Jump to Content Jump to Main Navigation
Transfer Pricing and the Arm's Length Principle After BEPS by Collier, Richard; Andrus, Joseph L (17th August 2017)

1 The Emergence of the ALP

Richard S Collier, Joseph L Andrus

From: Transfer Pricing and the Arm's Length Principle After BEPS

Richard Collier, Joseph L Andrus

Formation of contract — Interpretation of contract — Payment of price

(p. 6) The Emergence of the ALP

The tax system of the various countries has, indeed, been created piece by piece according to the needs of the day and along the lines which appeared to offer the least difficulty from the purely administrative point of view and which were, at the same time, the least likely to cause conflict with public opinion and accepted ideas. Contemporary tax systems are therefore the outcome of historical developments, economic and social considerations and administrative possibilities.

League of Nations Fiscal Committee, Report to the Council on the Work of the Eighth Session of the Committee

A.  Introduction

1.01  The current international tax system was developed through a process of international collaboration which began in the early 1920s. In the course of this process the ALP emerged as an international standard. Initially, the principal focus was on developing an approach to the allocation of taxing jurisdiction between states to remove, or at least materially reduce, the incidence of double taxation. Such double taxation involved the simultaneous taxation (p. 7) of the same income or profits by at least two states (as where, for example, a company headquartered or ‘resident’ in one state carries on business through a branch in another state and both states seek to tax the income arising from the branch operations). The task of dealing with the double taxation issue involved reconciling the competing claims of states to levy taxation. That reconciliation would inevitably limit some taxing rights and yet be broadly acceptable to all states. The early discussion was dominated by whether, in an effort to prevent double taxation, taxing rights should be allocated only to the headquarters state (the territory of the ‘domicile’ or ‘residence’ of the taxpayer concerned) or whether, and to what extent, taxing rights might also be granted to the ‘source’ state in which specific items of income might arise or be generated.1

1.02  Notwithstanding some early theoretical arguments in favour of an exclusively residence-based approach, it was always recognized that some level of source taxation was inevitable. This was primarily due to the widespread operation of source-based taxing regimes already in place at the time. The exercise therefore became a pragmatic attempt to develop a solution grounded on a synthesis of the many existing tax systems of individual states, virtually all of which applied ‘source’ or ‘origin’ taxation to some extent.

1.03  Ultimately the exercise evolved into a discussion of the appropriate limits of source state taxing rights. This rapidly led to broad agreement between the participant states that general business profits (usually referred to in the discussion as industrial and commercial profits) should be taxed in the state where they were ‘realized’ or ‘produced’.

1.04  This in turn led to the question of how the amount of income attributable to a branch by reason of being ‘realized’, or ‘produced’ by the branch should be determined. The goal was to develop a generally accepted approach so that overlapping source state claims (which would otherwise be a further cause of double taxation) would be avoided. There were also questions relating to the form of business presence in a state that would trigger source state taxing rights and how subsidiary companies should be treated for tax purposes.

1.05  The early discussion was dominated by the treatment of branch operations. The approach to allocation ultimately agreed was based on the ‘independent enterprise’ (also referred to as ‘separate accounting’) theory, which was essentially the equivalent of applying the ALP to branch operations. As the discussion of branches proceeded, the position of subsidiary companies was introduced, with initial uncertainty as to whether or not subsidiaries should be taxed in their own right as separate taxpayers or whether they should be treated as mere extensions of the parent company and so taxed in the same way as branches. It was determined that subsidiaries should be taxed in their own right as distinct taxpayers and the ALP was introduced to ensure that, for tax purposes, there was no diversion of profits resulting from any manipulation of the financial relations between parent and subsidiary companies.

(p. 8) 1.06  The following discussion will consider the main lines of the early discussion as it developed a consensus solution to the double taxation issue, in the process charting the emergence of thinking on the division of profits and the ALP.

B.  Developing the Founding Principles of International Taxation

1.07  Given the significance of the formative international tax developments in the period following the First World War, it is appropriate to start by considering why these developments occurred at that particular time.

1.08  There are three main factors that, in combination, explain the significantly increased focus in this period on the problems of double taxation. These are, first, the expansion of international trade that had taken place in the early years of the twentieth century, coupled with the recognition that a post-war revival of international commerce would be vital to the reconstruction of European economies after the First World War; second, the sharply increased rates of tax that were being levied by states after the First World War, with the result that double taxation had rapidly become significantly more costly to business; and, third, the engagement of the newly created international bodies whose mission it was to deal with any obstacles in the way of the post-war reconstruction process and which built, for the first time, a level of international consensus on an appropriate response to the problems of double taxation. These three factors are discussed briefly below.

1.  The Increase of International Trading Volumes

1.09  The latter decades of the nineteenth century had seen significant advances in the level of world trade. The period leading up to the First World War saw further significant growth in global economic integration, with world trade going from $8 billion in 1896 to $18 billion in 1913 (even corrected for inflation, this was nearly a doubling in a less than twenty-year period):2

The opening years of the twentieth century were the closest thing the world had ever seen to a free world market for goods, capital, and labour. It would be 100 years before the world returned to that level of globalisation. In addition, this integrated international economy grew at its most rapid rate in recorded history. Output and incomes rose, and not just in rich nations: many relatively underdeveloped countries also grew dramatically.3

1.10  As Keynes observed, ‘social and economic life’ had experienced an ‘internationalisation … which was nearly complete in practice’.4 Within a few years, the damage caused by the First World War inflicted a major economic reversal,5 and led to the need for a major rebuilding of Europe with organizations like the League of Nations and the International Chamber of Commerce (discussed below) playing an important role in the process.

