Problems in international division of the business

income tax base

Richard J Vann*

Dear Colleagues,

The attached draft is in its final stage so far as the text is concerned but there is
considerable more referencing to be done. The draft in part will be a response to OECD
documents due to be released before the end of the year and so there is more referencing
in particular to be done for OECD materials.

The argument in a broad sense is twofold. First, neither the country of the shareholder nor
the country of headquarters of a multinational corporate group collect much tax on the
international business income of the group for a variety of structural and policy reasons
(Section II of the article). The collection of corporate tax occurs in the countries to which
the profits of the group are allocated under transfer pricing rules. Secondly, current
transfer pricing rules have a number of structural flaws which permit the movement of
profits to tax havens (Section III of the article). Section IV brings the previous sections
together in proposing solutions. I suggest that you read Sections IIC and III.

Regards,

Richard

*
Challis Professor of Law, University of Sydney, William K Jacobs Jr Visiting Professor of Law, Harvard
Law School. My thanks for the generous support of the Harvard Law School Fund for Tax and Fiscal
Policy Research and [other acknowledgments].

1
 CONTENTS

Problems in international division of the business income tax base................................... 1

Richard J Vann............................................................................................................ 1
I Introduction .................................................................................................................. 3
II Context of division of the international business tax base .......................................... 4
A Corporate shareholder taxation ............................................................................... 4
B Residence and source taxation ................................................................................ 6
C Transfer pricing ....................................................................................................... 9
III Operation of the transfer pricing rules ..................................................................... 17
A The firm................................................................................................................. 18
B Market price benchmarks: transactions, dealings and legal entities ..................... 20
C Functional analysis and risk .................................................................................. 25
D Personnel............................................................................................................... 30
E Assets..................................................................................................................... 32
F Sales....................................................................................................................... 37
IV Possible solutions..................................................................................................... 40
A Fixing the transfer pricing rules ............................................................................ 40
B Are there other solutions for the growth of transfer pricing tax avoidance?......... 43
Appendix: Integration of corporate and shareholder taxation ...................................... 48
Part I Relative tax at corporate and shareholder level under integration.................. 48
Part II Integration where income derived by permanent establishment of corporation
................................................................................................................................... 49

2
I Introduction
In the US alarm bells are regularly sounded concerning the loss of corporate tax revenue
from international tax avoidance. A common culprit blamed for the loss is transfer
pricing.1 The OECD recently decided to investigate if there is a fire.2 This article seeks to
assess why current transfer pricing rules are the source of tax avoidance and to explore
some possible remedies.

To do this the article first places transfer pricing rules in the overall context of the
international division of the income tax base in the common situation where a widely
held corporate group operating in several countries derives business income. The purpose
is to show why transfer pricing is the dominant international issue as compared to
corporation and shareholder or residence and source taxation. The article then identifies
that it is structural issues in transfer pricing rules which are the main drivers of tax
planning, not difficulties of finding appropriate prices. Identifying the causes of tax
planning in turn suggests possible solutions. The theory, or rather theories, of the firm
stemming from the work of Coase provide the backdrop to the analysis.

There are three main problems in current transfer pricing law. First, the current rules for
defining whether a corporation is present in a country or not for tax purposes are too
narrow. Secondly, the transfer pricing rules permit corporations to structure intra-firm
contracts as they wish on an inappropriate market analogy. Third, the rules give too much
emphasis to risk and not enough to personnel and assets in dividing up international tax
revenues. These problems can be rectified directly but there also are other mechanisms in
tax and corporate law that assist in dealing with transfer pricing.

The rules with which the article deals are generally a mixture of national tax rules
applicable to international situations and bilateral tax treaty provisions. Over the 80 years
1
Two recent examples are Desai, Foley and Hines, “The demand for tax haven operations” 90 Journal of
Public Economics 513 (2006) and Altshuler and Gruber, “Governments and multinational corporations in
the race to the bottom” 110 Tax Notes 979 (2006). Specifically the former focuses on transfer pricing and
deferral while the latter targets the use of hybrids in international taxation but in the broader sense they are
all about transfer pricing related tax avoidance as appears hereafter. There is a long history of work on the
possible revenue impacts of transfer pricing.
2
The OECD has been revising its transfer pricing rules since 1992. Just as the exercise was approaching its
conclusion, the OECD became concerned with business restructures under which, in one view, relatively
minor organizational changes in the business lead to large shifts of profits from high tax to low tax
jurisdictions. The original announcement at the 2nd Annual Centre for Tax Policy and Administration
Roundtable: Business Restructuring held on January 26-27, 2005 was fairly mild and limited, “There was
general agreement that it would be useful to do further work in clarifying and perhaps re-examining the role
of the dependent agent definition.” See summary of proceedings from which this quote is taken at
http://www.oecd.org/document/20/0,2340,en_2649_201185_34535252_1_1_1_1,00.html. By 2006, there
was greater urgency, “[T]here has been massive restructuring, typically involving risk stripping and
intangibles stripping, over the past decade. Governments … need a theoretical framework to differentiate
legitimate restructuring from tax rearrangement, and they need to know how to treat the legitimate
version.” Comments of Caroline Silberztein, Head of Transfer Pricing Unit, Centre for Tax Policy and
Administration, OECD as reported by Sheppard, “OECD makes nice to Americans, Part 2” 111 Tax Notes
1200 (2006).

3
during which there has been coordination of income taxation in multilateral institutions, a
consensus has developed around the main rules of the international income tax system,
including transfer pricing. There is considerable room for national legislation both as
regards structural policy choices and elaboration of the agreed international norms, but
there has been increasing convergence of the national as well as the treaty rules. No
specific country will form the focus; rather the discussion will be relevant to many
countries.3

II Context of division of the international business tax base

There are three main sets of rules which allocate taxing rights among countries over
international business income derived through corporations: rules about corporate-
shareholder taxation, rules about residence and source taxation and transfer pricing rules.4
The first two sets of rules effectively cede primary taxing rights to the third set of rules
and the discussion here indicates how and why. As well as explaining the importance of
transfer pricing, this context will be relevant to the later discussion of solutions to current
problems.5

A Corporate shareholder taxation

Most corporate shareholder tax systems nowadays attempt some degree of integration of
the taxes at the corporate and shareholder levels, including the US since 2003.6 The
problem of integration would not arise if the only tax on income derived through
corporations were at the individual shareholder level. It is generally accepted that conduit
(look-through) taxation of income of widely held corporations is not practically possible
and that generally income of necessity is taxed on a realization basis. As a result a levy of
income tax at the corporate level is necessary, otherwise there would be potentially

3
Many of the general statements about various countries’ tax systems are sourced from Ault & Arnold,
Comparative Income Taxation: A Structural Analysis (The Hague, Kluwer, 2004) 2nd ed especially Parts 3
and 4, and Vann, “Trends in company/shareholder taxation: single or double taxation” General Report in
International Fiscal Association cahiers de droit fiscal international vol LXXXVIIIa 21-70 (2003)
(hereafter “Trends”).
4
Nowadays usually a fourth set of tax rules is involved arising from the fact that a great deal of individual
investment in listed corporations is intermediated by one or more collective investment vehicles (CIVs –
mutual funds, pension funds, etc). As CIVs are not generally taxed in their own right they are ignored in
this discussion, though they add considerable complexity to international taxation of corporate income in
practice.
5
While much is written about each set of rules from both theoretical and practical perspectives, much less
is written about their interrelationship and what there is tends to be from a particular perspective, for
example, the literature on corporate-shareholder taxation especially in recent decades has been much
concerned with the interaction of this set of rules with the residence-source rules of international taxation,
but generally residence and source taxation is taken as given and the question is how effectively to fit the
two sets of rules together. Hence in the analysis here, there will only be passing references to these
literatures.
6
Internal Revenue Code §1(h)(11) (capping the maximum tax rate on dividends at 15%). The recent
research and some account of the recent events in the US can be found respectively in Graetz and Warren,
Integration of the US corporate and individual income taxes: The Treasury Department and American Law
Institute reports (Arlington, VA, Tax Analysts, 1998), Vann, Trends, note 3.

4
infinite deferral of income tax by retention in the corporation of the income realized at
the corporate level.7

Most international business income is derived through widely held multinational

corporations, that is, substantial corporations that operate in many countries (the
traditional understanding of the term), generally have a headquarters in one particular
country and increasingly are owned by shareholders from many countries. If shareholder,
corporation and income may be located in different countries and there is to be a
corporate tax on income derived through the corporation, the major question is which
country or countries should levy the corporate level tax. It is the major question because
little income tax is collected from ultimate shareholders under the usual kinds of
corporate tax integration – the main tax on the income is at the corporate level because of
the need to deal with deferral. A more detailed explanation of this outcome can be found
in Part I of the Appendix.

The deferral justification for the corporate tax only requires levy of the corporate tax on
any retention of profits. Moreover it suggests that corporate tax should be levied by a
country with respect to the proportion of the corporation’s retained income attributable to
shareholders in the country. Levy of corporate tax only on retentions and in the country
of the shareholder are not the international norms (assuming that corporation and income
are located elsewhere). Taking the point about not taxing in the country of the
shareholder first, there are obvious practical reasons for this result. It is difficult to
enforce the tax against the corporation if it has no links with the jurisdiction and to detect
the proportion of shareholders in the country, especially if the concern is to find the
ultimate individual owner of the shares and there are interposed entities. Relatedly and
more fundamentally such a tax would contradict the underlying assumption of the system
already noted that it is not feasible to tax corporate income on a look through basis.
Although the tax would be on the corporation, the amount of tax levied in the country
could only be determined by looking through the corporation and in effect attributing
retained income to shareholders.

Similarly corporate taxes focused on retentions (systems which give tax deductions for
the payment of dividends or apply reduced rates of corporate tax on distributed income)
are rare historically and apparently extinct. A number of reasons explain the levy of
corporate tax on all corporate income, not just retained income. Most obviously it is a
convenient administrative device to ensure effective collection of tax on distributions as
well as retentions. Further, if the shareholder is in one country, and the corporation and

7
Realization at the shareholder level is generally taken to mean taxation of dividends and of realized gains
on corporate shares. The income realized by the corporation is, if retained in the corporation, only realized
when the shareholder disposes of the shares. If for any reason there is no effective capital gains tax at this
point, for example, shares held until death in the US under Internal Revenue Code (“IRC”) §1014, deferral
becomes exemption. It will be noted that this is a different concept of realization to that which applies at
the corporate level so that it is possible to have income realized at the corporate level but not at the
shareholder level. One of the appeals of a system under which income realized at the corporate level is
attributed and taxed to shareholders whether or not distributed is that it aligns the concepts of realization at
the corporate and shareholder levels. As noted in the text, for widely held corporations such a system is not
regarded as feasible.

5
its income in another country, a dividend deduction or similar system would mean that
the tax take of each country and its timing would be affected by the amount of profits
distributed. The corporation could shift tax between the countries at will. Germany in the
past used a split rate of corporate tax with a significantly lower tax rate on distributed
earnings and simply ended up giving a lower corporate tax rate to US owned corporations
compared to German owned corporations.8 Apart from the potential for manipulation, it
is not evident what policy would be served by making revenue division between the
countries depend on the level of corporate distributions.9

In summary of this section, corporate tax is necessary because of the problem of deferral
of tax on retained earnings at the shareholder level under a realization based income tax.
In the international context the corporate tax is necessarily applied to all corporate
income and not just retentions if the corporate tax is levied by a country other than that of
the shareholder. Otherwise the distribution of tax revenue between countries would be at
the whim of the corporation and the result would not serve any other purpose. In practical
terms, this tax cannot be levied in the country of the shareholder10 if the corporation and
income are located elsewhere.

B Residence and source taxation

To this point the article has referred to the shareholder, corporation or income being in a
particular country. The terms used in international taxation for such locations are
residence and source but the focus above was corporate shareholder taxation and so far as
possible residence and source issues were suppressed.

In a general sense a taxpayer’s residence is the country with which the taxpayer has the
closest connection and income’s source is the country with which the income has the
closest connection. The current international norm is that the residence country taxes the
worldwide income of the taxpayer and the source country taxes only the income of the
taxpayer sourced in that country for which the residence country provides relief to
8
[History of German system]
9
The result could be offset by the levy of a dividend withholding tax on the shareholder in the country of
the corporation equal to the corporate tax rate but from that country’s perspective such a complication
hardly seems worthwhile compared to taxing the corporation on all its income. Such a proposal was made
in the US in 1984 but did not proceed, [Treasury I proposal]. It is possible to separate the methods of
taxation of corporation and shareholder from the division of the resulting revenue between countries by
using treasury to treasury transfers of tax revenue. While such approaches have been used in the past, they
have not been particularly successful and have not survived. They are not sufficiently practical or
politically acceptable to be discussed further.
10
It has long been recognized that it is possible to levy an accruals income tax on portfolio shareholders in
listed corporations by taxing dividends received plus or minus the annual change in value of the shares. The
difficulty is that an accruals system is not practical for other assets such as closely held businesses and real
estate because of valuation and other issues with the result that there would be discrimination in the taxing
method among different classes of assets. Nonetheless several countries do levy accruals taxes on
shareholders in foreign corporations in cases where it is likely that the income of the corporation is
essentially passive fixed income, the income is retained in the corporation, and no corporate tax is paid by
the corporation, for example, a corporation based in a tax haven investing in fixed interest securities, see
Ault and Arnold, note ? at 386-389. Because this is a departure from the normal realization rule, countries
are careful to apply the tax to very limited cases which are not intended to include investment in
multinational corporations operating active businesses.

6
prevent international double taxation. The justification for worldwide residence taxation
is that the residence country being the country to which the taxpayer is most closely
connected is the one that properly takes account of the personal circumstances of the
taxpayer and provides redistribution to or away from the taxpayer under its tax and
transfer system. Hence personal tax benefits of various kinds are confined to residents
and the progressive income tax rate schedule is applied to residents. To make sensible
decisions about distribution it is necessary to take the full circumstances of the taxpayer
into account which includes their full (worldwide) income.

There is much less agreement or indeed discussion on what the justifications for source
taxation are and therefore what source means. Source taxation is often just taken as a
given in which case discussion revolves around solution to the double taxation problem
arising when two countries tax the same income (the residence and source countries).
One view would explain source taxation on the basis of economic rents, though exactly
what rent means in this context is usually not elaborated. The oil well provides the
standard example – few query the taxing rights of the country where the well is situated.
Another view explains source on the basis of benefit taxation – the amount a taxpayer
pays for the benefits received in a jurisdiction.11 It is widely accepted that income from
use of a taxpayer’s asset in a jurisdiction is sourced there: real estate rentals, mineral
royalties, chattel rentals, royalties on intangibles, interest on money lent. For sales and
personal services income there is less agreement on the source rules, a matter returned to
below.

Because source taxation does not reflect the full circumstances of the taxpayer, it is
generally accepted that issues of progressive taxation (or not) are irrelevant and that tax
should be levied at flat rates. For similar reasons, the tax levels should be lower than the
highest individual tax rates of progressive tax systems in residence countries as the higher
rates are there for redistributional purposes. Mainly for administrative reasons, taxes on
passive income at source are levied by final withholding on a gross basis, except
generally for income from real estate.

The theoretical literature on residence and source taxation addresses two main issues –
whether residence only taxation is the appropriate policy and what is the appropriate
method of relief in the country of residence for source taxation. The former literature
essentially assumes that corporations do not exist while the latter literature assumes that
the owners of the corporation are resident in the same country as the corporation. The
argument for residence only taxation is that source taxation is shifted to immobile factors
and/or residents of the source country and so does not achieve its purpose of taxing the
non-resident.12 While it is easy to show that such shifting occurs in particular situations
especially involving gross basis source taxation, it is by no means clear that the same
11
The literature is conveniently collected in Arnold, “Threshold requirements for taxing profits under tax
treaties” in Arnold, Sasseville and Zolt eds, The Taxation of Business Profits Under Tax Treaties (Toronto,
Canadian Tax Foundation, 2003) at ? note 4, and see Pinto, “The Need to Reconceptualize the Permanent
Establishment Threshold” 60 Bulletin for International Taxation 206 (2006).
12
The literature is briefly summarized in Kaplow, “Taxation” section 5.5 to be published in Polinsky and
Shavell eds, Handbook of Law and Economics (North Holland, Elsevier due 2007) available at
http://www.law.harvard.edu/programs/olin_center/papers/542_Kaplow.php.

