo-
usage, and that play a significant role in generating ratory indicator for other types of entities.
operating profit. That means that the ROA and
ROCE generally are used to test manufacturing Return on Total Cost
entities that employ significant assets in their op- The ROTC is the ratio of operating profit to total
erations. cost (OP/TC).12 It is a net profit margin that uses
Although the U.S. Treasury regulations do not costs as its base instead of assets. The ROTC is
exclude intangibles from the ROCE calculation, as commonly used for contract and low-risk manufac-
part of its advance pricing agreement program, the turers and service providers, as it generally provides
U.S. IRS recommends that the ROCE be calculated the most reliable measure of profitability for those
as total assets minus intangible assets such as entities.
goodwill, minus investments in subsidiaries, minus Return on Operating Expenses (Berry Ratio)
excess cash and equivalents (except those used as
working capital).8 The reason given for excluding The Berry ratio is the ratio of gross profit to
intangibles from the ROCE calculation is the inher- operating expenses (GP/OE).13 It is used for service
ent difficulty in including the tested party and providers and for routine or ‘‘pure’’ distributors, and
comparable companies’ intangibles on a consistent may be thought of as a markup on operating ex-
basis.9 penses. Applying precisely the same logic, it also
may be used to test whether service providers have
earned enough of a markup on their operating
The Berry ratio is the ratio of expenses. In many cases, when it is possible to
gross profit to operating expenses. clearly demarcate costs unrelated to service provi-
sion (that is, when such costs are classified in the
costs of sales of a service provider’s income state-
If this formula for the ROCE is used, care should
ment), the Berry ratio may actually be a better
be taken to exclude any intangible-related amorti-
measure of performance than the ROTC, which
zation expense incurred by both the tested party and
includes all the costs incurred by a service provider.
comparables from the operating profit used to calcu-
late the ROCE ratios. In addition, other asset-based In essence, the Berry ratio implicitly assumes
ratios — such as return on invested capital, return that there is a relationship between the level of
on total assets, return on fixed assets, return on operating expenses and the level of gross profits
current assets, as well as a combination of those earned by routine distributors and service providers.
ratios — also can be used, as long as the economic Consequently, it is appropriate to use the Berry ratio
rationale underlying their use is clearly explained. if the selling or marketing entity is a routine dis-
When testing entities such as distributors or tributor and is entitled to a return on its operating
service providers, which do not employ significant expenses alone, or if it is a service provider entitled
assets, PLIs using income statement items as the to a return on its costs of provision of its services.
base might be more appropriate. Typically, the
income-statement-based PLIs used for distributors
and service providers are the ROS and ROTC, re- 10
OECD guidelines, para. 3.27: ‘‘The net margins also may
spectively. One of the principal advantages of net
be more tolerant to some functional differences between the
margins such as the ROS and ROTC is that they are controlled and uncontrolled transactions than gross profit
less affected by differences in functions or account- margins.’’
ing classifications between the cost of sales and 11
Operating profit is defined as above.
12
Operating profit is defined as above. Total cost is defined
as the cost of sales plus operating expenses.
13
8
Gross profit is defined as sales less the cost of sales.
APA Study Guide, p. 14, available as of Nov. 4, 2005, at Operating expenses are defined as expenses related to busi-
http://www.irs.gov/pub/irs-apa/apa_study_guide.pdf. ness operations other than the cost of sales, interest, taxes,
9
Id. and extraordinary items.