(p. 9) 2.  Sharp Rise in Tax Rates and the Cost of Double Taxation

1.11  The First World War and its aftermath led to a massive expansion of government spending, and in consequence, to significant rises in the rates of taxation levied. The income tax rate in the UK rose from 5.8 per cent (plus 2.5 per cent super tax) in 1911–12 to a combined rate of 52.5 per cent by 1918–19.6 Rates also rose sharply in many other countries, with the highest rates of income tax including Australia (40.625 per cent); Canada (72.9 per cent); Japan (30 per cent); and the US (77 per cent).7 Many countries also introduced a progressive (i.e. graduated) income tax in the period 1900–20, including the US in 1913, and France in 1914.

1.12  The general rise in the level of taxation was a major concern to business. In 1923, the International Chamber of Commerce (ICC) commented: ‘Taxes have been raised everywhere to a level which would have seemed excessive before the war … [Countries] have all pushed fiscal pressure to its extreme limit’.8 Business concerns about sharply rising tax rates were compounded by the greater incidence of double or multiple taxation of the same income.9 This was caused by the uncoordinated (and so overlapping) taxation of cross-border business ventures. The ICC stated:

The cause of the trouble is not difficult to discover. It arises because each state creates its fiscal legislation to please itself, without asking if the rules which it is laying down correspond to those in practice in other countries.10

3.  The Rise and Engagement of International Organizations

1.13  Two organizations played a material role in the post-First World War discussions on tax, namely the ICC and the League of Nations, with the latter ultimately playing the key role. Both organizations came into being after the end of the war.

1.14  A forerunner of the United Nations, the League of Nations was an intergovernmental organization that came into being in 1920. Its role was to act as a forum for resolving international disputes by fostering cooperation between states and by providing security for its members.11

(p. 10) 1.15  To address the dire economic and financial conditions in Europe after the First World War, the League established an Economic and Financial Organization (EFO), the world’s first intergovernmental organization dedicated to the promotion of economic and monetary cooperation. The EFO subsequently organized a number of specialist committees (Financial, Economic, Fiscal, Statistical).12 The work of the Fiscal Committee is discussed below.

1.16  The founding of the international tax system is often ascribed exclusively to the work of the League of Nations. However, the ICC made a significant early contribution to efforts to deal with the problem of double taxation.

1.17  The ICC was also formed in 1920 following an initiative in 1919. The ICC was formed as a non-political organization of business interests in the various affiliated countries.13 The immediate objective of the ICC was to facilitate business participation in the reconstruction after the First World War. The ICC ultimately became a permanent organization focused on facilitating cross-border commerce.

4.  ICC Concerns on Double Taxation

1.18  Although the League of Nations and the ICC came into being at a roughly similar time, it was the ICC that initially took the lead on efforts to address double taxation. The ICC prioritized the problem of double taxation from its inception.14 In fact, even before the ICC had formally been created, the issue was raised at the ICC’s 1920 Paris organization conference. Countries were urged at that conference to take prompt action to deal with the issue.15 The Paris meeting also agreed on the formation of an ICC Committee on Double Taxation which was formed to consider the problem and present its views to the First Congress of the ICC, to take place in London a year later.16

1.19  By the time of that London Congress of June 1921, work on the subject had been carried out by the ICC Committee on Double Taxation and country reports had also been compiled and sent to the ICC. The Committee presented a report, ‘Double Taxation’, to the 1921 Congress.17 This included a lengthier and more strongly worded draft resolution on (p. 11) double taxation together with some proposed principles, which were intended to form the basis of a response to the matter.

1.20  As already noted, the early discussion on double taxation was dominated by the issue of whether, in an effort to provide a solution for dealing with double taxation, taxing rights should be allocated only to the headquarters state (the territory of the ‘domicile’ or ‘residence’ of the taxpayer concerned) or whether, and to what extent, taxing rights might also be granted to the ‘source’ state in which specific items of income might arise or be generated. The principles developed by the ICC sought to address this question by recognizing source state taxing rights. Specifically, one of the principles suggested by the ICC Committee called for rebates of tax levied in a country where that tax had been levied on income that had been ‘earned abroad’ and therefore already subjected to tax abroad. This principle was subsequently adopted by the 1921 ICC Congress.18 The draft resolution referred to the heavy burden of double tax on international trade and stated that governments ‘should be pressed to come to an understanding with a view to alleviating this burden’.19 At this early stage, the discussion of the ICC Committee on Double Taxation was concerned with the general issues of double taxation and how they might be dealt with. The ICC clearly contemplated that any solution to the problem would involve some level of taxation being levied in source countries.20

C.  Early Work of the ICC on Income Allocation

1.21  The discussion that took place at the 1921 Congress included an early recognition of the necessity of determining allocation rules for tax purposes. This followed from the acknowledgement that some source taxing rights would be part of any solution to deal with double taxation and from the fact that businesses would often be carried on through two or more establishments in different countries. The complexities of cross-border business were also acknowledged:

[C]‌onsideration must be given to methods of determining where income is earned. Administrative detail is here of paramount importance, since it is on this point that the legislative and administrative authorities of the different countries are likely to disagree. If a British corporation cuts and partly fashions timber in the United States, consigns it to its own factories in Scotland, and sells the product very largely from sales offices located in Holland, in what proportion is the final net income earned in the three jurisdictions involved? This is the fundamental problem, and general principles recommended to legislative bodies may fall on deaf ears unless specific proposals can be agreed upon.21