7
result applies to corporations under the current arrangements.13 If the result does apply,
then levying the corporate tax in any country other than the residence of the shareholder
will not achieve its purpose of preventing deferral at the shareholder level as the
incidence of the tax will not be on the shareholder. As already discussed levying the
corporate tax in the country of the shareholder is not feasible if the corporation and its
income are located elsewhere. It is assumed in what follows that the incidence of the
corporate tax is generally on the shareholder. The current international system that has
survived for over 80 years on a coordinated international basis and longer on an
uncoordinated basis is based on the corporate tax being levied at a place other than the
shareholder’s country and operates on an assumption of no shifting of the tax.14

The arguments over the method of relief of double taxation revolve around whether the
preferred policy is capital export neutrality, capital import neutrality and national
neutrality/welfare.15 The double tax relief systems associated with these concepts are the
foreign tax credit, exemption of foreign income from tax and deduction of the foreign tax
respectively. We can avoid discussion of this issue by noting that it elides the residence
of the corporation and its shareholders and so is inapplicable in the context of the
increasingly common case where a corporation’s shareholders are not necessarily resident
at the location of the corporation. Moreover, even if the corporation and shareholder are
in the same country, the ultimate result at the shareholder level is what is important and

13
The problem is most evident for interest on loans from financial institutions. If the country of the
borrower levies a flat rate gross withholding tax on the interest even at a low rate, the tax will exceed the
bank’s gross profit on the loan, that is, its spread between its borrowing and lending interest rates.
International lending agreement therefore typically contain a clause that grosses up the interest rate by any
interest withholding tax at source and shifts the tax to the borrower. For corporate tax on business income,
it may be that the tax holidays often provided by developing countries (and the issue of tax competition
between countries more generally) may be a signal of shifting, that is, countries give up or lower their tax
rates when they think shifting of the incidence is occurring. Similarly widespread tax planning by non-
residents to avoid source tax may be a signal that shifting is not occurring. Conversely if such tax planning
is by residents of the source country dealing with the non-resident as commonly occurs in international
lending transactions, this is an indicator that shifting has occurred. On this basis we might conclude that
shifting is a problem for many developing countries who therefore cut source tax rates significantly or
create special tax incentives, for certain types of income taxed on a gross basis and perhaps for corporate
tax rates that are high by international standards. Conversely, the studies about loss of revenue referred to
in note 1 may indicate that shifting does not generally occur in large developed countries. In the end result,
however, we do not know to what degree shifting occurs. Even if tax is shifted, the nominal taxpayer will
often be benefited if the tax can be avoided, which explains why some corporations seek to avoid sales
taxes.
14
The problem of shifting of source tax was a concern from the very beginnings of the coordination of the
international system and figured prominently in the report of the four economists which effectively
inaugurated international tax cooperation, see Economic and Financial Commission (Bruins, Einaudi,
Seligman and Stamp), Report on Double Taxation submitted to the Financial Committee Economic and
Financial Commission (League of Nations Document EFS73F, 19 April 5th 1923) at 7-16 (available
electronically at http://setis.library.usyd.edu.au/oztexts/parsons.html item 3 1st document). More recently,
Graetz, “Taxing International Income: Inadequate Principles, Outdated Concepts and Unsatisfactory
Policies” 54 Tax Law Review 261 (2001) has taken issue with the residence only argument.
15
The original discussion is usually attributed to Musgrave, United States Taxation of Foreign Investment
Income: Issues and Arguments (Cambridge, Harvard Law School, 1969); following Musgrave’s lead US
authors usually come down in favor of capital export neutrality, compare Desai and Hines, “Evaluating
International Tax Reform” Harvard NOM Working Paper No. 03-48 (2003) available at
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=425943 [get more recent version].

8
will often vary from that at the corporate level. The result at the residence of the
shareholder, wherever it may be, has as much to do with the method of corporate
shareholder taxation as the international double tax relief in the country of the
corporation. Appendix Part II provides more details of this result. To put it another way,
the traditional residence source analysis only makes sense in the case of individual
taxpayers.

In summary this section argues that residence and source taxation is under theorized in
the sense either that corporations have not been fitted into the framework which
essentially concerns individuals, or that it is assumed the corporation and shareholder are
in the same jurisdiction while the method and extent of corporate-shareholder integration
are ignored. The framework does not provide any significant information about which
country should tax income at the corporate level, if it is accepted that the country of the
shareholder cannot perform this task.

C Transfer pricing
The corporation is relatively invisible in discussions of corporate shareholder taxation
and residence and source. It is effectively assumed to be a pure conduit for income and
the focus is on the taxation of the ultimate shareholder, with taxation on the corporation
cast in a secondary supporting role. Yet the major part of tax revenue from international
business income is collected at the corporate level with little regard to the country of
residence of the shareholder. The purpose of this part is to explain why that is so and
exactly how it is achieved.

Theory of the firm

The suggestion here is that the corporation is seen in international taxation as the origin
of value and as more than a mere conduit of income. The international tax treatment of
corporations is linked, though not coherently, to the views of the firm broadly associated
with Coase. In these views firms form to make greater profits by directing the allocation
of productive resources instead of leaving resource allocation decisions to the market.
The explanations in this Coasean tradition of why the firm is able to make profits not
available from market transactions vary and include:
• transaction costs (Coase himself);
• exploitation of assets which because of their special characteristics cannot be fully
exploited in the market; and
• the role of the entrepreneur.
Firms differ from the market in directing resources rather than leaving allocation to the
market and involve hierarchies of personnel through which this direction occurs and in
being characterized by incomplete contracts.16 Under these theories, the firm acquires

16
Coase’s original statement of his position, “The theory of the firm” (1937) 4 Economica 386 (republished
often including in Coase, The Firm, the Markets and the Law (Chicago, University of Chicago Press,
1988)) went according to him at least virtually ignored for 40 years largely because it did not fit into the
dominant approach to microeconomics of the day, see Coase, “The Nature of the Firm: Influence” in
Williamson and Winter eds, The Nature of the Firm (New York, Oxford University Press, 1991) at 61.
More recently his position has been elaborated by a variety of scholars, notably Williamson who in addition

9
inputs from the market up to the point that it cannot produce the input internally and
make a profit on it compared to the market price, but the firm expects to make profits on
its use of inputs acquired from the market (as otherwise it would leave the next step to the
market). In terms of outputs, the firm sells to the market only at the point that it cannot
make any additional profit on internalizing the sale into the market.

Although it is not the purpose of the theories to explain where profits are generated in a
geographical sense, it is but a short step to see the profits as located where the firm
operates, that is, where its resources are located and directed.17 This does not mean that
firms should be taxed entirely separately from their shareholders. Although such separate
systems of corporate taxation have existed in most countries at some point in time, there
is now general agreement that it is the ultimate tax burden on the individual shareholder
that matters. But the idea of the firm generating profits beyond market returns by the
direction of resources has provided the justification for taxing the corporation on all its
profits in countries other than the residence of the shareholder and has indicated which
countries are the appropriate ones to collect the tax (the countries where the resources are
directed).

We can trace such a view to the very origin of the transfer pricing rules, even before
Coase published his theory. The author of the rules, Carroll, explained his view of them
as follows,18

[I]n the usual case where an enterprise has its principal establishment in one
country and secondary establishment in others … the real centre of management
is probably at the principal establishment. The control and management, financial
and technical, are centred there. At the meetings of the directors the decisions are
taken which make or break the enterprise. There the risks are centred. The profit
or loss results from all the activities of the enterprise taken together, but how can
the part attributable to the establishment in each country be most readily

to his own major works edited a series of essays on the variants on the theory to mark the 50th anniversary
of original publication of Coase’s work, ibid. A recent survey of the literature which expounds an
entrepreneurial view of the firm is Sautet, An Entrepreneurial Theory of the Firm (London, Routledge,
2000). For the purposes of this article, the interest is in what unites the modern variants on Coase rather
than what divides them. In this view they are able to stand together in the sense that they identify a number
of features which are characteristic of firms. No attempt is made here to rank or critique the theories; rather
existing transfer pricing rules and their development are examined to see whether they fit with the theories
and what the implications of the theories are for international division of the business income tax base.
There are other views which also explain important features of the firm (particularly as it is represented in
the modern corporation) but do not require as Coasean theories do that firms exist because of the profits
that can be earned from the direction of resources. They include the view of the corporation as a nexus of
contracts and the work on agency problems in corporations.
17
There is some work which links the theories directly to transfer pricing, Holmstrom and Tirole, “Transfer
Pricing and Organizational Form” 7(2) Journal of Law, Economics & Organization 201 (1991), Sautet,
note ? at .
18
Carroll, Taxation of Foreign and National Enterprises Volume IV Methods of Allocating Taxable Income
(League of Nations, Geneva, 1933) at para 677, available at
http://setis.library.usyd.edu.au/oztexts/parsons.html at item 5. An understanding of the firm in the Coasean
sense is evident in this passage. The details of the conclusion reached by Carroll were not accepted at the
time and are still a live issue in international taxation as explained below.

10
 measured? If we recognise the fact that the real centre of management, especially
if it is situated at the principal productive establishment, is the most vital part of
the enterprise, the most practical approach to the problem is to give it the
residuum of profit or loss after allocating to each outlying secondary
establishment compensation for the services it has rendered to the enterprise in
accordance with what would be paid to an independent enterprise rendering such
services.

Thus it is plausible to see the distribution of taxing rights over corporate profits in terms
of Coasean theories of the firm.

Residence of corporations

The way in which this distribution of rights to levy corporate tax among countries in
international taxation occurs is complex. Rather than allocating taxing rights over a
corporation’s profits to countries directly, the corporation was fitted into the residence
and source framework even though at a policy level as noted above all that matters is the
residence of the individual.

This apparent assimilation of individuals and corporations in international tax rules might
be explained on the basis that most or all of a widely held corporation’s shareholders are
resident in the same country as the corporation but that is less and less true for listed
corporations and the discussion that follows focuses on this development by assuming
that the corporation is centered in a different country from at least a significant number of
its shareholders. If shareholder residence were the focus, one would expect the corporate
residence rule to reflect that idea, such as a majority of shareholder’s resident in the
country. Residence rules of this kind are in fact very rare which may be explained by the
practical considerations mentioned above for taxing the corporation in the country of
residence of the shareholder, with the difference that the problem of detecting the
proportion of shareholders, including tracing through interposed entities, would arise in
the residence rule rather in calculating income subject to corporate tax. Indeed such a
rule, for reasons noted above, is equally contrary to the assumption that it is not possible
to apply a look through or conduit approach to taxation of shareholders in widely held
corporations.

At the level of the widely held corporation, corporate residence rules effectively rather
adopt the location of the corporate headquarters.19 This test may reflect a relatively

19
This is more evident in civil law countries where tests more or less directly refer to the headquarters. In
countries influenced by the UK the test is central management and control, which is generally taken to
mean where the board of directors takes its decisions but this usually equates to the corporate headquarters.
Many countries include a place of incorporation test as well as a headquarters type test but generally listed
corporations will ensure that they coincide to avoid dual residence problems. The US is the best-known
contrary example of a country using only a place of incorporation test for corporate residence. For a
description of the rules see Ault and Arnold, note ? at 349-350. Even for the US in practice for listed
corporations the place of incorporation coincides with corporate headquarters. As was seen recently in the
case of corporate inversions, Congress is likely to intervene if attempts are made to separate the place of
incorporation of widely held corporations from US located corporate headquarters, IRC §7874. The extract

11
unthinking anthropomorphic view of the corporation as a person20 and on its own would
suggest that the country of the corporate headquarters taxes the corporation on its
worldwide profits.

Transfer pricing rules

In fact very few countries,21 not including the US, really tax multinational corporations
on their worldwide profits by reference to the notional residence of the parent corporation
based on its headquarters, even though most countries nominally tax corporations on a
worldwide basis, including the US. The nominal residence rule for corporations performs
other purposes. To understand the actual operation of the system it is necessary to
introduce the transfer pricing rules. As well as a corporate residence rule based on the
location of headquarters, the transfer pricing rules are also built on another central
concept, the permanent establishment or PE,22 which is constituted by a corporation
having in a country a fixed place of business or dependent agent that carries on (part of)
the corporation’s business.

The transfer pricing rules then provide as follows:

(1) A corporation may not be taxed in a country other than its residence unless it has a PE
there.
(2) The profits which a corporation makes at its residence in relation to its dealings with
associated corporations resident in other countries are determined on the basis of the

quoted above note ? from Carroll who was from the US clearly adopts a headquarters view of residence.
The tie breaker in tax treaties for dual resident corporations is the “place of effective management,” see
OECD, Model Tax Convention on Income and on Capital Condensed Version (OECD, Paris, 2005) article
4(3) at 26.
20
The classic UK case on residence of corporations developing the central management and control test as
a judge made rule, De Beers Consolidated Mines Ltd v Howe [1906] AC 455, may fall into this category,
“In applying the conception of residence to a company, we ought, I think, to proceed as nearly as we can
upon the analogy of an individual. A company cannot eat or sleep, but it can keep house and do business.
We ought, therefore, to see where it really keeps house and does business. An individual may be of foreign
nationality, and yet reside in the United Kingdom. So may a company. Otherwise it might have its chief
seat of management and its centre of trading in England under the protection of English law, and yet escape
the appropriate taxation by the simple expedient of being registered abroad and distributing its dividends
abroad.” Under Coasean theories of the firm, it is possible to construct a (fairly weak) justification for
worldwide taxation by the country of the headquarters of a corporation. The justification would see the
direction of the firm’s resources driven by the corporate headquarters as supporting corporate tax on those
profits by the country of the headquarters even though the activities directed may have occurred in other
countries. Such a view is suggested by the passage from Carroll quoted above, note ? but the point that the
headquarters may warrant a special allocation of profits can be dealt with in other ways as noted below in
the text.
21
Sweden and New Zealand seem to be the only countries that seek to tax active business income of a
multinational corporation (including subsidiaries) on a current basis if the parent corporation is based there,
Vann, Trends note ? at 61.
22
This term is not very “English” in any version of the language. That is because it is a (bad) translation
derived from German via French, see Avery Jones et al, “Origins of Concepts and Expressions Used in the
OECD Model and their Adoption by States” 60 Bulletin for International Taxation 220 at 234-235 (2006).
The PE terminology is standard in international tax language (which courts have recognized as a specialist
language) and is used here. A more meaningful English rendering might be “branch or agency” but that
phrase is both cumbersome and inaccurate for reasons that will appear.

12
profits it would make if it were an independent corporation dealing on normal market
terms with the associated corporations.
(3) The profits which a corporation makes at a PE (and which may be taxed in the PE
country) are the profits which the PE would make if it were an independent corporation
dealing on normal market terms with the rest of corporation of which it is a part.23

In short the firm, whether constituted by a group of corporations each operating only in
its country of residence, or by a single corporation with PEs in other countries, or by a
combination of corporations and PEs is effectively divided up among the countries where
it operates. The profits that are treated as arising in each country are those that would
arise if the various parts of the firm were independent and dealing with each other in the
market on ordinary market terms. To take a very simple example, if a parent corporation
in one country manufactures goods there and transfers them to a subsidiary in another
country which sells the goods in that country, the manufacturing profit will be taxed in
the country of the parent and the selling profit in the country of the subsidiary as a result
of rules (1) and (3). If instead the goods were manufactured at the headquarters of a
corporation and transferred to a PE in another country and sold there by the PE, the
manufacturing profit would be taxed in the country of the headquarters and the selling
profit in the country of the PE as a result of rules (1) and (2).