Tax Notes International November 21, 2005 • 763

Special Reports

The next section is devoted to an examination of acted as an independent third party. However, DISA

(C) Tax Analysts 2005. All rights reserved. Tax Analysts does not claim copyright in any public domain or third party content.
the Berry ratio, which is exceptionally useful when was not a ‘‘typical’’ distributor and performed a
applied correctly, but also is by far the most misun- combination of functions, including market re-
derstood of any PLI used in transfer pricing analy- search, marketing consulting, and advertising. In
ses. addition, DISA performed logistics and accounting
functions for some product shipments that were
The Berry Ratio — A Historical shipped directly by DuPont, while in other cases, it
Overview took possession of the inventory and undertook
full-fledged distribution functions. Therefore, Berry
The Berry ratio has its origins in a case from the characterized DISA as providing services analogous
late 1960s, when Charles Berry, then-professor of to those provided by a combination of a market
economics at Princeton University, was consulted by research and management consultant, an adver-
the U.S. IRS and Justice Department to evaluate tiser, and a distributor.17
the economic circumstances underlying a dispute
To evaluate the returns that might be earned by
between the IRS and the E.I. DuPont de Nemours
third-party comparable companies, Berry evaluated
Co. of Wilmington, Delaware (DuPont). The case
DISA’s performance separately with reference to
was filed in the U.S. Claims Court by DuPont to
third-party market research and management con-
recover assessments by the IRS during the tax years
sultants, third-party advertisers, and third-party
1959 and 1960 under the provisions of IRC section
distributors. In the analysis involving market re-
482.
search and management consultants, he compared
The case was important for two reasons: It in- DISA’s ratio of gross profit to operating expenses to
volved a challenge by the IRS concerning the mar- the ratio of total income earned to total costs of
gins earned by a wholly owned related-party dis- service provision (in essence, the markup on total
tributor of DuPont, and it was the first major case to costs). In the analysis involving advertisers, DISA’s
follow on the heels of the newly released U.S. Trea- ratio of gross profit to operating expenses was com-
sury regulations in 1968.14 pared to the ratio of billed commissions that can be
The facts of the case were relatively simple. In considered analogous to the gross profit (excluding
essence, in 1958 DuPont established a new, wholly the costs of advertising placement, as such services
owned subsidiary in Switzerland, DuPont de Nem- were performed by the media rather than the adver-
ours International S.A. (DISA), which acted as a tisers themselves) to total operating costs of the
‘‘super distributor’’ in Europe on behalf of its U.S. advertising agencies. In the third case (distributors),
parent, DuPont.15 The functions performed by DISA DISA’s ratio of gross profit to operating expenses
included marketing and advertising in Europe to was compared to the third-party comparables’ ratios
establish a European presence for DuPont, as well of gross profits less interest and extraneous income
as distribution functions, which included purchas- to operating costs (excluding interest costs and de-
ing products for resale to other affiliates in Europe. preciation).18 That ratio of gross profit to operating
All products sold in Europe were routed through expenses was later named the Berry ratio in honor
DISA, which acted as the intermediary between its of Berry, who devised that approach of evaluating
parent company, DuPont, and other related Euro- distributors.
pean distributors. DISA was afforded a margin of Berry’s analysis revealed that, regardless of
approximately 20 percent on the selling price of the whether one considered the markup on costs real-
products it purchased from DuPont. That margin ized by third-party management consulting and
was considered too high by the IRS, which issued a market research firms, advertising agencies, or in-
deficiency notice in that regard.16 dependent distributors, DISA’s Berry ratio indicated
Berry was asked to determine whether DuPont that DuPont had compensated DISA for its services
transacted at arm’s length with its related-party at a level that was significantly above what could be
distributor, DISA. More specifically, he was asked to considered to be arm’s-length terms.19
determine whether the resale margin (the discount
from the final European selling price) afforded to
DISA by DuPont could be considered to be commen- 17
surate with the services performed by DISA if it had Id., p. 17.
18
Contrary to the U.S. Treasury regulations, which define
operating expenses as including ‘‘a reasonable allowance for
amortization and depreciation’’ (Treas. reg. section 1.482-
14
5(d)), Berry’s preference is to exclude depreciation from
E.I. DuPont de Nemours & Co. v. United States, 608 F.2d operating expenses altogether because of the inherent arbi-
445 (Ct. Cl. 1979). trariness in determining what might constitute this reason-
15
Id., p. 16. able allowance.
16 19
Id., pp. 16-17. Id., p. 18.