(p. 12) 1.22  The Committee on Double Taxation met again in Paris in December 1922 and adopted certain guiding principles including the requirement (consistent with its earlier position) for countries to give relief for tax paid on income ‘not acquired in its territory’. The guiding principles included some suggested rules as to the classification of various types of income, together with a direction as to where the relevant income was to be regarded as so ‘acquired’ (i.e. those rules constituted what we might refer to today as ‘source rules’). These rules included, in the case of income derived from a business of any kind, the approach of identifying ‘the country in which the business is carried on’. The Double Taxation report prepared for the second ICC Congress held in Rome in March 192322 (which was based on the work of the Committee on Double Taxation) again called for a response to the double taxation issue and suggested that the difficulty would largely be solved if the same income were taxed only once. The report went on to consider how this approach might be delivered and identified as a ‘widespread view’ an approach under which a domicile (residence) basis is applied, subject to a country having taxing rights over income ‘derived’ in that country. The report noted that such an approach would require a deduction in the state of residence for the foreign tax already paid to a foreign country on income derived in that foreign country.23

1.23  As in the earlier discussions, it was recognized that the obvious requirement of this approach is the need for a common understanding of ‘how the amount of income supposed to have been gained in a given territory shall be estimated’. The response proposed by the Committee on Double Taxation recognized the significant difficulties that could arise in addressing this question but sought to address them by proposing various principles, which, it suggested, could serve as the basis for a more complete set of international regulations.24 The principles relating to taxes on income included whole or partial relief for double taxation. There were also four principles that related to the allocation of profits. These may be summarized as: (i) income from a business of any kind should be regarded as ‘acquired’ (i.e. sourced) in the country in which the business is carried on; (ii) if a business is being carried on in more than one country, the profits derived by the business should be taxed in each country in proportion to the profits realized therein; (iii) if it is impossible to prove the exact amount of profits accruing in each country, the portions pertaining to each country should be determined according to rules to be established by the country concerned. A common approach to identifying relevant portions—e.g. by reference to turnover—was contemplated; and (iv) the portions of income taxed in the different countries should never exceed the total fixed by the competent authority in the state of domicile (residence).25

(p. 13) 1.24  The discussion at this early stage is fairly rudimentary but it is worth noting that principle (ii) above would have involved some form of profit split or ‘formulary’ approach in cases where business was carried on in two or more countries. As will be seen below, the approach ultimately adopted for allocating income to branches (though several years later) was based on the ‘separate enterprise’ approach, which is essentially the application of the ALP for dealings between branches. The two methods could lead to quite different results because, where a profit split methodology is used, all those participating in the methodology will share in the total profits arising, whereas under a ‘separate enterprise’ approach they may not.

1.25  The discussion in the 1923 ICC Double Taxation report also sets out the objective of the ICC work on the allocation of income for tax purposes. The ICC sought to have a certain number of countries, under the auspices of the League of Nations, undertake to accept principles for the allocation of taxing rights. These principles would be implemented by means of a set of working regulations, such as those proposed by the ICC, which the signatory countries would be obliged to incorporate in their national laws. It was noted that such principles could be delivered through bilateral tax agreements, and the discussion also contemplates the possibility of a ‘collective’ (i.e. multilateral) agreement.26

D.  Revised ICC Proposals of 1924 on Double Taxation

1.26  The four allocation principles discussed above were proposed for adoption at the ICC Rome Congress but, because they recognized source-based taxation, they were highly controversial. As a result, they were not formally adopted and the Committee on Double Taxation subsequently decided it would be too difficult and possibly premature to pursue them further.27 Instead, the Committee sought to confine itself to the formulation of general principles intended to lead to the complete abolition of double taxation. This led ultimately to the production of revised resolutions in March 1924, which were adopted by the Council of the ICC.28 These revised resolutions again expressed very strong concerns about the impact of double taxation:

[T]‌he present state of fiscal legislation in different countries imposes upon international commerce and industry taxes which individually are already high and, taken together, constitute burdens so heavy that they paralyse the development of international trade and consequently threaten to endanger the prosperity of Nations.29

(p. 14) 1.27  In line with the tack taken by the ICC from the beginning of its discussion of double taxation, the resolutions reaffirmed a preference for a residence basis approach but recognized that as a practical matter ‘origin of income’ (source) taxation will be applied by states. In that regard, the resolutions state that where origin taxation cannot be avoided, it should be restricted to ‘income accrued within [the origin, or source] territory’.30 Further, in cases where origin tax was levied the resolutions call for the state of residence to grant relief for the tax paid to the origin state.31 The discussions of the Committee contemplated the conclusion of some form of multilateral convention on the basis of these principles and, pending such a development, called for more bilateral agreements between states to reflect these principles.32

1.28  The League of Nations had increasingly been taking an interest in the topic of double taxation (in part, as a result of the promptings of the ICC and especially the Committee on Double Taxation with which it had been in regular contact).33 From around March 1924, when a delegation from the ICC went to work with the League’s experts who were then in session discussing these same problems,34 the League became the dominant forum through which discussion of the topic and actions to address it were pursued.

1.29  Before turning to the work of the League of Nations, it is worth pausing to note the contribution made by the ICC to this point. The double tax issue had been at the top of the ICC agenda from 1919. Running through the work of the ICC in this period was the recognition that the preferred residence basis would, for practical reasons if not otherwise, need to operate in tandem with some form of ‘origin’ or source taxation, which would necessitate some reduction of tax in the residence country. There was also an attempt to classify income into various categories, with accompanying source rules. This in turn led to early attempts to identify the scope or limits of that origin taxation on business income, (p. 15) with an attempt (though ultimately unsuccessful) to develop more detailed rules on how to address the problem of allocating income between states where a business was conducting its operations in two or more states. The work of the ICC also indicated that any solution developed to address the double taxation issue would, to be workable, need to be as good a fit as possible with existing tax systems (which explains the recurrent efforts by the ICC to gather information on the tax systems operated by states). The goal was to work towards a ‘general convention’ or multilateral treaty involving as many states as possible. The work of the ICC clearly paved the way for the major developments that were orchestrated subsequently by the League of Nations, including the League’s subsequent work on rules for the allocation of profits.