In recognition of the two critical aspects of the rules they are often referred to collectively
as the “separate enterprise arm’s length” principle or variants thereon. “Separate
enterprise” recognizes that the firm is effectively divided up on geographical lines and
“arm’s length” the use of market prices between the parts to allocate profits. The separate
enterprise part of the principle means that the international tax system is premised on

23
The typical tax treaty provisions state as follows:
Article 7 Business Profits
1. The profits of an enterprise of a Contracting State shall be taxable only in that State unless the enterprise
carries on business in the other Contracting State through a permanent establishment situated therein. If the
enterprise carries on business as aforesaid, the profits of the enterprise may be taxed in the other State but
only so much of them as is attributable to that permanent establishment.
2. [W]here an enterprise of a Contracting State carries on business in the other Contracting State through a
permanent establishment situated therein, there shall in each Contracting State be attributed to that
permanent establishment the profits which it might be expected to make if it were a distinct and separate
enterprise engaged in the same or similar activities under the same or similar conditions and dealing wholly
independently with the enterprise of which it is a permanent establishment.
Article 9 Associated enterprises
1. Where
a) an enterprise of a Contracting State participates directly or indirectly in the management, control or
capital of an enterprise of the other Contracting State, or
b) the same persons participate directly or indirectly in the management, control or capital of an enterprise
of a Contracting State and an enterprise of the other Contracting State,
and in either case conditions are made or imposed between the two enterprises in their commercial or
financial relations which differ from those which would be made between independent enterprises, then any
profits which would, but for those conditions, have accrued to one of the enterprises, but, by reason of
those conditions, have not so accrued, may be included in the profits of that enterprise and taxed
accordingly.
See OECD, note ? at 28-30.

13
taxing each corporation separately, and accordingly recognizing each corporation’s
separate residence by applying the normal residence rules for corporations.

In terms of the theory of the firm, the location of the value generated within the firm is
seen as an attribute of the firm’s activities more generally and so spread over the firm
wherever its resources are being directed. In that sense, the allocation of income under
the transfer pricing rules across countries where the firm operates fits with the theory.
The use of a standard of profits earned on market transactions to allocate that value,
however, seems to contradict the theory which is generally premised on the use of the
firm to earn additional profits compared to what is available from market transactions. I
return to these issues in the next section.

Relief of international double taxation for corporations

The final rules that it is necessary to notice in understanding the actual operation of the
transfer pricing rules are those for relief of international double taxation. Because
corporations are regarded as resident at the place of the headquarters, the headquarters
country under the normal operation of residence and source concepts would tax the
worldwide income of the corporation and be obliged to give relief for tax levied at
source. In fact in most cases, there is no tax levied in the country of the headquarters on
income that is treated by the transfer pricing rules as taxable in another country. The
route by which this result is reached varies from country to country. Most directly some
countries use a comprehensive exemption system for corporations so that profits of PEs,
dividends from foreign subsidiaries and capital gains on shares in foreign subsidiaries are
exempt from tax in the country of the headquarters (in the case of foreign PEs) or
residence of the parent (in the case of foreign subsidiaries).24

To the extent that a country uses a foreign tax credit for corporations (such as the US),
the lack of tax in the country of the corporation’s residence arises in the case of PEs
generally because of the rough equivalence of corporate tax rates around the world
combined with considerable foreign tax credit tax planning. For foreign subsidiaries, tax
in the country of the parent is generally only triggered when the subsidiary pays
dividends or the parent sells shares in the subsidiary. Moreover, in the case of dividends,
the foreign tax credit is extended to the corporate tax paid by the subsidiary not just
source taxes on the dividends. Foreign tax credit planning is thus available for dividends
in the same way as for PEs. More importantly both payment of dividends and sale of
shares are in the control of the parent and so it is possible to postpone tax in the country
of the parent, which is generally referred to as deferral.25

As already noticed, deferral is really an issue with respect to tax on the ultimate
shareholder. Deferral is only a problem in the country of the parent corporation’s

24
Some countries use exemption fairly generally for foreign income while others use it only for
corporations. Even in the countries which use the system generally there are special rules for corporations
which mean that there is different treatment in most countries compared to individuals.
25
Deferral is an endemic issue in the income tax but the use of the term in this specific sense has a long
history.

14
headquarters if the shareholder is resident in the same country. Otherwise application of
the foreign tax credit in this country does not make much sense as the residence rule for
the corporation is essentially operating through the transfer pricing rules to locate income
in the country, not to tax worldwide income. If it is thought that the activities of the
headquarters deserve special recognition, because, for example, the headquarters is
responsible for the strategic direction of the firm, that can be more clearly achieved by
allocating more of the firm’s profit to the headquarters under the transfer pricing rules
rather than residual taxing rights under the foreign tax credit mechanism.26 There is a
noticeable trend around the world to exemption systems for relieving international double
taxation of corporations, including in the US though only for the calendar year 2005 for
dividends from subsidiaries.27

The outcome of the way in which international double tax relief systems work for
corporations is that the application of the same residence and source regime to
corporations as to individuals is only apparent, not real. The residence of the corporation
is used along with the PE and transfer pricing rules to connect income (profits) to a
country, not generally to tax on a worldwide basis. In the sense that source is the
connection between income and a country, the residence rule for corporations, the PE
rule and the transfer pricing rules together are sourcing rules.

Transfer pricing and source taxation

It follows from the conclusions of the previous sections and here that the corporate tax on
listed corporations is essentially a source tax. The pitching of the corporate tax rate
around the world is consistent with this characterization. A standard policy prescription
in a purely domestic context is that the corporate tax rate should be equal to the top
marginal tax rates for individuals to eliminate deferral of shareholder tax entirely for
closely and widely held corporations. Nonetheless over long periods of time and in most
countries, the corporate tax rate has typically been significantly lower than the top
individual marginal tax rate. This is consistent with the view of the corporate tax as a
source based tax. It will be recalled that source tax rates are generally flat rate and lower
than top progressive tax rates.28 The US is currently an exception with alignment of the
federal corporate and top marginal rates at 35% but over the longer term the US has also
fitted the standard pattern.

26
At least one country, Switzerland, has long expressly incorporated a return to management at the
headquarters in its tax system. The quote from Carroll, note ?, adopts a similar view.
27
IRC §199. The push for a general exemption system for corporations in the US remains strong but the
debate is generally framed in the traditional terms of capital export neutrality, capital import neutrality and
national welfare. It is not the purpose of this article to argue for the exemption system as such but rather to
explain why corporate tax is not collected to any degree in either the country of the shareholder or the
country of corporate headquarters if the income is allocated by the transfer pricing rules to another country.
The implication, however, is that the current debate over exemption is misdirected to the extent that it
ignores the residence of the shareholder and the method of integrating corporate and shareholder income
taxes, see Vann, Trends note ? at 53-63.
28
In many countries the corporate rate is nowadays also similar to the flat rate charged in national law on
passive income taxed on a source basis.

15
Whether this be the explanation for the corporate tax rate or not, to the extent that the
residence of the shareholder and/or the headquarters of widely held corporations are
different from the country to which income is attached by the combination of the
corporate residence, PE and transfer pricing rules, it is the last country that will collect
the bulk of the tax on the income generated by the corporation. That tax will be generated
by the corporate tax levied on the income allocated to the country under the transfer
pricing rules.

It should be noted that the international system tackled the problem of how to divide
taxing rights between countries when transactions occurred across borders such as
manufacture of tangible goods in one country and sale in the other as an essentially stand-
alone exercise. Combined with the fact that the residence rule for corporations is
essentially a sourcing rule as just described, not surprisingly the current international tax
system exhibits both complexity in its handling of residence and source issues29 and does
not reconcile the principles underlying traditional flat rate final source taxes described
earlier and the transfer pricing rules. Transfer pricing rules differ in a number of respects
from simple sourcing rules of the kind described. Traditional source rules see all income
from a particular productive resource as located in one country; by contrast, transfer
pricing rules involve allocating profits from a particular productive process between
countries. The source taxes are often levied on a gross final basis while the transfer
pricing and associated rules effectively require taxation on a net basis.30 The transfer
pricing rules were for many years seen as effectively anti-avoidance rules while the
traditional source rules were seen as systemic. Tax treaties either significantly reduce or
suppress entirely the traditional source taxes except in the case of real estate but require
the application of the separate enterprise arm’s length principle to allocate business
income.31 These differences will be dealt with at various points in the next sections, but
no final solution is offered though some possible directions are indicated.

The recognition of corporate residence as a sourcing device also throws light on another
common criticism that the separate enterprise part of the transfer pricing rules fails to
recognize the economic unity of the corporate group. If corporate residence is viewed as
a true residence rule as the formal nature of the rule for corporate residence rule suggests,
then it is artificial to break up the residence of members of the group based on whether
there are separate corporations or not. We do not divide up the residence of an individual
based on where the individual’s income earning activities occur – that is the function of
29
Vann, “Reflections on Business Profits and the Arm’s-Length Principle” in Arnold, Sasseville and Zolt
eds, The Taxation of Business Profits Under Tax Treaties (Toronto, Canadian Tax Foundation, 2003) 133
at 142-146.
30
Tax treaties generally permit gross basis taxes for dividends, interest, royalties and rent but effectively
require net basis taxation for business income, see OECD, note ? articles 6, 7(3), 10-12, 24(3), (4) at 28-33,
39-40.
31
It is common to tie a number of these features to a distinction between active and passive income, for
example, Avi-Yonah, “The Structure of International Taxation: A Proposal for Simplification” 74 Texas
Law Review 1301 (1996), that is, source taxation of passive income (apart from real estate) is very limited
and taxed at final flat tax rates; business and services income is active and is taxed at source without limit
once a basic threshold is passed but the tax is levied on a net basis. While true and reflective of the history
of international taxation in a broad sense, the distinction does not explain all the differences noted in the
text.

16
source principles. If we recognize the corporate residence rule as at bottom a sourcing
rule like the PE rule, this form of the criticism is not compelling.

What is a problem rather is the lack of robustness of the corporate residence rule within
the group. We have noted that at the level of listed corporations residence will generally
coincide with the headquarters, despite the variety of expression of corporate residence
rules. Within the group, all forms of corporate residence rule are the subject of tax
planning, because the place of incorporation, headquarters or board meetings of
subsidiary corporations can be easily manipulated, especially if the corporation in
question carries out few activities. Hence the creation within the corporate group of
corporations resident in tax havens is easily organized. In itself this corporate residence
manipulation is only really a concern for highly mobile income like interest.32 If the
transfer pricing rules for allocating business income were robust, little income could be
allocated away from countries where business is actually done. But the transfer pricing
rules are not robust for the structural reasons explored in the next section.

Summary

This section has introduced the transfer pricing rules in international taxation and linked
them to Coasean theories of the firm. Contrary to the appearance of their being at least in
part a residence rule based on the headquarters of widely held corporations, the principle
in fact provides sourcing rules which allocate business income between countries. The
effect, when combined with mechanisms for relief of international double taxation, is that
most corporate tax is collected in the country to which income is allocated under the
transfer pricing rules. Relatively little tax is collected by the country of (nominal)
residence of the corporation or the country of the ultimate individual shareholder if the
rules allocate the income elsewhere. Hence the effectiveness of the transfer pricing rules
is central to the effectiveness of the corporate tax in relation to multinational
corporations.

To the extent that the residence of the ultimate shareholder is not in the same country as
the headquarters of the corporation, we have also seen that some current international tax
rules have a questionable basis. The headquarters country of the corporation has little
policy claim to tax more than the income allocated to it under appropriate transfer pricing
rules.

III Operation of the transfer pricing rules

This section seeks to isolate the structural weaknesses of current transfer pricing
principles that make them ripe for tax planning. Over recent years the detail of the rules
32
The source of interest is simply a question of which financial institution where is chosen for making
deposits. Tax systems respond to this problem within corporate groups by controlled foreign corporation
(CFC) regimes that tax passive income derived by corporate groups at the place of the headquarters
(residence) of the parent company, see Ault and Arnold, note ? at 377-386. Like many other rules discussed
above, there is an implicit assumption that the ultimate individual shareholders will be resident in the same
country. If they are not, the taxing claim of the country of the headquarters of the corporation is not
obvious. In some countries, the CFC rules are also extended to transfer pricing situations. I return to these
rules in the discussion of potential solutions to transfer pricing tax avoidance.

17
has expanded as a result of OECD work and parallel elaboration in national legislation
and practice.33 The details themselves provide many tax planning opportunities but that is
not the focus here. If it is possible to get the framework right, the details can be left to the
courts or other action at the national level to ensure that the details match the framework.

The discussion starts with the definition of the firm to which the rules are applied. It then
discusses the significance of market prices between independent enterprises as the
standard of allocation, particularly its implications for value that may be created above
market prices in the firm and for using legal transactions (contracts) as the basis of the
rules. Finally, the functional analysis at the heart of modern transfer pricing practice is
examined. This methodology requires that profits be allocated among countries on the
basis of the “functions performed in the light of the assets used and risks assumed”
(commonly referred to as FAR for functions, assets, risks). Rather than being a unifying
and robust methodology as intended, it has been at the centre of much of the tax
avoidance activity. The treatments of the following areas are considered in relation to the
functional analysis: risk, personnel, assets and sales. In each of them the methodology has
been manipulated in practice to allow diversion of profits to low tax jurisdictions.

A The firm
Because we are discussing the international division of the business income tax base, we
need to divide the firm up in a geographical sense and relate the firm to specific
countries. As noted above such a geographical connection is not the general concern in
theories of the firm. Nonetheless the Coasean theories posit a distinction between the firm
and the market based on the direction of resources rather than market allocation of
resources and this essential feature can be used geographically – if we find resources in a
country that are being directed in this sense, then the firm is present in the country.

The current international rules in part seem to adopt this approach and in part not to. It
was noted above that the rules use the residence of the corporation and the PE to describe
the necessary connection with the country. At this point it is necessary to expand slightly
on these concepts. The definition of the PE includes not only fixed places of business,
that is, facilities in a country occupied in a legal sense by the corporation itself,34 but also
dependent agents who enter into sales contracts with third parties on behalf of the
corporation. Dependent agents cover employees and non-employees, including other
corporations. The dependency test turns on whether the agent is legally and economically
independent of the principal or not. The idea is one of a dependency attachment to the

33
The current OECD rules are contained in Transfer Pricing Guidelines for Multinational Enterprises and
Tax Administrations (Paris, OECD, 2001) which deal with associated corporations. The position on PEs is
currently being developed in a four part discussion draft which is referred to in part later in the article; the
final version of the first three parts is due for publication before the end of 2006. For the US the rules are
contained in the regulations under §482. Together this material amounts to many hundreds of pages. More
generally see Ault and Arnold, note ? at 420-425.
34
The notion of permanence implied by PE terminology has two aspects. The location must be relatively
static (not moving equipment for example) and last for generally 12 months. For agents the permanence
idea means that the agent must act for the principal “habitually.”