764 • November 21, 2005 Tax Notes International

 Special Reports

Application of the Berry Ratio in operating expenses, and is analogous to the ROTC

(C) Tax Analysts 2005. All rights reserved. Tax Analysts does not claim copyright in any public domain or third party content.
Transfer Pricing Analyses used for service providers. That is best illustrated by
deconstructing the ROTC as follows:
As Berry points out in his article,20 the key
insight to be drawn from the DuPont case is that the ROTC = OP/TC
Berry ratio is merely a variant of the cost-plus ROTC = OP/(COGS+OE)
method. Indeed, if one were to think of the gross wherein COGS = cost of goods sold.
margins earned by a distributor as analogous to a In the above ratio, the COGS generally can be
firm’s total revenues available to a distributor, and excluded from the cost base in the denominator,
the operating expenses incurred to distribute prod- because for distributors COGS indicates the mea-
ucts as analogous to the firm’s total costs, then the sure of the value of the product distributed, rather
ratio of gross margin to operating expenses would than the costs incurred for distribution. In the case
capture the markup on operating expenses that is of service providers, unlike manufacturers, the
afforded to the distributor. Conceptually, the Berry COGS may not even be applicable, or may relate to
ratio represents a return on a company’s value costs other than service provision. In that event, the
added functions and assumes that those functions ratio becomes analogous to the ratio of operating
are captured in its operating expenses.21 In other profit to operating expenses, or the markup on
words, the Berry ratio can be a useful measure of the operating expenses alone. Therefore, for service pro-
markup earned on a distributor’s distribution activi- viders, assuming that the COGS = 0, ROTC =
ties. In that context, it may be useful to further OP/OE.
deconstruct the Berry ratio to understand the impli-
cations of its use in specific situations.
The key insight to be drawn from
When evaluating distributors, given that the ra-
tio has gross profit in the numerator and operating
the DuPont case is that the Berry
expenses in the denominator, a profitable distribu- ratio is merely a variant of the
tor would invariably show a Berry ratio greater than cost-plus method.
one, if shown in units, or 100 percent, if shown in
percentage. If that is not the case and the Berry Therefore, as illustrated in the above formulae,
ratio is less than one unit or 100 percent, as the case the Berry ratio can be applied to both distributors
may be, there may be some evidence of excessive and service providers, as long as the cost categories
operating expenses that need to be curtailed in the are demarcated and classified appropriately. In fact,
long run. In essence, any distributor or service transfer pricing economists with the U.S. IRS and
provider with a Berry ratio of less than one unit or other OECD member tax administrations, such as
100 percent cannot sustain its operations indefi- the CRA, have routinely begun applying the Berry
nitely. ratio to test the margins of distributors as well as
As mentioned previously, the Berry ratio also can service providers, and to conclude APAs with tax-
be applied to service providers, as it can also be payers.22
conceptualized as the markup earned on the costs of
provision. To better understand that relationship, it Revisiting the Misuses of the
may be prudent to reduce the Berry ratio in terms of Berry Ratio
operating profit by subtracting one from the Berry
ratio expressed in unit terms as follows: Although the Berry ratio is a conceptually simple
profitability measure, it is probably one of the most
Berry ratio - 1 = GP/OE - 1 misused ratios in the context of transfer pricing
= (GP-OE)/OE analyses. Its misuse stems primarily from the fail-
= OP/OE ure to understand its limitations when evaluating
wherein GP = gross profit; OP = operating different types of entities. On a fundamental level,
profit; and OE = operating expenses.
The above result — the ratio of operating profit to
operating expense — is merely an alternative way to
conceptualize the Berry ratio as the markup on 22
An APA is an arrangement between a taxpayer and a tax
administration that confirms the appropriate transfer pricing
method to establish an arm’s-length price for transactions
between related parties. The U.S. IRS has been concluding
20
APAs using the Berry ratio for some time now, and the CRA
Charles H. Berry, ‘‘Berry Ratios: Their Use and Misuse,’’ recently completed one APA using the Berry ratio, with two
Journal of Global Transfer Pricing, April-May 1999, re- more in progress. See the CRA’s 2003-2004 APA Report,
printed by CCH Inc. released by the Competent Authority Services Division, In-
21
APA Study Guide, p. 16, available as of Nov. 4, 2005, at ternational Tax Directorate, Compliance Programs Branch,
http://www.irs.gov/pub/irs-apa/apa_study_guide.pdf. for additional details.