E.  The First Involvement of the League of Nations

1.  The 1923 Report of the Four Economists

1.30  In relation to the topic of double taxation, the first act of the League of Nations was to refer the study of the question to its Financial Committee which in turn (in 1921) asked four eminent economists (Professor Bruins (of the Commercial University of Rotterdam), Professor Senator Einaudi (of Turin University), Professor E. R. A. Seligman (of Columbia University), and Sir Josiah Stamp (of London University)) to prepare a report dealing with double taxation.35

1.31  The terms of reference of the Report call for an assessment of the economic consequences of double taxation and for general principles, if they can be identified, that might be applied in a multilateral treaty on the topic, as well as a consideration of purely bilateral and domestic law responses. The Report responds to the brief primarily from a theoretical perspective, which is clearly a different approach to that taken by the ICC work that preceded it. A second important difference to the preceding work is the introduction into the discussion of the topic of ‘fraudulent claims’ (which is also referred to as ‘evasion’, corresponding to something that might be regarded in modern terms as the spectrum of aggressive tax avoidance and evasion). However, this point is of less relevance in the context of the Report because, though included in the terms of reference, the topic was largely ignored by the economists in their response. Nonetheless, the discussion of ‘evasion’ became an ever-present feature of subsequent League of Nations reports.36

(p. 16) 1.32  The Report was delivered in April 1923.37 Not surprisingly, the four economists’ fairly lengthy discussion of double taxation led to the conclusion that it is highly damaging. Their discussion then proceeded to the general principles by reference to which taxation is levied and the analysis concluded that economic allegiance is, or should be, the appropriate basis for the allocation of taxing rights.38 In the subsequent analysis of economic allegiance, the report proceeds to the conclusion that the two most important factors are the origin of wealth (i.e. the source, or place where wealth is produced) and the residence or domicile of the owner who consumes wealth.39 In what seems a rather broad-brush approach, the discussion leads on to apportion economic allegiance as between origin and domicile in the case of a number of different types of assets or wealth.40 For example, land and commercial establishments are regarded as a category of wealth where origin is the preponderant element, while domicile is regarded as the preponderant element for shares and bonds. In the course of the discussion on this classification of income approach, the difficulties that could be caused by intermediate or interposed companies were recognized but these were not explored in any detail.41 The focus of the discussion then turns to the possible methods for avoiding double taxation, noting that this will involve the surrender of revenue by governments. Before addressing how governments ought to view this matter, consideration is given to the practical reality and a conclusion is swiftly arrived at: ‘A survey of the whole field of recent taxation shows how completely the Governments are dominated by the desire to tax the foreigner.’42 This might suggest, the Report notes, that ‘Governments would be prepared to give up residence rather than origin as establishing the prime right [to tax]’ though the Report argues that this would be a misconceived approach leading to a number of difficulties.

(p. 17) 1.33  The Report then considers four possible methods for dealing with double taxation. These may be summarized as: (i) a reduction of tax in the residence country for tax paid in ‘origin’ locations; (ii) no origin taxation—i.e. a source country exemption for non-residents; (iii) a division of taxing rights between origin and residence countries; and (iv) specific categories of income only taxed at origin.

1.34  In drawing conclusions on applying these principles to the taxation of income, approach (ii), allocating exclusive taxing rights to the residence jurisdiction, is identified as ‘the most desirable practical method’ (because it would allow taxation of total income based on ability-to-pay principles). It was therefore recommended for adoption wherever countries felt in a position to operate the approach.43

1.35  However, it was recognized some debtor states would not accept the absence of origin or source taxation,44 in which case method (iv), possibly as modified by method (iii) was recommended. Method (i) is barely discussed but flatly rejected because the economists saw source taxation as discouraging investment by foreigners.45 The Report goes further in predicting, in rather patronizing terms (and as it turns out, quite wrongly), that the incidence of origin (source) taxes will diminish as semi-developed countries become more industrialized, leading to a wider application of exclusively domicile (residence)-based taxation.46

1.36  The 1923 Report included a short addendum which is concerned with the identification of appropriate criteria to deal with the allocation of earnings in cases where different parts of a single business are conducted in two or more states, such as where manufacture or production takes place in one territory but the relevant sales are made in another. The Addendum noted various allocation criteria that were generally adopted in practice (e.g. sales, capital employed, salaries and establishment, and costs) and observed that the suitability of criteria depends entirely on the nature of the business. The discussion offered no solutions but suggested that the practical experience of some countries (such as the US where US states had experience of different methods for dividing the profits of franchise companies such as railways and general business) might be collected and collated and used by states with (p. 18) less experience as the basis of treaties. In due course, this suggestion would be taken up in a significant way by the League of Nations.

2.  The 1925 Technical Experts Report

1.37  Before the four economists finished working on their report, the Financial Committee of the League of Nations proposed that a special committee of government representatives should be formed to consider possible ‘common action’ on the questions of double taxation and tax evasion.47 Once formed, the special committee (which comprised senior representatives of the tax administrations of the seven constituent countries of the League of Nations) met five times between June 1923 and February 1925 and then produced a series of resolutions, preceded by a report explaining the basis on which they were arrived at, together with a commentary. This is the 1925 Technical Experts Report. It took account of the 1923 economists’ Report (and the work of the ICC) but it represented an independent assessment ‘from an administrative and practical point of view’ of viable solutions to the problem of double taxation.48

1.38  The approach of the Technical Experts was firmly rooted in the then current operation of tax systems and, in contrast to the previous work of the economists, it was practice not theory that determined their approach.