18
principal and corresponds to the direction of resources idea in Coasean theories.35 In the
case of separate corporations, the required relationship is that the corporations be
“associated enterprises” which in turn is expressed in terms of control. Although control
could be viewed as another way of stating the dependency attachment idea, in fact it is
generally taken as an ownership test (that is, parent and subsidiary and similar
relationships).36

There are two fault lines in these rules. So far as the PE is concerned, the dependency
attachment test is subordinate to the agency test so that if agency is avoided (and agency
here is the legal concept, not the commercial concept of the term), it does not matter if
there is dependency. In the case of the associated enterprises, the dependency idea in the
PE rule has been replaced with an ownership test. While ownership will usually imply a
dependency attachment, dependency is possible without ownership. It is obvious in the
case of employees that “ownership” by the firm is not required for the employee to
constitute part of the firm. The direction of the supply of the employee’s labor is
sufficient. Coase came to regret that employees loomed so large in his original exposition
of direction of resources being the hallmark of the firm as attention was diverted from
other relationships in which the necessary direction was present.37

The combination of these fault lines means that some dependency situations will fall
outside the definition of the firm in the international tax rules. Some possible examples
are certain kinds of franchise and distribution arrangements. If the franchisee or
distributor is a separate corporation, it will not be viewed as part of the firm constituted
by the franchisor or producer if it is not owned by the franchisor or producer of the
products being distributed and is not an agent. As franchisees and distributors typically
sell in their own right and not as agent, the test essentially comes down to ownership. The
significance of this structural problem in the rules will become evident as we proceed
particularly regarding the significance of transactions and sales. It is not often highlighted
as one of the problems in transfer pricing rules which may mean that a Coasean policy
basis is implicitly rejected for defining the firm. If so, it is not clear what the underlying

35
The distinction can be seen clearly in two extremes. If a firm uses a large stockbroker with many clients
to sell the some of the firm’s portfolio shareholdings in listed corporations, then assuming that the broker is
acting legally as agent which is the case in some countries at least, the broker would not constitute a
dependent agency PE as the broker is independent. Both the broking and the sale transaction are normal
market transactions. By contrast if a firm which manufactures unique goods sells them through agents in
various countries and the agents only act for that firm and under strict guidelines, the agents will be
dependent agents as they are a part of the firm not much different from employees. The agency contract
will not be a normal market transaction in the sense that the firm expects to make at least some of its profit
out of its direction of the selling activity of the agent, in addition to its manufacturing profit. The term
“integration” can be used to describe the idea as it conveys that the person or asset is part of the firm.
Because the term “integration” has been used above as is also common to describe systems that try to
overcome double taxation of corporation and shareholder, the terms “dependency” and “dependency
attachment” are used here to describe when a person or asset is considered to be part of the firm.
36
The ideas in this part of the article are considerably expanded in Vann, “Tax Treaties: The Secret Agent’s
Secrets” [2006] British Tax Review 345 (hereafter “Secret Agent”).
37
Coase in Williamson and Winter note ? at 64-65.

19
policy basis of the rules is.38 While it is possible to posit some cases where dependency is
present but ownership and agency are not, it is more problematic to provide a direct test
of dependency that is sufficiently detailed but not subject to manipulation.

In summary this part of the article has suggested that a dependency attachment should be
the organizing principle in defining the firm for international tax purposes. Current rules
adopt dependency but add requirements of ownership or agency which have the effect of
bringing the rules up short of the principle.39

B Market price benchmarks: transactions, dealings and legal entities

To paraphrase from earlier discussion, the benchmark for allocating a firm’s profits
across the countries where the firm operates is the profit that would arise if the part of the
firm in each country dealt with each other part of the firm in other countries as if they
were independent corporations transacting in the market in the normal way. As already
noticed, the benchmark appears to contradict the Coasean theory of the firm which posits
that a firm organizes because it is able to make profits which are not available in market
transactions – the reasons for the additional profit varying to some degree in different
versions of the theory. The firm does contract with the market for inputs and outputs but
the profit that arises after taking account of these contracts is attributed to the internal
activities of the firm. It is these internal activities of the firm to which the market price
benchmarks are being applied in allocating profits across countries for corporate tax
purposes.

The application of the market price benchmark may be an implicit denial that firms exist
for the reasons expressed by Coasean theories. Certainly that denial has been expressed in
other contexts and may underlie theories that do not require there to be any additional
profits for firms to form. This style of argument has been deployed in recent times in
relation to international taxation of dependent agency PEs to suggest that even if such a
PE exists, there is no profit to tax to the principal in addition to the separate taxation of
the agent on its profit.

38
The PE concept is frequently criticized as a threshold test for source taxation, see the work of Arnold and
Pinto referred to in note ?. The criticism is not, however, formulated in the terms set out here as the PE
concept is seen as all or any of mechanical, outdated, or administration based and hence a barrier to source
taxation, rather than as a principled, if flawed, concept for sourcing business profits.
39
Further, the PE rules have exceptions which were originally intended to be de minimis as indicated by
their description of “preparatory or auxiliary.” While the exceptions do not cover firms whose very
business is the activity in question, it is not clear if there is an overall preparatory or auxiliary limit on the
exceptions, and nowadays the listed activities include significant value adding elements – purchasing,
warehousing, delivery, advertising, collection of information and research, but not after sale service. In the
case of purchasing activities, no profit is attributable to the activity even if there is otherwise a PE. These
exceptions hark back to a different world when international trade consisted mainly of raw materials and
finished goods which more or less sold themselves. They are the subject of considerable manipulation and
indicate the problem of using proxies instead of direct rules. If it is desired to have a de minimis rule so that
firms are not subject to the considerable compliance costs entailed in filing a tax return in a country, the
rule would best be expressed directly through some kind of threshold such as the level of turnover or of
purchases (the former is common in value added taxes).

20
The argument runs as follows. Assume that a dependent agent in a country is selling
goods manufactured abroad by a foreign principal. The agent is a separate legal person
from the principal and enters into an agency contract with the principal. The price in that
contract will reflect the market value of the agent’s services, or, if the agent is an
associated corporation of the principal in an ownership sense (such as a subsidiary), the
agency contract can be adjusted to the market price under the transfer pricing rule that
apples to associated corporations. To the extent that the agent constitutes a part of the
principal’s firm under the dependent agency PE rule and the principal is to be taxed in the
PE country, the revenue of the principal attributable to that country will be the market
value of the agent’s services because that is the activity of the principal that is carried on
there and hence is the amount of revenue that the principal should have attributed to the
PE. This amount is given by the actual contract price with the agent or the adjusted price
of that contract if the agent is an associated enterprise. The expense of the principal at the
PE will be the fee paid to the agent under the contract (or the adjusted fee if the agent is
an associated corporation). Revenue of the PE thus equals its expense and nothing is
taxable to the dependent agency PE, which is sometimes described as the nil sum view.40

Although it would seem quixotic for the officials entrusted with devising the international
tax system to spend 80 years creating rules that it now turns out leave nothing to tax in a
country, this argument is widely accepted in the international tax profession and is the
source of a significant amount of tax planning. It has not been the subject of an adequate
official reply.41 My response is twofold: first, that the international system has from the
beginning accepted that there is an additional profit to allocate above the market prices of
the actual external (non-dependent) inputs of the firm, and secondly, that the profit may
be allocated in whole or in part in many cases by positing appropriate transactions within
the firm and using market prices for those transactions effectively to allocate the profit
generated by the existence of the firm. The most explicit recognition of this response is
the OECD sanctioned transfer pricing method called the “residual profit split.”42 Under
this method the total profit is allocated in part by market prices of (presumed)
transactions among the parts of the firm and as to the remainder by some appropriate
apportionment “keys” representing the contributions of various parts of the firm in
addition to the transactions. In my view it is clear that the international system accepts the
Coasean theory of the firm despite the apparent oddity of the market price benchmark.

40
For the detailed history of the relevant rules and my response to the nil sum view, see Vann, “Secret
Agent” note ?. The nil sum argument at the moment is mainly deployed in relation to subsidiaries. It has
not been generally used in relation to employees who constitute an agency PE or to fixed place of business
PEs though it could equally be used there with the result that there would not be any taxable profit in any
PE.
41
The current official response appears in OECD, [Revised] Discussion Draft on the Attribution of Profits
to Permanent Establishments: Part I General Considerations (2004) available at
http://www.oecd.org/dataoecd/22/51/33637685.pdf at 60-65 which is criticized in Vann, Secret Agent, note
? at 376-380.
42
In this article I use profit or income to refer to the firm’s overall profit. In the usage ‘residual profit split’
the term residual refers to the profit remaining after allocation of some of the profit by the market prices of
transactions.

21
The apparent oddity is to be explained by the assumptions that were made by most
countries during the many years over which the transfer pricing principles were
developed and refined. Returning to the situation of the agent above, the assumption was
that the dependent agency PE would have allocated to it as revenue the selling price of
the principal’s goods rather than the value of the agent’s services. The principal through
its dependent agency PE is selling its goods and the revenue of the PE is derived from the
sales not from the agency activities. In other words there is a presumed sale by the parts
of the principal outside the PE country to the PE in the country at the (manufacturing)
market price and then a sale by the PE to the ultimate buyer at the (selling) market price.
It is the market price of that presumed contract which is relevant under transfer pricing
rules, not the market price of the agency contract.

Under this approach the overall profit of the firm is effectively split between the country
of manufacture and the country of sale with the manufacturing profit taxable in the
former and the sales profit taxable in the latter. In this simple case the full profit is
allocated between the manufacturing and selling country and there is no need to go any
further. It will be noted that this is the same allocation between countries as would apply
if the transactions were structured with manufacture by a foreign parent, sale by the
parent to a local subsidiary and sale by the local subsidiary. The rules in other words
were structured on the basis of presumed ways of dealing among the parts of the firm
which by a convenient short cut allocated the Coasean profits in what were the common
cases in past years.43

In the case of PEs the contracts generally have to be presumed because the PE is simply
one geographical part of the same legal entity. This is true for both fixed place of
business and dependent agency PEs although in the case of the latter there tends to be
confusion of the agent in its own right with the PE (and hence of actual contracts with
presumed contracts). Although the PE is constituted by the dependent agent’s activities, it
is those activities viewed as part of the firm of the principal, which they are because of
their dependency attachment to that firm, that constitutes the PE. This confusion in part
underlies the arguments that no profits are attributable to dependent agency PEs
considered above.

In the case of associated corporations, however, there can be actual contracts in the legal
sense between the parts of the same firm (parent and subsidiary in the simple example
here). The issue then is whether those actual contracts trump the presumed contracts that
would be used in a PE situation. Little attention was given to this issue in the original
framing of the rules for associated corporations as at the time the problems of allocating
profits between countries seemed mainly to arise through PEs.44 It was no doubt assumed
that actual and presumed contracts would line up. In fact corporations quickly exploited
the possibilities of varying the actual contractual arrangements from the presumed ones.
It was easy to read the way in which the transfer pricing rules were formulated as

43
It was at this critical point that country practice departed from the views of Carroll, quoted above n ?.
Carroll rejected such presumed sales and viewed a PE as providing services to the head office.
44
Corporate laws in some jurisdictions did not permit corporations to own shares in other corporations
which may partly explain this position.

22
permitting this freedom of contract which is the hallmark of the market and the rules have
been interpreted as based on freedom of contract within fairly broad limits.45

Thus it is possible to vary the contracts in the parent and subsidiary situation so that the
parent legally sells the goods in the country of the subsidiary and the subsidiary simply
provides assistance in making the sales without rising to the level of an agent in the legal
sense. The only contract to price for taxation purposes in the country of the subsidiary is
the sales assistance contract between the parent the subsidiary. The parent ensures that it
does not have a PE and so is not taxable in the country of the subsidiary while the
subsidiary performs an apparently minor role which is appropriately rewarded with a
market contract price. In reality little has changed but the nil sum view is effectively
achieved by different contracts.46 The choice of switching the pattern of the contracts
means that the allocation of the total profits is essentially within the power of the
corporation.

The theory of the firm indicates that the source of the problem is the assumption of the
market freedom of contract. The way in which the subsidiary is acting in these cases is at
the direction of the parent and the subsidiary’s activities are dependent on those of the
parent. Another way of expressing the point which is common in the literature is that
firms are characterized by incomplete contracts (though arguments range back and forth
about whether incomplete contracts are characteristic of firms or a much more general
market transaction). On the incomplete contract version, the only actual contract between
parent and subsidiary here is incomplete in that it is really a contract for much more than
minor selling services.

There are other dimensions to what is essentially the same problem. The contracts
between associated corporations may be obviously incomplete but nevertheless are priced
on the basis of a different complete market transaction. This occurs in the case of
licensing of intangibles. Contracts between associated corporations may be written as
non-exclusive licenses so that powers of enforcement remain with the licensor
(depending on intellectual property rules in different jurisdictions) but are treated by the
corporations for transfer pricing purposes as if they were exclusive licenses as that is the
equivalent market transaction. While it is possible in some cases to find comparable
market transactions, even though they do not represent the actual incomplete contract, in
other cases there will not be a comparable market transaction because separate firms

45
It was only in 1995 that principles were formulated for the disregard of transactions between associated
corporations but despite their very broad expression, the principles were stated to be exceptions to be
utilized rarely. The exceptions are “where the economic substance of a transaction differs from its form
[and] where, while the form and substance of the transaction are the same, the arrangements made in
relation to the transaction, viewed in their totality, differ from those which would have been adopted by
independent enterprises, behaving in a commercially rational manner and the actual structure practically
impedes the tax administration from determining an appropriate transfer price.” Transfer Pricing
Guidelines, note ? at para 1.37.
46
For more details, see Vann, Reflections, note ? at 154-157.

23
would not have entered into the transaction in question as the nature of the contract
depends critically on the dependency attachment of the parties to the same firm.47

Whether expressed as a lack of freedom of contract or as incomplete contracts, in the end

the clear object of the transfer pricing rules was not to produce different allocations of
profit in what are essentially the same circumstances. Either associated corporations have
to be limited in the type of contractual arrangements that are acceptable for transfer
pricing purposes or the actual contracts have to be replaced by presumed contracts. The
attraction of the former approach is that it is using actual contracts which feel familiar to
taxpayers and tax officials. Even so in the associated corporations case it will not always
be possible to find equivalent market transactions. Current transfer pricing detailed rules
specify circumstances where only certain kinds of contracts are acceptable. The view
which is being stated here is thus not new but it must be much more widely applied. The
current freedom of contract is one of the most significant structural weaknesses in
transfer pricing.

Returning to the PE situation, that fact that contracts have to be constructed as no real
relevant contracts generally exist (with an exception noted below),48 does not necessarily
preclude notional freedom of contract. The early development of the rules contained a
number of rules of thumb which limited any notional contractual freedom. In the recent
and on-going reconsideration of PE transfer pricing rules, it appeared for a time that the
OECD was going to extend the same freedom of contract to constructed contracts. More
recently as the OECD has begun to finalize its views on the application of transfer pricing
rules to PEs it is creating or preserving more and more presumptions in the PE rules
which effectively amount to restrictions on the freedom of contract. In later parts of this
section of the article a number of these presumptions will be noted as they are the best
current indication of how restriction on freedom of contract might be implemented. To
the extent that these presumptions appear sensible, they should be extended to associated
corporations.

That the same or very similar standards should apply to PEs and subsidiaries is necessary
as the choice of a PE or subsidiary is effectively within the election of the corporation. In
recent times the electivity has become even greater because of hybrid entities. Originally
these were generally exotic entities that different countries treated in different ways for
tax purposes. One country treated the entity as a corporation taxable in its own right and
the other country treated it as a part (and usually PE) of the entity or entities which
owned the hybrid. When the US introduced the check-the-box rules for the classification
of entities as corporations or PEs, it was no longer necessary in cases involving the US to
use an exotic entity to achieve this outcome.49 What two corporations for the purposes of
another country under its corporate, contract and tax law will for US purposes be two

47
As explained later, contracts involving intellectual should in any event be constrained and not permit
licensing transactions at all.
48
Within the corporation there are usually only the various activities of its different parts, which the OECD
refers to as “dealings” to distinguish them from the legal contracts between associated corporations, which
the OECD refers to as “transactions.”
49
Regulations §301.7701-2

24
separate corporations in corporate and contract law but one corporation with a PE for the
purposes of tax law if the corporations so choose. Thus it is possible to have actual legal
contracts (because the corporations are separate legal persons for corporations and
contracts law purposes), but a single corporation with a PE for tax purposes. If freedom
of contract were permitted for PEs it would be possible by this means to remove one of
the difficulties of manipulating PE taxation – ambiguities about what the contractual
arrangements are.50

In summary this part of the article has argued that while the market price benchmark may
be thought to contradict Coasean theories of the firm, in fact those theories do underlie
current rules. However, the freedom of contract permitted under current transfer pricing
rules for associated corporations means that the allocation of the profit, which can be
appropriately allocated under certain market priced contractual arrangements, may be
subverted by different contractual choices. This is a significant structural flaw in current
transfer pricing law which is also inconsistent with the theory of the firm. It is necessary
to constrain freedom of contract in transfer pricing rules in order to allow the market
price to allocate profit appropriately and even then such prices may not fully exhaust the
profit in which case other methods of allocation will be necessary. With PEs, contracts
generally have to be constructed and current and developing rules in the area are moving
in the direction of much less freedom of contract. The kinds of presumptions existing and
emerging in the PE area may provide the way forward.