Tax Notes International November 21, 2005 • 765

Special Reports

the Berry ratio relies on the fact that there is some tors that incur similar selling, general, and admin-

(C) Tax Analysts 2005. All rights reserved. Tax Analysts does not claim copyright in any public domain or third party content.
consistency between the level of gross margins and istrative expense-to-sales ratios as the tested party
operating expenses (that is, the greater the operat- are used to develop the arm’s-length range of Berry
ing expenses, the greater the gross profit needs to be ratios (operating expense intensity screen). That
to sustain a similar level of operating profit). None- would effectively ensure that any distortions to the
theless, that consistency can be expected only if the Berry ratio analysis caused by radical differences in
operating expenses capture all of the value added of operating expense intensities among the compa-
the functions performed by the distributor. In other rable distributors and between the comparable dis-
words, the Berry ratio simply captures the markup tributors and the tested party are minimized. None-
that should be earned on operating expenses, as- theless, it is also necessary to ensure, to the extent
suming that those expenses reflect all of the value possible, that functional and product comparability
added by a distributor. is not sacrificed in favor of implementing an operat-
ing expense intensity screen, because ignoring the
For that precise reason, the Berry ratio cannot be
degree of functional or product comparability be-
applied to integrated distributors (that is, distribu-
tween the tested party and third-party comparables
tors that also perform manufacturing functions), as
also could derail the spirit of the analysis.
the Berry ratio would not be able to capture the
additional return earned by the manufacturing
functions. In addition, when integrated distributors Application of the Berry Ratio —
are concerned, given the accounting conventions and Some Practical Insights
flexibility of classifying costs between the costs of
goods sold and operating expenses, the cost base In the authors’ experience, the Berry ratio is
used in the Berry ratio (the operating expenses) also rarely, if ever, applied in isolation to test routine
may contain costs related to manufacturing, which distributors, especially in the OECD countries. In
is certainly a perversion of the original intent of the many cases in which the United States and another
Berry ratio. OECD jurisdiction are involved, taxpayers often use
the ‘‘modified’’ resale price or cost-plus methods to
Therefore, when applying the Berry ratio to test distributor or service provider margins, respec-
evaluate distributors, care should be taken to apply tively, and to corroborate their analysis using the
the ratio only to distributors that perform only CPM or TNMM, as the case warrants. For example,
routine distribution functions, and do not engage in when transactional data is unavailable for analysis,
any additional value-added assembly or manufac- the OECD guidelines provide the option of using
turing functions. Needless to say, that limitation ‘‘modified’’ methods, which use external, potentially
must be kept in mind when considering both the comparable companies. The analysis is performed
tested party and the third-party comparables to on their aggregate-level data in a manner that is
ensure that functionally similar entities are being quite similar to the application of the CPM or
compared. In essence, the Berry ratio is best applied TNMM.
to test routine distributors, and only when there is a
high degree of functional comparability between the Specifically, paragraph 3.2 of the OECD guide-
tested party and the third-party comparable compa- lines states: ‘‘The only profit methods that satisfy
nies. the arm’s length principle are those that are consis-
tent with the [profit-split method] or the transac-
In that context, it may be worthwhile to point out tional net margin method as described in these
that the application of the Berry ratio is not without Guidelines. . . . In particular, the so-called ‘compa-
its problems. For example, many empirical studies rable profits methods’ or ‘modified cost plus/resale
have shown that distributors with exceptionally low price methods’ are acceptable only to the extent that
operating expense intensity (that is, operating ex- they are consistent with these Guidelines.’’
penses relative to sales ratios that are less than 10
percent to 15 percent) show inordinately high Berry Therefore, as an example, a tested-party distribu-
ratios when compared with distributors with higher tor’s gross margins first may be evaluated using the
operating expense intensities.23 Therefore, consider- ‘‘modified’’ resale price method by deriving the
able caution should be exercised when comparing arm’s-length range of gross margins of a set of
the Berry ratios of distributors with low operating comparable distributors. Subsequently, the TNMM
expense intensities and distributors with higher or CPM is applied to compare the Berry ratio and
operating expense intensities. That problem can be ROS ratios of those comparable distributors with
corrected, however, by ensuring that only distribu- the tested party’s Berry ratio and ROS results to
corroborate the earlier gross margin analysis. Given
the lack of explicit guidance in most OECD countries
as to which profit indicators to use in specific cases,
23
APA Study Guide, p. 16, available as of Nov. 4, 2005, at performing such corroboratory analyses using mul-
http://www.irs.gov/pub/irs-apa/apa_study_guide.pdf. tiple PLIs may perhaps be the best approach to