1.39  After noting the very wide application of origin taxes (and recognizing that some countries would find it difficult to dispense with them), the starting point for the Technical Experts was that origin taxes were typically in the form of ‘impôts réels’ (also referred to as ‘real’ or ‘impersonal’ taxes) i.e. separate taxes on particular types or classifications of income or assets, such as income from real property, or income from commercial and industrial business activities. Such taxes applied without regard to the personal characteristics of the owner of the income (hence, ‘impersonal’ taxes). A number of European countries imposed such taxes at this time. Such taxes stood in contrast to the personal income taxes levied on the total (i.e. aggregated) income of taxpayers, often at graduated rates depending upon the circumstances of the taxpayer concerned, by countries such as the US and Great Britain. The approach adopted by the Technical Experts was centred on the distinction between these two categories of taxes that states in practice imposed.49 A major difference to the previous work of the economists was the receptivity of the Technical Experts to source or origin taxation, though only in the case of impersonal taxes.50

(p. 19) 1.40  The 1925 Report accepted that source state impersonal taxes could be imposed in certain designated cases or categories of income. This was a significant difference to the approach suggested in the 1923 Report where taxing rights over impersonal taxes were divided, according to the classification or category of the income concerned, between residence and source countries by reference to the economic allegiance approach. One of the categories of income on which impersonal taxes in the source state were accepted related to profits from industrial and commercial business activities.51 It was further clarified that, where a business carries on its activities in several states, the principle of the division of income between these states should apply.52 The discussion went on to refer to different methods for apportioning income, drawing on the comments made by the economists in their Report, and particularly in its Addendum. Reference was also made to then current tax treaties which already had provisions on this point and which ‘provide a sufficiently accurate basis for computing the division of profits’.53 For example, reference was made to the Treaty of March 1924 between Danzig and Poland, which ‘takes the kilometric length as the basis or index of division in the case of transport enterprises, and gross receipts and profits in the case of other business enterprises’. It was noted that such apportionments were carried out in several countries already and also that these provisions in treaties concluded in Central Europe applied to both personal taxes and impersonal taxes.54 The relevant resolution that was adopted in the 1925 Report provided that if the enterprise has its head office in one of the states and in another has ‘a branch, an agency, an establishment, a stable commercial or industrial organization, or a permanent representative’ then each state would tax ‘the portion of the net income produced in its own territory’.55 The treatment of subsidiary companies was not discussed, so at this stage it was not clear whether a subsidiary company was intended to fall within the scope of this rule. The point was clearly important given that some countries treated a subsidiary as a permanent establishment of the parent in their domestic law.56

1.41  For personal (i.e. income) taxes, the principle preferred was that the state of domicile (residence) only should levy taxation, though the report—and the specific resolutions proposed—recognized that the state of source might also impose taxation, in which case double taxation relief was to be provided in the state of domicile (residence). The precise method for achieving such relief was left to the states concerned in any particular case but two possible methods were mentioned in the Report. The first method involved a (p. 20) deduction from the tax levied in the domicile state (calculated by either excluding the income derived in the state of origin—as in the exemption system approach—or alternatively taxing it but allowing a deduction or credit for the origin tax already suffered—as in a credit system approach). Under the second method, the income would be divided between domicile and origin states with each state taxing only the allocated portion.57 Further, the scope of any source taxation was not demarcated and the Report noted it may be enlarged by states concluding bilateral treaties, though the Report indicated that source taxation would apply in specified cases, including in the case of industrial and commercial concerns.58

1.42  In relation to the operation of a domicile (residence) basis of taxation, the Report also noted the imprecision of the ‘domicile’ test for individuals and corporate bodies. In response, it proposed more specific tests including, in the case of a company, ‘the place where the concern has its ‘effective centre’, i.e. the place where the ‘brain’ management and control of the business are situated’.59 The Report also contained a lengthy section on tax evasion, noting the scale of the tax evasion issue and also that the topic had a clear interaction with the double taxation issue.60 To deal with such evasion, the Report calls for the exchange of information and administrative assistance between tax authorities. The Report also makes the point that, unlike the double taxation issue, which may be addressed by bilateral means, the problem of tax evasion needed to be addressed by general international cooperation between all tax authorities.61

3.  The 1927 Report of the Expanded Committee of Technical Experts on Double Taxation and Tax Evasion

1.43  The 1925 Report suggested that the League of Nations should consider gathering together a wider group of technical experts to produce, if possible, draft or model tax treaties. In consequence, an expanded Committee of Technical Experts on Double Taxation and Tax Evasion was convened and, following its work in 1926 and 1927, drew up four draft tax treaties based on the 1925 Technical Experts report.62 The draft treaties addressed: (i) the prevention of double tax in relation to both impersonal and personal taxes; (ii) double tax and succession duties; (iii) administrative assistance between tax authorities; and (iv) (p. 21) judicial assistance in the collection of taxes.63 The report of the Committee, including the text of the four treaties, was published in April 1927. The Committee expressed a clear preference in favour of ‘collective’ (i.e. multilateral) treaties but concluded these would be too difficult to deliver given the variety of the fiscal systems of the various countries. The four draft treaties were therefore drawn up as model bilateral treaties.

1.44  The model on direct taxes allocated taxing rights in the manner suggested by the 1925 Technical Experts’ report and so dealt separately with the allocation of taxing rights relating to impersonal and personal taxes.