C Functional analysis and risk

Tax avoidance problems in the transfer pricing area are not new. As noted above the
functional analysis and the accompanying FAR jargon (“functions performed in the light
of assets used and risks assumed”) that came into use in the 1980s was intended to
prevent, not create, tax avoidance opportunities, but the emphasis on risk had precisely
the opposite effect.51 There are two stages in the development though historically there is
considerable overlap. The first stage flowed from the failure to tackle freedom of contract
between associated enterprises. As risk can be assigned by contract more or less at will, it
was possible for associated corporations to move risk around a corporate group as they
wished and with the risk went a substantial proportion of the profit. Some simple
examples are contract manufacturing and captive insurance.

50
Altshuler and Gruber, note 1 have recently identified the check-the-box rules as one significant cause of
loss of corporate tax revenue in the US. As the brief discussion here indicates, the problems are essentially
ones of transfer pricing. The problems mainly show up in the parts of transfer pricing involved in financing
of firms. These issues will not be dealt with in the article. [Tax calculation details for PEs and of
corporations of which they are part will still differ from parents and subsidiaries but not in ways that affect
the point made here– include brief technical discussion in full version].
51
In the US 1986 tax reform, some important changes were made with respect to intangibles and a study of
the whole area was mandated. There followed a 1988 US Discussion Paper on transfer pricing which
introduced the functional analysis and the economist into the transfer pricing area. After this shift in
transfer pricing analysis found its way into US draft regulations, the OECD got involved as many countries
considered that the new US approach was not consistent with the then orthodoxy (as concerned
methodologies rather than theory). The final 1995 “compromise” in the Transfer Price Guidelines
represented a victory for the US approach with the FAR jargon appearing prominently in the Guidelines.

25
The main difference between an ordinary manufacturer and a contract manufacturer is
that the latter takes no inventory risk. It manufactures to the order of another corporation
which carries the risk of being able to sell the manufactured items. Such arrangements are
now commonplace between independent parties. If a member of a multinational corporate
group is carrying on manufacturing activities in a country, at the stroke of a pen it can be
either a full risk manufacturer or a contract manufacturer vis à vis other members of the
group as the group desires, even though with respect to the market generally the group is
a full risk manufacturer. Product liability and related issues have produced a whole
industry of captive insurance which pulls such major risks out of countries where they
arise generally to tax havens. The corporation which otherwise would carry such risk
enters into an insurance contract with an associated corporation which thereafter carries
the risk. Tobacco firms in Australia have diverted 10% of sales revenue through this
route52 and given the current litigation and insurance situation in the tobacco industry,
this figure can probably be upped considerably.53

These kinds of tax planning had been around well before the 1980s development of the
functional analysis but the highlighting of risk in that analysis certainly gave
encouragement to them and they are now standard transfer pricing tax planning fare.
Limits on freedom of contract in the transfer pricing area of the kinds suggested above
should be enough to deal with such basic paper shuffling devices. The second stage of the
developments is more substantive. The main impact is on the three areas to be considered
next but the development of the approach to risk is discussed here.

This next stage in the rise of the importance of risk in transfer pricing can be traced to
work at the OECD, though there was considerable parallel work in many OECD
countries, including the US.54 For present purposes, four developments in the approach to
risk emerged. These developments occurred in the context of PEs and apparently under
an assumption that freedom of contract was not to be applied to PEs, even though
freedom of contract for PEs was actively being contemplated.55

First, analysis of financial innovation involved the decomposition of a broad range of not
just financial transactions and assets into two components: a standard risk-free loan and a
bet with all the risk obviously residing in the bet. Financial innovation was generally all
about the risk side of the transaction – whether taking or eliminating risk through a

52
WD &HO Wills (1996) 32 ATR 168 which held that the Australian general anti-avoidance rule (which
overrides tax treaties) did not apply in this situation; presumably the revenue did not argue transfer pricing
on the basis that the price was within the bounds of market prices.
53
See Vann, Reflections note ? at 153-157 for elaboration.
54
There were three interrelated events at the OECD during the 1990s: work on the tax implications of
financial innovation, work on the taxation of PEs and work on taxation of e-commerce. The
interrelationship was that the PE work was initially concerned with financiers which were the main
businesses to operate in branch form, while e-commerce raised questions about whether the PE threshold
was still appropriate as the foundation of international taxation.
55
It is not surprising that the various relevant documents produced by the OECD over several years in
trying to find consensus on PE transfer pricing issues do not speak with the one voice even in the same
document on freedom of contract and other issues. The documents are drafted by committees and seek to
satisfy simultaneously the positions of as many countries as possible. Consensus is the goal not avoidance
of contradiction.

26
variety of derivatives etc.56 The analysis suggested that returns on assets were relatively
standard and that major profits (and losses) were generated by the exploitation and
management of risks related to assets and liabilities.57

The second development linked the exploitation and management of risk to specific
personnel in corporations. The particular issue was the allocation among countries of the
profits from 24-hour trading in capital markets by global financial institutions involving
personnel in several countries and the profit was seen to arise largely from the activities
of the traders who took or eliminated the risks involved through various trading
strategies.58 This idea was elaborated and generalized as the “key entrepreneurial risk
taking functions” (KERT functions) understood as the “people” functions “which require
active decision making with regard to the most important profit generators of the
business.”59 As for the rest of the firm’s employees, the treatment was the same as assets
– a basic (small) reward for routine functions.

The third element links equity capital of corporations with risk in the determination of the
allocation of interest deductions for firms between the countries where the firms operate.
The route here was via capital allocation of banks and the international use of the risk
weighting of assets found in banking regulation.60 Apparently inconsistently this

56
The OECD produced a 1994 report Taxation of new financial instruments which was concerned with
domestic as well as international issues. Many in the academy were at the same time working on financial
innovation, for example, Warren, “Financial Contract Innovation and Income Tax Policy” 107 Harvard
Law Review 460 (1993), Strnad, “Taxing New Financial Products: A Conceptual Framework” 46 Stanford
Law Review 569 (1994), Edgar Income Tax Treatment of Financial Instruments: Theory and Practice
(Toronto, Canadian Tax Foundation, 2000).
57
An analogous result was reached in the area of e-commerce. Hardware is an asset which generates little
return; all the profit is generated by the (writers of) software, see OECD, Taxation and Electronic
Commerce (Paris, OECD, 2001) which brought together some of the early OECD work on e-commerce to
this effect, see ch 4.
58
The 1994 OECD report note ? led to the creation of the OECD Special Sessions on Innovative Financial
Transactions which started to work on the taxation problems of global trading in financial instruments,.
Detailed analysis of global trading by the OECD led to two draft reports in 1997 and 1998 which
emphasized the role of the traders, see The Taxation of Global Trading of Financial Instruments (Paris,
OECD, 1998).
59
OECD, note ? at 23
60
Another part of the OECD (dealing with transfer pricing) started work on the taxation of PEs after the
work on the 1995 Guidelines was substantially completed in the late 1990s, being particularly concerned
with capital allocation within banks. The capital allocation effectively determines how much equity and
debt are treated as located within particular countries and therefore how much interest expense can be
deducted in each country. The OECD looked in this process at the regulation of banks which had come to
be governed internationally by the Bank for International Settlements Basel accords (now Basel II Revised
international capital framework available at http://www.bis.org/publ/bcbsca.htm). Under this approach the
home jurisdiction of each bank is responsible for its regulation, but using generally uniform standards. The
minimum capital requirement of banks is set by reference to the value of their assets weighted by risk and
so it was natural for the OECD to seek to adopt risk-weighted assets as the basis of their approach. The
2002 draft on the taxation of banks proposed that the capital allocation approach based on risk-weighted
assets be used for tax purposes and introduced the KERT terminology. In the event it turned out not to be
possible to get consensus on this single approach to capital allocation, but there was consensus on the
KERT functions approach. In the meantime the work on global trading was amalgamated with the PE
transfer pricing work and KERT terminology was included in the third draft in this area which appeared in
2003. Transfer pricing concepts are applied to the financing of firms as well as their activities. As the focus

27
approach linked risk to assets rather than people but the fourth element seemed to
eliminate the inconsistency in that it “located” assets in the country where the risk with
respect to the assets was managed, that is, where the KERT personnel were.61 Inexorably
it seemed that all profit derived from risk and that risk could be identified with a
relatively small number of people in the firm – the KERT people.62 While more
substantive than the freedom of contract approach to risk, the idea that a small group of
people attracted the major share of the profit was quickly seized on by tax advisers and
used in the restructuring of firms by moving such personnel to more attractive tax
climates.

Such restructures were not necessarily tax driven in the sense that the structure of
multinational firms has continued to evolve with business conditions, new technologies
and changing views on management.63 Many manufacturing and other processes have
been shifted to low labor cost countries and firms have increasingly adopted regional
structures and the consolidation of common functions such as treasury and back-office
activities. This process has involved devolution of certain functions from central
headquarters to regional headquarters and the movement of particular functions that were
scattered in different locations to a common, typically regional, location. That is, a new
regional level of management has been created in many firms. While structural evolution
in firms has been occurring for many years, the emergence of the KERT functions
approach in transfer pricing meant that a considerable amount of tax planning has
occurred under the cover of the recent restructures and significantly affected the way they
were implemented.64

here is on the activities of firms, financing issues are left aside as being in a different category. They tie
back in part to the taxation of corporations and their shareholders that was discussed in Section II.
61
This link appeared in the second draft on general principles for attributing profits to PEs (that is, outside
the financial sector as well as within it) which also adopted KERT terminology generally, OECD, note ? at
47.
62
As this work was generalised to all PEs, assets of all kinds, not just financial assets, are now seen as
involving a risk-free rate of return. Virtually all value and profit generated by firms is seen to reside in the
management of risk in relation to their assets and indeed all their activities generally. The generalisation of
the downgrading of assets and the rise of KERT functions (and people) was extended by the OECD work
on taxation of e-commerce which took place contemporaneously in the period 1998-2005. Among other
things this work concluded that the current central place of the PE concept in international tax
arrangements is still justified, “[A]t this stage, e-commerce and other business models resulting from new
communication technologies would not, by themselves, justify a dramatic departure from the current rules”
OECD, E-commerce: Transfer Pricing and Business Profits Taxation (OECD, Paris, 2005) “Part II Treaty
Rules and E-commerce: Taxing business profits in the new economy” at 151. It is true that technology
developments have not meant that corporations can simply abandon their physical presence in countries
and do business remotely, contrary to earlier expectations, but the concept of the multinational firm more
generally of which the PE concept is part is a major structural problem in current transfer pricing rules as
argued above and further developed below.
63
The e-commerce debate was premised on a view that business was now operating in completely different
ways from the past, despite the meltdown of the tech sector in 2000-2001. Restructuring of old economy
businesses to meet the challenge of the new economy was seen as natural and indeed necessary. What was
and is happening is an on-going evolution rather than a revolution and one that is only partly related to e-
commerce and new technology, for example, Sautet note ? at 108ff.
64
In particular it affected the country where regional headquarters were located (countries with favorable
headquarters tax regimes) and the people who were relocated to such headquarters. Tax advisers had to
decide who the KERT people were.

28
It may turn out that the tax planning was premature. The transfer pricing developments in
relation to risk outlined above have not yet been finalized and current signs are that there
will be significant modification of the positions that seemed to be emerging. Rather than
risk being treated as a separable and the chief key to corporate profit, the direction may
well become that risk is an overall attribute of the firm. That is, risk cannot be separately
located but is rather co-located with all the activities of the firm. At this stage it is not
possible to be definitive but some of the signs of such an approach which have broad
ramifications are referred to in the following parts of this section. At the most general
level, the major sign of retreat is the indication by OECD officials that KERT
terminology will be dropped from the general principles for attributing profits to PEs and
confined to the finance sector where it originated.65 What is not so clear is if this is more
a strategic move to abandon unpopular terminology or really means that risk generally
will not be treated separably.

The emphasis on risk in transfer pricing reflects developments in the theory of the firm
highlighting the entrepreneur as the explanation for the existence of firms and as the
generator of value within the firm which cannot be captured by market transactions.
Recall the KERT terminology – “key entrepreneurial risk taking” functions. It is natural
in the light of the growth of the technology and allied sectors to emphasise the
importance of key personnel to the emergence of new firms – Bill Gates and the early
decades of Microsoft et al – but the generalisation to all firms is not so obvious either in
theory or in the way it has been developed in transfer pricing. With regard to theory it
was noted above that the variants on the Coasean view of the firm are not contradictory
and it has been suggested that entrepreneurship is important to the start of the firm but
becomes less important as the firm matures, or becomes dispersed throughout the firm,
which leads to similar conclusions for our purposes.66 Most, but by no means all,
multinational firms are relatively mature. In the transfer pricing area, the KERT
personnel are not identified as the Bill Gates of the world – the makers of the initial
breakthrough or high level firm strategy – but in a general sense mid level personnel on
the ground in management and operations. Why the value generation in the firm should
be specifically located at this mid level has not been clearly articulated and is taken up in
the next part.

In this part of the article I have argued that the growing emphasis on risk in transfer
pricing analysis has been counter-productive as it has facilitated tax avoidance rather than
prevented it. Theories of the firm which stress the entrepreneur seem to lie behind the
focus on risk. The structural problem is not this focus so much as the way it has been
implemented. The assumption of freedom of contract is a particular problem here as it is
more generally and reinforces the conclusion in the immediately preceding part that
contractual approaches to allocation of value within the firm need to be severely
constrained if they are to work appropriately. To the extent that the assumption of

65
Comments of Mary Bennett, Head of OECD Division on Tax Treaties and Transfer Pricing as reported
by Sheppard, “OECD Makes Nice to Americans, Part 3” 113 Tax Notes 37 (2006).
66
Baumol, “Entrepreneurship, Innovation and Growth: The David-Goliath Symbiosis.” 7(2) Journal of
Entrepreneurial Finance and Business Ventures 1 (2002)

29
freedom of contract is not operative (or less so), as in the PE context, the identification of
value creation with risk and of risk in turn with a narrow range of personnel have
provided significant scope for group (re)structures that seek to identify and locate these
personnel in favorable tax locations.

D Personnel
In Coasean theories of the firm, it is the direction of resources that distinguishes the
firm’s activities from the allocation of resources by the market. It is natural to conceive of
this direction in terms of the entrepreneur or managers, which is probably what drives the
thinking behind the KERT functions and personnel in transfer pricing. The consequences
of this thinking have been spelt out above but not the structural flaw involved.

The problem with this approach is that it conceives of the firm in a very primitive form
with one or a few directing the many. The large modern firm is a series of hierarchies
with the direction of employees going far down the structure. The mid level view
reflected in the KERT approach ignores the hierarchies and direction at the top and also
at lower levels of the firm. More importantly as noted above Coase himself regretted that
the idea of directing resources was interpreted (implicitly by himself as well as explicitly
by others) purely in the context of employees. It is the direction of the resources of the
firm, human and otherwise, that is relevant and even low level employees are involved in
that direction of resources of the firm through the use of firm assets in their work. The
quality of work on the factory floor has a significant impact on firm profit.