766 • November 21, 2005 Tax Notes International

 Special Reports

ensure that taxpayers’ analyses withstand the scru- in non-U.S. jurisdictions, as profit-based methods,

(C) Tax Analysts 2005. All rights reserved. Tax Analysts does not claim copyright in any public domain or third party content.
tiny of the various tax administrations. including the TNMM, are looked at with some
disdain by many tax administrations within the
Conclusion OECD.
In sum, PLIs applied in the context of transfer
Although tax administrations may wish other-
pricing analyses using profit-based methods in the
wise, taxpayers do have some leeway in their choice
United States or the OECD in general should be
of methods and profit indicators to prove the arm’s-
chosen with particular reference to the economic
length nature of their related-party transactions,
rationale underlying their application, and with a
although that leeway is limited, of course, by the
clear understanding of the specific facts and circum-
constraints of data availability and the reliability of
stances of the related-party transaction being exam-
the data that is available. Nonetheless, such leeway
ined. That is especially true when applying PLIs
should not be interpreted as carte blanche to apply
such as the Berry ratio that require an examination
transfer pricing methods, and especially PLIs, in
of not only the type of functions performed by a
situations where they may not be economically jus-
distributor or service provider, but also the level of
tified or applicable. When in doubt, taxpayers would
intensity at which those functions are performed.
be well advised to consider the economic fundamen-
When applying such a PLI, it is also quite important
tals that may have motivated their choice of PLIs.
to understand its limitations (which in the case of
Ultimately, transfer pricing is more art than science,
the Berry ratio is the fact that it categorically cannot
more judgment than precision, and the final objec-
be applied to distributors who perform value-added
tive of any transfer pricing analysis is to prove the
manufacturing or assembly functions).
arm’s-length principle.
Furthermore, it is imperative that taxpayers and
practitioners carefully evaluate the type of entities To quote Berry himself, ‘‘If we are to be consistent,
chosen as third-party comparables to determine the ultimate test is the arm’s length test, and not the
whether the PLI being used may be distorted by existence of a necessarily incomplete example or
issues such as operating expense intensity, asset some arbitrary rule that gives a mistaken aura of
intensity, or account classification issues. In that precision to what is inherently an inexact and highly
context, it is also advisable to keep in mind that judgmental process.’’24 ◆
whenever possible, taxpayers should use more than
one PLI to corroborate their transfer pricing analy-
ses, especially if doing so will strengthen the results
of the primary analyses. That may be especially 24
Charles H. Berry, ‘‘Berry Ratios: Their Use and Misuse,’’
prudent when performing transfer pricing analyses supra note 20, p. 23.

Tax Notes International November 21, 2005 • 767