(a)  Impersonal Taxes

1.45  For impersonal taxes, the model follows the recommendations of the 1925 Report in designating various categories of income (rentals from immovable property, agricultural undertakings, and industrial and commercial establishments) in respect of which taxing rights are allocated to the source state. Article 5 provided: ‘Income from any industrial, commercial or agricultural undertaking and from any other trades or professions shall be taxable in the State in which the persons controlling the undertaking or engaged in the trade or profession possess permanent establishments’. This was the first use of the permanent establishment concept.64 Against the background of the uncertain scope of impersonal taxes, it sought to clarify for the first time the limited scope of source taxing rights on business income.65 There was also included in the draft treaty an explanation of what was intended: ‘The real centres of management, affiliated companies, branches, factories, agencies, warehouses, offices, depots, shall be regarded as permanent establishments’.66 By including the reference to ‘affiliated companies’ in the definition of the term ‘permanent establishment’, the text of the draft was addressing an open point from the Technical Experts’ 1925 recommendations, and now clarifying that affiliated companies, such as subsidiaries, were to be included in the definition of permanent establishments.

1.46  In another significant innovation, the model also provided an exemption from the permanent establishment definition for business dealings through ‘a bona fide agent of independent status (broker, commission agent, etc.)’. The Commentary explained what was meant by the reference to ‘bona fide’—‘the words “bona fide agent of independent status” are intended to imply absolute independence, both from the legal and economic point of view. The agent’s remuneration must not be below what would be regarded as a normal remuneration.’67 This exemption for agents of independent status is important for (p. 22) two reasons. First, the exemption introduced an important limitation on source country taxing rights and the terms of the exemption would be heavily discussed in subsequent years.68 Second, and in the present context more important, in clarifying the scope of the exemption there is introduced, for the first time, what amounts to an arm’s-length test. Specifically, the explanation in the Commentary indicates that, in construing whether an agent meets the terms of the exemption from permanent establishment status, it is necessary to take account of the agent’s role and remuneration. Qualification as an independent agent is therefore dependent on this first application of the arm’s-length test, with the remuneration test being whether it is ‘normal’ in the sense of at the same level as a genuinely independent agent.

1.47  The treaty provisions on permanent establishments also included an approach to dealing with the measure of taxable income: ‘Should the undertaking possess permanent establishments in both Contracting States, each of the two States shall tax the portion of the income produced in its territory.’ The reference to ‘the portion’ was taken from the 1925 Report but for the first time there was a clarification of how the required approach was to be applied: ‘In the absence of accounts showing this income separately and in proper form, the competent administrations of the two Contracting States shall come to an arrangement as to the rules for apportionment.’69 Also, the Commentary provided an additional clarification.

These rules will vary essentially according to the undertakings concerned; in certain States account is taken, according to the nature of the undertakings, of the amount of capital involved, of the number of workers, the wages paid, receipts, etc. Similarly, in cases where the products of factories are sold abroad, a distinction is often made between ‘manufacturing’ and ‘merchanting’ profits, the latter being the difference between the price in the home market and the sale price abroad, less cost of transport. These criteria are, of course, merely given as indications.

(b)  Personal Taxes

1.48  The treatment of personal taxes in the model was rather more straightforward. For personal taxes, the taxing rights were allocated to the state of residence, though with a reduction of tax in the residence state to secure double tax relief for any impersonal taxes suffered in the source state.70

1.49  The approach suggested by this 1927 report is similar in a number of respects to the current approach. Though expressed in terms of impersonal and personal taxes, the approach recommended in 1927 called for the taxation at source of business profits in countries where there is a permanent establishment, with the level of taxing rights determined by reference to the ‘income produced’. While this is similar to the approach that applies currently for (p. 23) the taxation of branches of an enterprise (though the current version of the OECD Model looks to the ‘profits attributable’ rather than the ‘income produced’ by the branch), what is completely different is the wide definition of a permanent establishment to include affiliated companies. It will be seen that this aspect of the proposals was short-lived. Had such an approach been followed, it would presumably have involved the taxation of multiple legal entities within a multinational group, using apportionment or division of profit methods of the type which have been referred to in the discussion above (though on which there had by this time been little detailed discussion and no set of proposals on which there was a consensus). Depending on the precise profit split methodology adopted, this would potentially have materially reduced the ambit of transfer pricing rules by removing the need to price inter-company transactions.

1.50  The draft treaties produced by the Committee on Double Taxation and Tax Evasion are also of significance in introducing a mechanism which allows the financial administrations (what soon came to be referred to as the ‘competent authorities’, the term still used today) to confer and agree how the treaty may be applied in any cases not expressly provided for and in providing a dispute resolution process.71

1.51  It is also worth noting that, in the Commentary to the draft treaty on administrative assistance, comment is made on the relationship between the efforts to remove double taxation and efforts to counter tax ‘evasion’:

From the very outset, the Committee realised the necessity of dealing with the questions of tax evasion and double taxation in co-ordination with each other. It is highly desirable that States should come to an agreement with a view to ensuring that a taxpayer shall not be taxed on the same income by a number of different countries, and it seems equally desirable that such international co-operation should prevent certain incomes from escaping taxation altogether. The most elementary and undisputed principles of fiscal justice, therefore, required that the experts should devise a scheme whereby all incomes would be taxed once, and once only.72

1.52  It has been suggested that these comments support the single taxation principle, which requires that income should be taxed only once, with the result that double taxation and double non-taxation are equally unacceptable.73

1.53  The 1927 Report also noted that, in submitting the draft treaties, only the first step had been completed and that the value of the exercise depends on the degree of use made of the drafts by governments. It commented further that the revision and expansion of the model treaties to keep pace with changes in tax systems and to reflect experience of the draft treaties in practice would lead to wider use of the draft treaties. It was also noted that it would be useful to deal with certain other matters and one of these is the drawing up ‘of model rules for the apportionment of taxation applicable to the profits or capital of undertakings working in several countries’. For all these reasons, a standing committee was proposed (p. 24) to work on these issues and generally to act as an advisory committee and reference point on matters of double taxation, and administrative and judicial assistance.74 That special group, the League of Nations Fiscal Committee, came into being in 1929, but only after a discussion involving a much wider group of members and non-members of the League of Nations on the proposed model treaties.