Another problem with KERT in practice to date, if not in theory, is that it tends to ignore
heterogeneity in firm structure. The structures of firms have evolved over time and at any
given time there is a wide variety of firm structures including very flat and highly
hierarchical management. Trying to identify relatively few very specific personnel in
these structures as responsible for a major share of the profits will not reflect reality for
many firms as well as being very difficult and contentious for the firm and the tax
administration.67

The only practical approach and one that will cohere best with the theory of the firm is to
regard all employees as contributing to the profit that the firm makes. Salary levels reflect
market judgments as to value contributions of particular employees and these are a more
reliable guide to contribution than picking and choosing among employees depending on
exactly what they do in the firm.68 It is not only total salaries of employees that are

67
As tax administrations have started to challenge the claimed tax effects of various firm restructures, the
tax profession has become disenchanted with KERT functions partly because it is very difficult to get a
common interpretation on what it means.
68
Using the cost of all labor inputs does not imply that uniform profitability per dollar of salary need be
assumed. The firm will have accounting records that divide profits on transactions with independent parties
in the market among the various product categories and individual products often at a highly disaggregated
level. The same records will also necessarily have allocated costs, including employee salaries, across the
categories and products. This information can be combined to work out the contribution of labor to the
various elements of the firm’s profits. There are various measurement problems involved but they are not
anything new. Current practice recognizes that it is often necessary to aggregate in transfer pricing analysis,
Transfer Pricing Guidelines, note ? paras 1.42-1.44. Nonetheless there are important areas where full

30
relevant which is why the more generic term personnel has been used in the heading. As
noted above the firm includes dependent entities which are not in the same ownership and
account needs to be taken of such dependent entities’ contributions. Some of these
contributions will often be little distinguishable from salaries of employees and in that
event should be taken into account for the purpose of using salaries if this is a key being
used for the process of allocating the firm’s profits among countries. The same point
applies to other areas and is not repeated, except for sales where it is most important.

There are some presumed transaction rules in this area which at first sight look to be out
of line with the recent development of the idea of KERT functions but in fact provide a
clue to the solution that can be generalized across several areas. In the PE area it was the
case that no profits were to be allocated to internal management activities to the extent
they were provided at a separate location from other activities (if provided at the same
place they would generally be captured by the market pricing of the outputs there when
transferred to other parts of the enterprise).69 The OECD is proposing to change this
presumption so that in future management services will be treated in the same way as for
associated corporations.70 For this case a transfer price is established which generally
provides a profit for the internal management activity, but the rules are relatively
prescriptive so that only a modest part of the overall profit is awarded to the country
where the services are rendered.71 The issue is becoming more important because of the
centralization of internal management functions under restructures of the kind discussed
above.

This presumed approach seems to be in some tension with the KERT functions approach
because in many cases the internal management services may be seen as KERT functions
yet they are only rewarded with a relatively small part of the overall profit.72 In fact it is
thought that another important principle is at play here. The underlying assumption in the
transfer pricing discussion of services above is that profits attributable to labor income
are appropriately allocated to the place where the services are performed. This is also
generally the approach in traditional source rules for services income. Increasingly the
place where services are rendered is much less tied to other production processes than
was the case previously. Moreover individuals can provide services to the same
geographical part of the firm from a variety of locations as they move around the world

international agreement is lacking, most particularly the measurement and treatment of compensation in the
form of equity in the corporate group (stock options etc), OECD, Tax Policy Studies No 11: The Taxation
of Employee Stock Options (2005) ch 4. Entrepreneurs and many managers in particular will be partly
rewarded in this way and it is important if labor cost is being used in the allocation process that this amount
be fully captured.
69
OECD note ? Commentary on article 7 paras 17.5-17.7 at 123-124.
70
OECD, note ?, at 58-59.
71
Corporate groups are effectively allowed to use the PE rule of no profit for the internal management
services if they wish, Transfer Pricing Guidelines, note? para 7.37. The US is currently revising its
regulations in this area in ways which are more prescriptive and do not permit an effective election to use
the PE rule. The proposed changes have got a mixed reception.
72
The way only a modest profit is normally allocated is by calculating the profit allocated to the service
provider on a cost plus basis with a relatively low cost plus percentage. It may be in the circumstances that
a greater share of the reward should be allocated, see the discussion in note ? below.

31
for meetings and other activities related to their work. This is one instance where
communication and transportation technology have a real impact.

There is an alternative source rule (as opposed to a transfer pricing rule) that would
allocate income from technical services and the like to the place where the services are
used but this rule is widely resisted in treaty negotiations.73 This rule partly reflects
general difficulties tax systems have long had with the way in which services merge into
property (especially intellectual property). Technology has given many more services
some of the characteristics of property in that services can be used in a location other than
where they were performed even if they are not regarded as having become property. It is
considered that this source rule is implicitly being adopted for internal management
services on the basis that it is the user (place of use) of the services which takes the risk
with respect to the services. The provider of the services is treated similarly to a contract
manufacturer, in this case appropriately, as the risk of loss (and opportunities for profit)
are in an economic sense more truly located in the part (place) of the corporation that
uses the services, not the part (place) that provides them.

This view of internal management services has wide application as will become apparent
in the discussion of property below. The idea provides a transfer pricing explanation for
allocating profits in such cases between the locations of performance and use and
reconciles the apparently all or nothing conflicting source rules for services (all at the
place of performance or all at the place of use).

In summary of this part the theory of the firm combined with the heterogeneity of firms
suggests that profit allocation relating to labor is best done on the basis of total labor
costs rather than under the KERT functions approach (trying to identify particular
individuals who are regarded as contributing the major share of the firm profit). All the
individuals contribute to the profit and their relative salaries are the most reliable measure
of contribution. As elsewhere, in some cases profit allocation relating to labor can be
done implicitly by market pricing of other (presumed) transactions. If the place of
performance and use of services is different (as with certain internal management
services) it is appropriate to regard the major part of the profit as allocated to the place of
use of the services as the better reflection of the economic position. Transfer pricing rules
for internal services provide this outcome and can be generalized as further developed in
relation to property.

E Assets
The approach that emerged in the finance area regards assets as giving rise to relatively
uniform basic risk free market returns and allocates the rest of the profit to where risk is
managed in relation to the asset. By contrast Coasean theories of the firm often
emphasize asset specificity as one of the reasons for the formation of the firm and the

73
India routinely seeks to insert this rule in its tax treaties, including in its treaty with the US. The US has
for many years been seeking to whittle the rule away, see India US Tax Treaty, article 12, and
Memorandum of Understanding concerning Fees for Included Services, May 15, 1989, OECD note ? at 98-
101.

32
capturing of profits on the assets which are not possible in normal market transactions.74
The result is that many of a firm’s major assets are specific to the firm and produce above
risk free rates of return. This view applies particularly but not only to the intangible assets
of the firm.

In this regard the transfer pricing issue for property has similarities to personnel and
services. Risk should not be regarded as separate from firm specific assets but as
integrated with them. The loan and bet view of the world may be appropriate for many
financial instruments but not for the kind of assets that characterize many multinational
corporations. The move by the OECD to confine the KERT functions approach to the
finance sector is understandable (though its need even there is doubtful for reasons given
above) if it allows a proper appreciation of the role of firm assets in the allocation of
profits between countries. There are some particular signs that risk will not be treated
separably in the case of non-financial assets.

Property nevertheless presents more issues than personnel. As already noted for the
merger of services into assets, there are questions whether profits are appropriately
located where assets are created (if created by the firm as is the case for most intangibles)
or where used. In addition as assets can also be the subject of separate ownership in a
place different from the places of creation or use, there are three possible countries to
which profits related to firm assets can potentially be located. Moreover we can
distinguish several types of assets to which different considerations may apply: assets
used in the productive processes of the firm which may be tangible or intangible and
created by the firm or acquired by it; and assets in which the firm deals which may be
subdivided in a similar way. The discussion starts with assets created by the firm and
used in the productive processes of the firm.

With respect to ownership of such assets, allocation of profit to ownership as distinct to

the place of creation or use of assets raises the now familiar transfer pricing problem of
freedom of contract. If ownership matters, then like contractual assignment of risk in the
case of associated corporations, the location of profit will depend on a stroke of the pen.
In fact we see a close analogue of the contract manufacturer in the area of intangibles for
shifting profits, the contract researcher. If the firm wants to locate profits away from the
place where the work creating an intangible occurs, the corporation in the group doing
the research is treated as a contractor which does not take the risk of loss if the research
fails. If it is wished to spread the profit around other parts of the group whether including
the researcher or not, the relevant corporations in the group will enter into a cost
contribution arrangement under which they share the costs (risks) and benefits of the
research. If it is desired to locate the profit where the work is performed, the researcher
takes all the risk.

74
As in other areas of the theory of the firm, there are several interpretations of the reasons why firms are
necessary to protect and maximize the value of assets, for example, Joskow, “Asset Specificity and the
Structure of Vertical Relationships: Empirical Evidence” in Williamson and Winter note ? at 117, Sautet
note ? at 32-34, 79-81.

33
While contracts of all these types exist in the market, freedom of contract within the firm
makes allocation of the profits arising from the research a matter of election for the firm.
This is another case where freedom of contract needs to be constrained to produce a
robust and meaningful allocation of profit within the firm. A significant part of transfer
price structuring involves the location of ownership of intangibles within the corporate
group. Even if intangibles are originally located where the research is done, they can be
relocated by transfer of ownership within the group. There may be some upfront tax cost
involved in such transfers. Still they figure significantly in current transfer pricing tax
planning and are often associated with business restructures which were discussed above.
The firm seeks to centralize the management of its intellectual property interests
(coincidentally in a tax favorable location) and rolls the transfers into a larger
restructuring project which gives cover for the tax planning. Stripping of intangibles from
higher tax jurisdictions in this way has been nominated by the OECD as one of its current
concerns.75 Accordingly freedom of contract in the ownership of asset area is a
significant structural flaw in current transfer pricing rules.76

This discussion has concerned situations where assets used in production are created by
the firm. In the case of acquisition of assets used in production, similar issues arise.
Profits can potentially be allocated where the asset is acquired, owned or used. In this
case the place of acquisition and ownership will often merge. Moreover, acquisition
activity may be relatively trivial compared to creation situations though much will
depend on the circumstances. The place of use of such assets generally should be the
major factor in allocation of profit; ownership should be generally irrelevant in its own
right. For assets in which the firm deals, the same considerations apply to ownership. The
allocation of profits arising from such assets, whether created or acquired, is postponed to
the later section on sale.

If the place of ownership is (largely) eliminated, the remaining question is allocation of

profit between the place of creation or acquisition and the place of use of assets. Because
current and proposed rules distinguish between tangible and intangible assets, the
following discussion does likewise. The focus is on PE situations as the current rules for
associated corporations accept that ownership matters which is not a sustainable position
for the reasons given above.

After initially leaning to freedom of contract for constructed transactions of PEs so that
the status of an asset used by a PE would be at the choice of the corporation, the OECD is
now going to take the position that assets are generally owned by the PE where they are
used (owned in the sense that the PE will have whatever interest in the asset that the
corporation does). The effect is that if the asset is transferred to the PE from elsewhere in
the corporation, there will effectively be a transfer of the interest to the PE. In other

75
See note ?.
76
The 1986 changes in the US were driven by this problem. Rather than adopting the approach referred to
in the following text, US law sought only in minor ways to constrain freedom of contract – by seeking to
tax outbound transfers of intellectual property fully and allowing adjustments of royalty rates over time
even if the original royalty rate were based on market prices when it was struck. Neither has proved
effective to deal with the restructuring activity.

34
words there will be a presumed sale transaction of the kind that has been discussed earlier
(manufacture at head office, sale by PE, presumed sale by head office to PE). Depending
on relevant national law, gain may be recognized at the other part of the corporation that
transfers the asset physically to the PE on the basis of the market value of the asset.
Again depending on relevant national law the PE will take the asset as owner and
depreciate it or follow any other similar treatment that applies in the PE country.77 If the
asset is created by the corporation, any gain recognized in the transferring country is
likely to be more significant than if the asset were acquired but otherwise this approach
can work equally well for created and acquired assets. It follows from this approach that
if the asset is real estate in the PE country, it is treated as held by the PE always.78

For intangibles, the issues are not so easily resolved. Initially it is important to distinguish
production and marketing intangibles. The discussion here focuses on the former; the
latter are left to the discussion of sales below on the preliminary assumption that they
relate exclusively to the country of sale and so do not raise issues of allocation between
the country of sale and another country. As noted above such property is usually created
by the firm rather than acquired. The position before the recent review of PE taxation was
that a PE could only deduct its appropriate share of royalties paid on intangibles by the
corporation to third parties, that is, no royalties on intellectual property created by the
corporation as opposed to acquired were permitted. This was justified on the basis that,79

In the case of intangible rights, the rules concerning the relations between
enterprises of the same group (e.g. payment of royalties or cost sharing
arrangements) cannot be applied in respect of the relations between parts of the
same enterprise. Indeed, it may be extremely difficult to allocate “ownership” of
the intangible right solely to one part of the enterprise and to argue that this part
of the enterprise should receive royalties from the other parts as if it were an
independent enterprise. Since there is only one legal entity it is not possible to
allocate legal ownership to any particular part of the enterprise and in practical
terms it will often be difficult to allocate the costs of creation exclusively to one
part of the enterprise. It may therefore be preferable for the costs of creation of
intangible rights to be regarded as attributable to all parts of the enterprise which
will make use of them and as incurred on behalf of the various parts of the
enterprise to which they are relevant accordingly. In such circumstances it would
be appropriate to allocate the actual costs of the creation of such intangible rights
between the various parts of the enterprise without any mark-up for profit or
royalty.

77
One of the benefits of this approach is that it is not necessary to rely on constructed transactions
producing a different kind of interest in the asset for the PE from the interest held by the corporation.
Constructed transactions would be particularly difficult in this area as the concept of ownership of tangible
assets for tax purposes is a very nuanced one that would require very precise definition of the constructed
transaction. If the PE takes the same interest as the corporation, the nature of the interest will be sufficiently
defined for the application of national tax rules.
78
Real estate posed a particular problem under the freedom of contract approach originally proposed.
[Technical explanation to be inserted].
79
OECD, note ? Commentary on Article 7 para 17.4. The “rule” about no deduction for internal royalties
was stated in the 1950s but the rationale quoted was only inserted in the Commentary in 1994.

35
The property in effect is treated as an attribute of the corporation as a whole. The non-
rival nature of intellectual property means that it is not possible to treat the property as
owned where it is used in the same way as for tangible property. Rather all parts of the
enterprise that make use of the intangible deduct a share of the costs of creation of the
intangible and are effectively its owners. The presumed transaction is a cost contribution
arrangement with all contributions valued at cost. It is considered that this is the
appropriate outcome with one modification – the research location should be rewarded
with some of the profit as a contract researcher as discussed above for internal
management services.80 The outcome is appropriate as the risk inheres in the intellectual
property and is borne by those parts of the enterprise that use the property (that is, those
parts for which it was created) but the place where the services of creating the intellectual
occurred should have a part of the profit.81 This outcome is also consistent with the
theory of the firm. The firm specific assets are located where they are used in the firm’s
productive processes and services involved in their creation are rewarded.82

The contrast with source rules noted above in relation to services is even more striking
here. It is generally accepted that income from intellectual property is sourced under
traditional source rules where it is used, rather than where it is created yet current transfer
pricing rules for associated corporations which allow freedom of contract locate the profit
where the property is owned. As a result intellectual property is at the centre of
international arguments over division of taxing rights among countries. The solution
offered here is that in a broad sense the traditional source rule is the appropriate approach
for transfer pricing;83 ownership is irrelevant but the place of creation of intellectual
property should share in the profit as well. This approach can be extended to tangible
property. It was noted above that the place of use of tangible property should be treated
as its owner. Based on the discussion of intangible property, the place of acquisition or
creation of the property should be rewarded for that activity if it were performed for
another part of the corporation.

80
If the research location also uses the intellectual property, effectively its contribution to that part of the
intellectual property is valued at cost and it is not rewarded for that part of the services. The discussion in
the text implicitly assumes a steady state firm with the use of the intellectual property by the various parts
of the firm remaining constant over time. The analysis becomes more complicated in the real world where
steady state does not apply. Current transfer pricing rules have to deal with this problem and have
mechanisms for doing so which can be adapted to the approach in the text.
81
It is likely that the reward to the researcher will be greater than the normal cost plus amount that goes to
the contract researcher for a number of reasons but mainly because the researcher itself will often be using
intellectual property of the firm in its research.
82
The position that the OECD will take on this issue is not yet clear. Currently the approach is driven by
KERT functions and for PEs attributes ownership of intangibles to the parts of the enterprise which bear
those functions in relation to the intangible, OECD, note ? at 50-58. The shift from the KERT approach and
the direction adopted for tangible assets discussed above may suggest an ultimate outcome that the place of
use of intangibles will be the main place to which profits are allocated under a cost contribution presumed
transaction.
83
There remains the significant difference that source taxation of royalties is on a gross basis whereas the
transfer pricing rules tax on a more appropriate net basis.

36
The measurement issues for property are very difficult if market prices have to be
determined, particularly for firm specific assets like intellectual property. One benefit of
the approach suggested above is that the problem is largely avoided as cost is generally
used for property. The allocation of profit between the parts of the firm occurs indirectly
through the pricing of products that ultimately are intended for sale to third parties as
discussed in the next part on sales. The main problem area is allocating profits between
the place of creation and the place of use of property but that involves detailed issues that
are beyond the scope of this article. The major point made here is that risk should not be
treated as a separable issue and should be seen as residing in the property and attached to
the place where the property is used.

In summary the major structural flaws of transfer pricing law in the assets area are the
separation of ownership from creation or acquisition and use of assets, and the now
familiar problem of freedom of contract as regards ownership. Risk is part of this story as
ownership in effect allocates the risk attaching to assets. If ownership by contract choice
is eliminated from consideration as it should be, the issue is division of profits between
the place of creation or acquisition and the place of use of assets. The view is expressed
that the place of use should determine ownership/risk and that risk should not be able to
be separated from the place of use. This is consistent with the theory of the firm in the
emphasis on use of firm specific assets as a main source of value in the production
process. The place of creation of assets has a claim to a share of the profits but not on the
basis that that place takes the risk of the creation process. That risk is an attribute of the
firm as a whole.

F Sales
The end point of the firm is the sale of its products into the market. As noted above under
Coasean theories the firm exists to direct resources as it can make profits that are not
available from leaving the allocation of resources to the market. It acquires those
resources from the market and ultimately returns the products produced from directing
those resources into the market through sales. The sale is as much a part of the direction
of the firm’s resources as the production process (as is what accompanies the sale which
will depend on the nature of the sale, for example, after sales service, warranties and
product liability in the case of common consumer products). Following from the two
previous sections, part of the profit in the sense of division of the international tax base
should be allocated to the place or places where the firm’s labor and assets participate in
the sale process.

Because the sale is the end point of the firm it needs to be identified carefully in the
international context in a geographical sense. This takes us back to the boundaries of the
firm discussed previously. It was noted there that the firm is identified geographically by
what was called the dependency attachment. While that attachment includes employees
and assets of the firm, it goes further to include non-employees and other assets which the
firm directs without necessarily owning them in any legal sense, if the necessary
dependency exists. The current PE definition is defective in not being fully based on
dependency. Dependency does apply in the agency context and considerable tax planning
occurs to avoid a legal agency for what is essentially agency in a commercial sense. In

37
this way the firm is kept out of a country for international tax purposes and so the country
is not allocated any profits of the firm. This tax planning is directed at the sale of the
firm’s products into the market and so is a major structural fault in the rules that affect
sales. The fault was noted earlier but is repeated to emphasize its significance for sales.84

The tax planning is necessary because it is in fact difficult for firms not to establish a
presence in a country if they have a significant market there for their products. Even
though it was thought that e-commerce would make such presence less common in
future, that outcome has not occurred significantly to date. Other variations on the earlier
themes are used to move as much of the profit out of the country of sale as possible –
freedom of contract among associated enterprises, shifting of risk by contract, movement
of KERT personnel to other more tax favorable places and ownership of intellectual
property in tax haven corporations. Together these strategies have been used in corporate
restructures of the kinds described previously to remove profits primarily for the country
of sale – in many cases the country of manufacture and sale.85

Hence if the structural flaws in transfer pricing that have already been identified are
fixed, then the jurisdiction of sale will have much of the profit restored to it that has been
stripped out in recent years. The one remaining issue to consider is the class of asset that
for many firms nowadays is the most important – marketing intangibles like trademarks
and goodwill. While these are subject to similar forms of tax planning to try to locate
creation and ownership elsewhere, they are inherently connected to the sales market. The
large part of their creation in a particular market by definition almost must occur in the
country of the market (with advertising there) and that is where they are used. To the
extent that profits are attributable to such intangibles, the profits belong at the place of the
market for the firm’s sales as they are an asset attached to that market.

The way in which firms have dealt with this issue has been implicit rather than explicit.
The incompleteness of contracts within firms allows marketing tangibles to be simply left
out of the analysis of profit allocation if they are not dealt with by contract and not
recognized or properly valued for financial accounting purposes. In cases where
trademarked goods are imported, the transfer price for the goods will often include the
value of the trademark which effectively allocates it to the place of manufacture, not sale.
Yet while the trademark was attached at the point of manufacture, its value is not
generated there so far as the sales market is concerned. In recent years the importance of
marketing intangibles and the circumstances of their creation and use has come to be
recognized in transfer pricing as well as the fact that they are inherently connected to the
market of sale. While freedom of contract, risk and other problems of the kinds noted
above are present in the OECD discussions, the direction of official thinking is similar as

84
As noted earlier, the rules are even more defective on the purchase side as purchasing activities on their
own do not give rise to a PE, and if there is otherwise a PE, no profit is attributable to it based on the
purchasing activity.
85
See Vann, Reflections note ? at 153-157.

38
for assets with the qualification that any difference between the places of creation and of
use is not present to any degree.86

The recognition of marketing intangibles as created and used in the market of sale (that
is, generally where the buyer is) in turn suggest that the boundary of the firm should be
extended to include the market of sale. The firm has assets there whose use is directed by
the firm and so the assets have a dependency attachment to the firm and constitute part of
the firm. In the case of tangible assets it is recognized that a fixed place of business PE
can exist where the assets are used in the firm’s business even if there are no firm
personnel where the asset is located.87 The same principle should be applicable to
intangible assets of a firm which are used in a jurisdiction but this is not possible under
current PE law which requires a physical presence in the jurisdiction in the form of
tangible assets or personnel.

For now this particular limitation on the boundary of the firm is not greatly troubling as it
is in fact difficult for a firm to avoid a physical presence of the kinds discussed above in a
market where it sells its products. Recognition of the importance of marketing intangibles
and allocation of part of an appropriate part of the profit to them will represent a
considerable advance in transfer pricing that will deal with the great majority of cases.88
If, however, the original predictions come to pass that e-commerce will allow significant
sales of a corporation’s products in a market where it has no physical presence, it will be
necessary to consider this extension to current PE rules. There are other more pressing
problems in PE rules that require attention now.

I have discussed at a number of points the relationship of transfer pricing to traditional

source rules. Traditional source rules in relation to sales are quite varied across countries
but they have in common some connection with the market of sale. The analysis above
again can unify transfer pricing and traditional source concepts. Although it has not been
highlighted before, the discussion of the relationship of source and transfer pricing rules
has been in the context of intra-firm dealings, not sales by the firm into the market.
Unlike transfer pricing rules, source rules apply to market transactions whether or not
firms are involved in the transactions. In that sense they are broader in operation than
transfer pricing rules and that is why no final reconciliation of transfer pricing and source
rules is attempted here. Up to and including the point of sale by the firm it is possible in
my view to have consistency of transfer pricing and source rules. This implies that
general consistency is achievable since traditional source rules do not seem to vary
depending on whether the income is intra-firm or not.

86
OECD, note ? at 50-58 represents a considerable advance from my perspective over the Transfer Pricing
Guidelines, note ? at paras 6.36-6.39
87
OECD, note ? Commentary on article 5 paras 8-9, 42.1-42.10. The tendency to downplay the
significance of assets to value is apparent here, however.
88
The recent $3.4 billion transfer pricing settlement between the IRS and GlaxoSmithKline indicates the
magnitude of the issue as the amount concerned profits from selling Zantac in the US market and
significant re-allocation to the place of sale, see Nutt, “Glaxo, IRS settle transfer pricing dispute” 112 Tax
Notes 1020 (2006)

39
The final point in relation to sales concerns the method used to achieve the allocation of
profits to the sales jurisdiction. As the end point is reached with the sale by the firm into
the market, there will be a true market price that can be used in the allocation process and
so sales revenue is the obvious starting point for the allocation. The presumed transaction
that naturally goes with the sale is the purchase of inputs by the part of the firm in the
country of sale from the parts of the firm outside the country. Much of current transfer
pricing tax planning for the sales jurisdiction involves trying to move the starting point of
the allocation away from sales revenue. Although the point is not stressed in exactly this
form above, it underlies the debate around freedom of contract and attribution of profits
to dependent agency PEs. The presumed transactions permitted for allocation of profits in
the sales jurisdiction should require sales revenue as the starting point.

In this part of the article it is suggested that many of the same flaws in transfer pricing
rules identified previously are relevant to sales. Focus on one type of asset that is unique
to sales, marketing intangibles, yields some further important conclusions about the
current rules. First, marketing intangibles are often invisible in the analysis which is
possible because of the incompleteness of contracts within firms. Secondly, the existence
of important firm specific assets in the market of sale suggests that the definition of the
firm needs to be expanded to include that market, though this will only be a pressing
issue if it is possible for firms to make substantial sales in markets without any other
presence there. As sales do involve an actual market price, that price should be the
starting point of the allocation of profits to the sales jurisdiction.

IV Possible solutions
This section first brings together the discussion above of the major structural flaws in
transfer pricing and explores how those flaws may be remedied. It then considers if there
are other rules that can support the specific rules in dealing with tax avoidance arising
from transfer pricing or overcome flaws in transfer pricing rules which cannot be or are
not remedied directly. In keeping with the rest of the article, the solutions are not
provided in detail rather the direction and framework of possible changes are indicated.

A Fixing the transfer pricing rules

The current definition of the firm in transfer pricing rules, particularly the PE rule and
ownership rule for associated corporations, falls short of the dependency attachment
principle that is derived from the Coasean view of the firm. Extending the definition to
include fairly clear dependency situations like many distributors and franchisees which
could be done by requiring only a commercial rather than legal agency would seem to be
feasible as a practical test. Going further to include any country where the firm has sales
(above a certain level) may create enforcement problems and is probably not necessary
while firms continue to have a physical presence in countries where they have significant
sales.

A major structural flaw in current rules is the freedom of contract that is permitted to
associated corporations. Firms are often considered to be characterised by incompleteness
of contracts and this freedom allows firms to fill in the details or not as they desire and
whether or not the details reflect the economic substance of what is occurring. Profit

40
allocation can work through transactions but only if freedom of contract is abandoned
and if the permitted transactions are constrained, or certain types of transactions are
simply presumed. In terms of the Coasean theory of the firm, the idea is the direction of
resources within the firm towards an ultimate sale in the market. The officials who
formulated transfer pricing rules for PEs until recently implicitly took the idea of
direction as applying to the transactions used in transfer pricing analysis. The direction of
supplies of assets or services under intra-firm contracts, whether constrained or
presumed, was to be from inputs to outputs. Although it may seem formalistic, this linear
concept of contract direction was assumed to produce the appropriate allocation of profits
as it would be implicitly rewarding assets and services used at each point in the
production process with part of the profit generated by the firm from its direction of
resources.89

The discussion concerning dependent agency PEs and whether the appropriate contract
for transfer pricing purposes is the contract of agency or a presumed sale of goods by the
principal to the agent is an example of the issue. Much of the tax planning in the transfer
pricing area involves reversing the direction of contracts as is possible with freedom of
contracts generally applied to associated corporations so that they are away or sideways
from the ultimate sale by the firm in the market and allocate the largest part of the profits
backwards out of the country where the part of the firm making the sale is located or off
to the side in a tax favored location.90 While relatively crude and dependent for its
effectiveness on the prices placed on the presumed contracts (as demonstrated in the
discussion of trademarked goods), this restriction on contractual freedom reduces the
scope for tax planning under freedom of contracts and seems a sensible starting point in
using transaction prices as a method for allocating profits.

The next major structural problem identified in the rules was the heightened emphasis
given to risk recently and the further views that risk can be freely assigned by contract
between associated corporations within the firm, that only a small subset of firm
personnel are responsible for risk (the KERT personnel) and that risk is generally
separable from assets. Limitation on freedom of contract can deal with the first issue. The
others require more specific remedies. For personnel the assumption should be that their
salaries reflect their relative contributions to the firm and that no specific group of
personnel should be privileged. To the extent that different products of the firm have
different profit levels, accounting records will be available to allocate profits and salaries
of personnel appropriately, to the extent that this is not done through constrained or
presumed transactions.

For assets the primary allocation of profit should be based on their place of use with the
place of creation of assets, if different, being given usually a lesser reward. Risk should
89
It is possible to trace the idea of the direction of contracts back to the origin of transfer pricing rules, yet
it is hardly ever clearly articulated. The nearest to a clear expression occurred only in 1994.OECD, note?
Commentary on Article 7 at paras 17-17.3. It will be noted from the passage quoted, n ? that Carroll took a
different view. His view was not for freedom of contracts but rather a presumed contract. Nevertheless
reversal of the direction of the contract had the effect (which he wished to achieve) of moving profit out of
the country of sale.
90
See the discussion in Vann, Secret Agent note ?.

41
not generally be treated as separable from the place of use of assets, especially firm
specific assets that are often regarded as one of the hallmarks of firms under Coasean
theories. This approach can be applied to both tangible and intangible assets but is
particularly important to the latter. For marketing intangibles this means that the
appropriate share of profits is allocated to the market of sale. In a transactional setting for
allocating profits among the parts of the firm, these views imply the presumption or
constraint of a complete sale of the asset or services rather than some other form of
provision from one part of the firm to another. This sale presumption also underlies the
linear view of transactions from inputs to outputs discussed above. These various
constraints or presumptions seem to be workable, indeed they already exist in various
areas.

Transactions may not allocate the full profits of the firm, or the construction of
transactions may be thought artificial or difficult in particular cases. In this event it will
be necessary to have recourse to some apportionment methodology. Such a method
already exists in the form of profit splits, either of all the profits, or of the residual after
allowing for partial allocation of profits by transactions. The discussion above which has
been couched in terms of personnel, assets and sales will no doubt be read by some as
simply another form of argument for formulary apportionment of the kind practiced by
the American states. These systems allocate overall profits on the basis of payroll, assets
and sales. In fact it is not such an argument.

Formulary apportionment has become more problematic in practice over the years for a
variety of reasons.91 Because the formula is arbitrary and is not based on any accurate
assessment of the relative contributions to profit for firms generally, let alone specific
firms, there is always a great temptation for states to change the formula when that seems
to be in their favor. Moreover it has not been possible to get any enduring agreement on
the formula. Even if a formula based on the average firm is adopted, the heterogeneity of
and evolution of firms means that it will not be accurate for any of them except by
accident. In turn firms will feel less constrained in manipulating the formula on the basis
that it is essentially unfair to individual firms. The use of sales in the formula potentially
involves double counting. While the sales market should and will be allocated profits to
tax under appropriate transfer pricing rules, this is because of the personnel and assets
there involved in the sale (especially marketing intangibles). Using a fixed formula also
means that measurement of the factors in a jurisdiction becomes critical and there are a
host of measurement problems. In any event it is clear that there is not going to be any
international agreement on simple formulary apportionment.

The profit split is a much more flexible apportionment methodology that tries to reflect
the actual position of the firm. It also allows measurement and manipulation problems to
be dealt with in part by choosing apportionment factors that are relatively robust and
measurable. Moreover if constrained or presumed transactions are maintained as part of
the transfer pricing framework, firms can allocate all or much of the profit by that means.

91
Bankman, Roin and other recent work

42
The outcome in one sense is not far removed from the current transfer pricing rules as it
generally follows the contours of the rules by using transactions and focuses on similar
issues. What is different and important are the constraints or presumptions that are
introduced into the transactional framework. To some they may seem relatively minor
changes but for the transfer pricing specialist they will be recognized as major even
though articulated in a similar framework.