4.  The 1928 General Meeting of Government Experts on Double Taxation and Tax Evasion

1.54  This General Meeting of government experts, at which twenty-seven countries were represented, took place in 1928 and was convened to test the level of wider support for the proposals which had been developed to address double taxation.75 There was unanimous support for the main principles of the model draft treaties as being ‘a useful basis of discussion’ but this support did not extend as far as a multilateral treaty. It was agreed that, while a multilateral agreement would be preferable, it was not in practice feasible at that time.76 Notwithstanding this support for the previous work of the Technical Experts, there were two material developments at this meeting. First, certain changes were made to the texts of the 1927 versions of the model treaties that had already been prepared by the Technical Experts. Second, the texts of two further alternative model treaties dealing with direct taxes were added in addition to the version already prepared by the Technical Experts.77

1.55  Dealing first with the changes made to the earlier Technical Experts draft from 1927, the most important change was the removal of any reference to ‘affiliated companies’ in the definition, or more accurately, clarification, of the term ‘permanent establishment’.78 The inclusion of the reference to ‘affiliated companies’ in the 1927 draft had clearly proved controversial and was opposed by the US delegate, Dr T. S. Adams, who was supported by the representatives of Belgium and Switzerland. The US argument against the inclusion of ‘affiliated companies’ in the definition was that it was out of step with the general approach of many tax systems, given that the mere holding of shares by a parent in a subsidiary company would not normally lead to the parent being regarded as having a taxable establishment in the jurisdiction of the subsidiary. However, the delegates from Spain and Italy objected on the basis that precisely that result could arise in their countries.79 The discussion was resolved, with agreement on the deletion of the words from the permanent (p. 25) establishment definition, on the basis that any concerns over the manipulation of the profits of the subsidiary company could be dealt with in a different way, given that a state would always have the right to examine the accounts of its taxpayers.80

1.56  A further change, in connection with the treaty provisions on the quantum of income attributable to the permanent establishment, was the removal of the default to ‘accounts showing [the permanent establishment’s] income separately and in proper form’, reflecting an early unease about relying exclusively on the relevant accounting treatment and the possibility of an absence of accounts. The change meant that the sole approach mentioned in the situation where there were permanent establishments in both states was the rather impractical solution of the competent authorities having to ‘come to an arrangement as to the basis for apportionment’. At this stage, the approach to be used was not defined and some form of formulary apportionment could, depending on the circumstances, have been an acceptable methodology.

1.57  Two additional draft treaties dealing with direct taxes were added by the General Meeting. These were added on the basis that they could be applied more readily for the fiscal systems of certain states or groups of states. The original Technical Experts’ treaty had been prepared for countries that levy impersonal taxes and also a personal tax, though it was noted the original draft could be adapted for other situations. One of the additional treaties was designed for use between countries using a domicile (residence)-based approach. The other was designed for countries with different fiscal systems. Both of the new drafts were shorter than the original 1927 version prepared by the Technical Experts and, unlike that text, a significant difference is that they both made no distinction between personal and impersonal taxes, though the principles (and much of the drafting) remained the same.81 The removal of the distinction between personal and impersonal taxes reflected US concerns that the right to tax based on source would be lost if the US taxes otherwise levied were classified as personal taxes, rather than impersonal taxes, and there were also British concerns about double taxation and the impact on competitiveness of British business abroad given the very limited tax relief available under the British tax system.82

1.58  This meant that, following the General Meeting of Government Experts, there were now six draft treaties: the original Technical Experts’ version on direct taxes, though as now amended by the General Meeting, together with the two new alternatives prepared by the (p. 26) General Meeting, as well as the amended texts of the Technical Experts’ drafts on double taxation and succession duties; administrative assistance in matters of taxation; and assistance in the collection of taxes. These 1928 model treaties, and in particular the versions dealing with direct taxes, have proved highly influential as the benchmark for very many tax treaties. The fundamental structure used in 1928, under which, broadly, passive income is allocated by residence and active (business) income by source, forms the basis for virtually all of the treaties now in force throughout the world.83 The work in 1928 is also significant for its bilateral rather than multilateral approach, and for its emphasis on the twin goals of eliminating double taxation and preventing under-taxation or non-taxation (or what today might be referred to as ‘double non-taxation’).84

1.59  The proposals made by the Technical Experts for a standing committee were strongly endorsed by the General Meeting. That endorsement also proposed a wider remit to include all questions connected with the study of fiscal problems; and identified a small number of priority issues to be addressed. One of these priorities was work on rules for the apportionment of the profits or capital of undertakings operating in several countries.85

F.  Early Work of the League of Nations Fiscal Committee 1929–33

1.60  The standing committee of fiscal experts, first proposed by the 1927 Report, came into being from 1929. The League of Nations Fiscal Committee met for ten sessions over the period 1929–46. An early task was to begin work on the specific matters identified by the 1928 General Meeting of Government Experts. This included work on specific rules for the allocation of profits to supplement the provisions of the 1928 model treaty.86 At its first meeting in 1929, the immediate conclusion of the new Fiscal Committee was that, to progress the matter, it would need a good knowledge of existing practices in various countries. It therefore contacted its members with a questionnaire, asking for detailed information on the subject. The Committee also asked the ICC for help on this point, as well as for advice on what would be the best methods of apportionment.87