B Are there other solutions for the growth of transfer pricing tax
avoidance?
There is a range of other possible measures to deal with transfer pricing tax avoidance
which are considered in this part, starting with the most specific and moving to the more
general possibilities. One issue that they raise is which country should take the
adjustment action. It was assumed in the previous part of this section that the action was
being taken by the country from which profits are being shifted by transfer pricing. Based
on the earlier discussion of the reasons for taxing corporations, there are two other
countries that could deal with transfer pricing, the country of the headquarters of the
corporate group or the country of the shareholder even if they are not the country directly
affected by the transfer pricing. If it is accepted that the country in which profits are
generated under transfer pricing principles is the appropriate country to levy the corporate
tax, it needs to be asked why the other countries should take action.

Many countries adopt controlled foreign corporation (CFC) rules to deal with deferral of
tax by shifting income to corporations with headquarters in low tax countries. The
regimes are typically targeted at mobile passive income and transfer pricing. As the name
of the regime suggests, it applies to situations of control of corporations in other (usually
low tax) countries. Typically, the most significant application of the regimes will be to
corporate groups, rather than corporations controlled by individuals. The regimes operate
in most cases by taxing the income in the “residence” country of the parent corporation of
the corporate group (that is, its headquarters). Even if the shareholder is resident in the
same country, it is not clear on the basis of the discussion in Section II why that country
should make transfer pricing its concern unless the transfer pricing is out of that country
to another country. The application of CFC rules to transfer pricing is often limited to
transfer pricing out of the country applying the CFC rules..

One problem with CFC rules is that they may have multiple operations if corporate
groups consist of more than one tier of subsidiary as they typically do. For example, if a
listed US corporation has a subsidiary in Australia which in turn has a subsidiary in a tax
haven deriving income targeted by CFC regimes, the CFC rules of both the US and
Australia may be engaged. In the case of transfer pricing, multiple applications will not
be a problem if the CFC rules are limited in each case to transfer pricing out of that
country. On the other hand this limitation may reduce the efficacy of CFC rules. It may
be clear when profits have been shifted to a tax haven by transfer pricing that another
country is not collecting its appropriate share of tax, but it may be less clear which
country has suffered from the transfer pricing.

43
It has recently been suggested that countries with CFC regimes (typically OECD
countries with significant numbers of multinational firms based there) should coordinate
their regimes so that the CFC regime of only one of them would apply in a given case in
a uniform way.92 Even though such overlap may not be an issue for transfer pricing if the
CFC rules of each country are limited to transfer pricing out of that country, the idea
could usefully be extended to cover transfer pricing more generally. Each country with a
CFC regime would become the enforcer of transfer pricing out of all countries in the
coordinated regime, rewarded by the revenue from enforcing transfer pricing rules more
generally. The justification for such an extension would be on a knock for knock basis,
that is, each country would on average collect at least the same amount of revenue if it
limited its CFC regime to transfer pricing only out of that country. There is some
voluntary coordination of CFC regimes at the moment but nothing more concrete than
that.

In the discussion of cause and effect in relation to revenue concerns about erosion of the
international business tax base and problems in transfer pricing, there is often a post hoc
ergo propter hoc assumption, that is, the rule being studied is defective in some way and
because of that defect it is abused by taxpayers to lower revenue. The cure obviously is to
fix the rule. So far in this article this has been the approach – to describe the problems
that exist in transfer pricing rules in the context of the international tax system, why they
have led to abuse and what the possible fixes are, including CFC fixes. The assumption
and prescription are problematic, however.

If we lift our gaze from the individual trees for a moment to look at the forest, we should
note that the US tax system has been plagued in recent years by corporate tax shelters,93
in which the US system is not alone.94 Although the shelters often involve international
tax issues in the plan to reduce tax, the use of such shelters applies across the whole
spectrum of the system, domestic and international, and indeed may be thought mainly to
apply to domestic income. It is true that corporate tax shelters often are devised after the
event to eliminate large taxable gains that have already arisen or are reasonably certain
and have one-off effects, whereas transfer pricing tax planning is applied in an on-going
way to reduce tax on future income that it is hoped but not certain will arise. Nonetheless,
it seems plausible to assume that if multinationals are prepared to enter into one-off
corporate tax shelters, they are also likely to plan in a systematic way to reduce tax on
future income.

The implication is that while fixing the particular rules is one possible strategy to deal
with transfer pricing as well as other abuses, other possible rule based strategies include
the use of business purpose and economic substance judicial tests and/or the adoption of a

92
Burnett, “Replacing CFC regimes with a collective attribution system” 38 Tax Notes International 1109
(2005).
93
Bankman, “The New Market in Corporate Tax Shelters” 83 Tax Notes 1775-1795 (1999).
94
In 2005 Australia, Canada, the UK and the US established the Joint International Tax Shelter Information
Centre for the sharing of information on the latest tax shelter activity, see http://www.irs.gov/pub/irs-
utl/jitsic-finalmou.pdf.

44
legislative general anti-avoidance rule.95 It was noted above that transfer pricing rules in
the past were regarded as anti-avoidance measures, though in recent years they have
become systemic rules. Form over substance and similar rules could be applied to transfer
pricing avoidance strategies where nothing of economic substance happens such as risk
shifting by contract within the corporate group. In many cases, however, there is
economic substance. As noted above corporate restructures often have commercial
purposes as well as tax purposes. In that event the application of general anti-avoidance
rules becomes more problematic.

As with CFC rules there are also issues of which country would apply its anti-avoidance
rules. Generally the attitude of most countries is that its anti-avoidance rules only apply
to avoidance of its own tax. Coordinated action of the kind discussed for CFC rules
would be necessary if application of anti-avoidance rules by other countries were to be a
possibility. This outcome seems even more unlikely than coordination on the CFC front.

Beyond substantive tax rules, it is also possible to address the incentives to enter into tax
motivated transactions through tax administration related measures such as increased
auditing, higher penalties and greater information requirements (as to both record
creation and reporting). In recent years many countries have taken action of this kind
directed at transfer pricing in particular. The most important of these measures probably
are requirements of contemporary documentation of transfer pricing practices.
Nonetheless it may be thought that these rules have increased compliance costs without
having much impact on tax avoidance through transfer pricing, if we are to judge by
claims of continuing tax avoidance with which the article started. Further there are issues
of the kind already noted of which country requires which documentation given that the
countries are concerned with their own transfer pricing issues, not the problems of other
countries, and of international coordination. In contrast to the CFC and anti-avoidance
areas, international cooperation in this area is well advanced so that corporate groups can
use common information and common formats of information to satisfy the
documentation requirements of several countries.96

Continuing the metaphor of trees and forests, if we look beyond the tax forest to
corporate behavior more generally, there seems to be a fairly clear link between the
recent problems in corporate governance and the growth in corporate tax shelters.97 If a
firm is reporting fake profits for financial and corporate law purposes, it certainly does

95
In 2003 the OECD took the controversial position that tax treaties are subject to domestic anti-avoidance
doctrines, see OECD, note ? Commentary on article 1 paras 7-26 at 53-64 which have attracted
considerable debate, for example, Ward et al, The Interpretation of Income Tax Treaties with Particular
Reference to the Commentaries on the OECD Model (Amsterdam, IBFD, 2005) at 78-92. Whether
domestic anti-abuse rules can be used in treaty cases is an issue that will take some time to reach
finalization in many countries.
96
Pacific Association of Tax Administrators (PATA – Australia, Canada, Japan and the United States)
Transfer Pricing Documentation Package available at
http://www.irs.gov/businesses/international/article/0,,id=156266,00.html, EU Code of Conduct on transfer
pricing documentation available at
http://ec.europa.eu/taxation_customs/taxation/company_tax/transfer_pricing/forum/index_en.htm.
97
[Enron Congress reports]

45
not want to pay tax on them. And if the managers of the firm are rewarded by reference
to the (after-tax) profits of the firm, they certainly have an incentive to reduce the tax
paid by the firm which may encourage aggressive tax avoidance, including transfer
pricing. Perhaps the remedy lies in corporate law either generally or in some tax-specific
elements of corporate law (such as special corporate procedures or reporting in relation to
corporate tax risk). There is some evidence of tax administrations taking this approach98
but whether corporations are becoming more sensitized to the issue and more cautions as
a result is unclear.

The issue once more is which country will apply the necessary corporate governance
measures. In practice corporate governance strictures will be at their greatest at the level
of the parent, usually listed, multinational corporation.99 In this case risky behavior by
firms from a corporate governance perspective would seem to be a concern even if it
involves tax avoidance in another country. If corporate governance rules of countries take
what is a selfless view from the tax perspective, then corporate governance rules may be
an additional (and potentially more effective) international enforcement mechanism of
transfer pricing rules.

Clearly it is important to keep these other mechanisms in mind when considering the
ways in which transfer pricing can be counteracted. Generally, however, they would
seem to provide additional rather than substitute protections against transfer pricing
abuses. In terms of which country should take action to prevent abuses, these other areas
suggest that it may be possible for countries to take action and so cooperate in transfer
pricing enforcement even though they do not have a direct revenue interest at stake.

In conclusion, in terms of own tax revenue it is important to recall that not all countries
may have the same interests. It is a fundamental policy assumption of the levy of
corporate tax in a country other than that of the shareholder that the tax is not shifted. It
has been suggested that developing countries in particular are wary of enforcing transfer
pricing tax rules because of a fear that the tax will simply be shifted to immobile factors
in the developing country.100 This may make the countries less interested in and thus
likely to cooperate effectively with developed countries in dealing with transfer pricing.
For this reason it is in the interest of developed countries to resolve a constant source of
tension with developing countries, the nature and extent of source country taxation on a
traditional source basis. This article suggests, without finally resolving, that a revision of

98
In 2005 the Commissioner of Taxation in Australia wrote to the chairs of the boards of major Australian
corporations about tax risk and the Australian Taxation Office since has been very active in linking tax
compliance and corporate governance, see
http://www.ato.gov.au/corporate/content.asp?doc=/content/56224.htm. The issue recently featured at the
OECD Forum on Tax Administration in September 2006, including remarks for the IRS Commissioner.
99
Clark, “Corporate Governance Changes in the Wake of the Sarbanes-Oxley Act: A Morality Tale for
Policymakers Too” Georgia State Law Review (2006) forthcoming sets out the various levels at which
corporate governance rules operate so that for a listed corporation in the US the total package is quite
formidable.
100
Baistrocchi, “The Structure of the Asymmetric Tax Treaty Network: Theory and Implications” paper
presented to conference on Tax and Development, University of Michigan, November 2006 at 15.

46
transfer pricing rules as proposed above may go a considerable way to resolving the
differences between transfer pricing and source taxation principles.

47
Appendix: Integration of corporate and shareholder taxation
Part I Relative tax at corporate and shareholder level under
integration
For purposes of illustration of various points we will take two kinds of integration
systems as representative: firstly, a dividend tax credit system under which corporate tax
is attached to distributions from corporations and credited to shareholders (much like
withholding from wages), and secondly a low flat rate tax on dividends. Under the
imputation system capital gains on shares are taxable in full while under the low flat rate
system the same tax rate applies to capital gains on corporate shares. These capital gains
arrangements are necessary for producing equivalent treatment under each system for
retentions of corporate income which are reflected in gains on sales of shares. The
assumed corporate tax rate is 30%, the (maximum) individual tax rate is 40%, and in the
case of the low flat rate system the tax rate on dividends and capital gains is 15%. The
first system has a resemblance to Australia and the second to the US. Nonetheless most
countries’ integration systems can be aligned with these prototypes, for example, the first
can be thought to represent broadly Canada and the UK, and the second broadly France
and Germany.

Table 1
System Imputation Low flat rate
Distribution Retention/sale Distribution Retention/sale
Corp income 100 100 100 100
Corp tax 30 30 30 30
Corp after tax 70 70 70 70
Shareholder 100 (70 distrib 100 (price cap- 70 70
income/gain + 30 tax credit) tures tax credit)
Shareholder tax 10 (40 at 40% 10 (40 at 40% 10.5 (70 at 10.5 (70 at
less 30 credit) but 30 tax 15%) 15%)
credit available
to buyer)
Shareholder net 60 60 59.5 59.5

The main points of this example are twofold. The income derived through the company
gets taxed in the end result at the shareholder’s ordinary tax rate. In the case of
imputation the correspondence is exact while for the low flat rate it is close. The latter
result is simply an artifact of the tax rates chosen but nonetheless is typical of rates found
in systems that try to achieve rough and ready full integration for individual maximum
tax rate shareholders. Further, most of the tax in each case is collected at the corporate
level. Again this is a result of the relative corporate and individual tax rates but again the
rates are broadly representative of real world rate structures in OECD countries.

Assuming that there are substantive reasons to levy the corporate tax in the country where
the corporation is located and to integrate that corporate tax with taxation of the

48
shareholder in the country of the shareholder, the calculations in Table 1 are still
practicable but the result is that the bulk of the total tax goes to the country of the
corporation, not the country of the shareholder. It is easiest in an administrative sense to
achieve this result by the low tax rate method in Table 1 rather than imputation, as indeed
the US system does; the low tax rate of 15% applies to dividends from foreign
corporations as well as domestic corporations (and the low capital gains tax rate applies
to gains on shares in foreign corporations).

Imputation is much more difficult to operate in this context as it generally tries to tax
distributed corporate income at the actual marginal tax rate of the shareholder which
requires refund of the corporate tax if the shareholder’s tax rate is below the corporate
rate. The country of the shareholder will not wish to refund foreign corporate tax.
Mechanisms to deal with this problem are complicated but countries were experimenting
until the European Court of Justice unfortunately and needlessly held that imputation was
contrary to EU non-discrimination norms and killed the system off in Europe where it
had begun.101 The ripples from Europe spread and only relatively few countries like
Australia maintain the system nowadays. In its place have sprung up systems which are
more approximate in their relief of double taxation, like the US system.

Part II Integration where income derived by permanent establishment

of corporation
If the corporation derives the income from a PE in a third country, then the result in the
shareholder country depends on a number of factors of which the method of relief for
double taxation in the country of the corporation is relatively minor. The following table
illustrates these results. It assumes a corporate tax rate in the PE country of 25% and
considers a foreign tax credit system in the country of the corporation and an exemption
of branch income in that country. For the imputation system the alternatives of giving tax
credits or not in the country of the shareholder for the foreign corporate tax mean there
are four permutations.

101
Graetz and Warren. “Income Tax Discrimination and the Political and Economic Integration of Europe”
115 Yale Law Journal 1186 (2006).

49
Table 2
System Imputation Low flat rate
Imputation credits for No imputation credits
foreign tax for foreign tax
Relief of Foreign Foreign Foreign Foreign Foreign Foreign
foreign tax tax credit exemption tax credit exemption tax credit exemption
PE income 100 100 100 100 100
PE tax 25 25 25 25 25 25
PE net 75 75 75 75 75 75
Head office 5 (30 less 0 5 0 5 0
tax 25 credit)
Head office 70 75 70 75 70 75
net
Shareholder 100 (gross 100 (gross 75 (gross 75 (no 70 75
dividend up foreign up foreign up gross up)
income and tax) domestic
domestic tax)
tax)
Shareholder 10 (40 15 (40 less 25 (30 30 10.5 (70 11.25 (75
tax less 30) 25) less 5) at 15%) at 15%)
Shareholder 60 60 45 45 59.5 63.75
net

Under an imputation system the result depends entirely on whether the imputation system
in the shareholder’s country grants any relief for foreign tax and is independent of the
method of relief of international double taxation in the country of the corporation. If the
shareholder country grants relief for foreign tax in the same way as for domestic
corporate tax, then the result is the same as in the domestic case. If the shareholder
country does not grant imputation relief for foreign tax, then the outcome in the
shareholder country is the equivalent of giving the shareholder a deduction for the foreign
tax – the shareholder is taxed on the net income after foreign tax (the head office net in
the table).

For the low rate system, the method of relief in the country of the corporation has some
impact but the impact is relatively minor if the difference between the corporate tax rates
in the country of the PE and the country of the corporation is not significant. If the
country uses a foreign tax credit, the ultimate outcome for the shareholder depends on the
corporate tax rate in the country of the corporation. If the country of the corporation uses
the foreign income exemption, the result for the shareholder depends on the corporate tax
rate in the PE country. If the corporate tax rates of the three countries are similar, the
variations in outcome are small whichever country levies the tax.

50