(p. 27) 1.61  The Fiscal Committee decided to add the status of affiliated companies for the purposes of the permanent establishment rules to its work agenda.88 The Committee recognized that the principle at issue was of great importance, though its application ‘only affected a small number of cases’. The Committee took the view that this issue was part of the wider discussion on the allocation of profits of industrial and commercial undertakings. It was therefore decided to take no decision on the point and postpone consideration until the information gathering exercise referred to above had been completed and the results analysed. This decision suggests that the removal of the reference to ‘affiliated companies’ in the 1928 General Meeting drafts had by no means disposed of the issue and that there was no ready consensus on the matter at this stage. Rather, it was intended that the allocation of profits exercise would help resolve all allocation matters of general relevance to industrial or commercial undertakings, including dealing with the position of affiliated companies.89

1.62  The Fiscal Committee also identified additional areas for consideration. These included work on clarifying the circumstances in which the activities of an agent would crystallize a deemed permanent establishment of the principal and also the requirements for an ‘autonomous agent’ which would be excluded from the operation of this rule.90

1.63  In 1930, at the second meeting of the Fiscal Committee, the focus on the allocation of profits issue was materially increased. By the time of that meeting, some twenty responses to the Fiscal Committee’s questionnaire had been received and these revealed that almost all of the countries levied source taxation.91 The responses were summarized in an appendix to the Fiscal Committee’s report, which is discussed further below. The Committee had also received a response on the matter from the ICC. This second meeting decided to tackle the profit allocation issue head on:

Nevertheless, while fully realising the difficulty of the task, the Fiscal Committee is of opinion that the moment has come to deal with the real substance of the question, (p. 28) since, until this is settled, one of the principal causes of double taxation will continue to exist.92

1.64  The Committee decided to use a recent Rockefeller Foundation grant primarily for the project and a subcommittee was formed to prepare the discussion for the next meeting.93

1.65  The second meeting also included further discussion of the possibility of multilateral treaties being used to deal with certain specific categories of income.94 One of the suggested items for a multilateral treaty related to the treatment of profits derived by a company from industrial, commercial, or agricultural undertakings. The suggested approach was based on that used in the draft treaties discussed above, namely that such a company should not be taxable in a country other than that in which the real centre of management of the company is situated unless the company has one or more permanent establishments in the other country. It was noted that further thought would be required by the Fiscal Committee on three points: first, whether the term ‘company’ should be defined; second, whether the provisions should suggest that the competent authorities would ‘come to an arrangement’ on the relevant apportionment of ‘income produced’ where there is a permanent establishment in each of two contracting states; and, third, whether it would be useful to add the clarification (based on the terms of a resolution at the 1929 ICC Congress) that:

The fact that an undertaking has business dealings with a foreign country through a local company, the stock of which it owns in whole or in part, should not be held to mean that the undertaking in question has a permanent establishment in that country.95

1.66  The discussion on the possibility of framing multilateral treaties was revisited one year later at the next (third) meeting of the Fiscal Committee. By this time, a draft multilateral treaty for the prevention of double taxation of certain categories of income had been produced,96 though it was recognized this would not be acceptable to all countries. This was because, in an effort to avoid double taxation altogether, that draft dealt with the position of residents and non-residents and some countries (such as Great Britain) were not prepared to accept any source taxing rights in respect of overseas income and profits derived by their own (p. 29) residents.97 In attempting to find a solution to the problem, a further draft was prepared that addressed the position of non-residents only. The approach of this draft was to protect the position of residents of one state by limiting the permitted source taxation to certain defined categories of income only. However, it was recognized that there was at this stage no complete solution for a multilateral treaty that would achieve a clear consensus, but that it would be a measure of progress if some countries were able to use one of the drafts.

1.67  Drawing on the previous drafts and principles dealing with the cross-border operation of an industrial, commercial, or agricultural enterprise, both multilateral drafts incorporated the same (by now, familiar) approach of stipulating no source country taxation except to the extent of the ‘income derived’ by a permanent establishment in the source state.98 The measure was supplemented by provisions directing the competent authorities to come to an agreement if necessary, regarding the methods of apportionment and directing that the use of a genuinely independent agent should not lead to a finding of a permanent establishment. The accompanying report refers to the fact that consideration was given to adding the clarification that where an undertaking has business dealings with a foreign country through a local company the stock of which it owns should not mean the undertaking has a permanent establishment in that country. After discussion, the Subcommittee considered ‘this addition might lead to considerable difficulties and decided to reject it’.99 At this stage, therefore, the position of affiliated companies remained unclear and presumably the Fiscal Committee still intended to deal with the point in the work on the apportionment of profits.

G.  Detailed Work on the Allocation of Profits, 1930–3

1.  Introduction

1.68  Very extensive work on the allocation of profits took place in the period 1930–3, following the appointment by the League of Nations of Mitchell B. Carroll to carry out a multi-country research project on the topic. However, a significant amount of effort had by then already been devoted to the gathering of reports on country law and practices on the allocation of profits. As noted above, the ICC had devoted significant time and resources to the matter from its inception, leading to a large number of country reports being prepared as the ICC developed its own principles for dealing with the allocation of profits. The ICC data and the experience of its members was also shared with the League of Nations, whose own Technical Experts were appointed in part due to their practical knowledge of their own country tax systems. The 1923 economists report had also briefly considered the issue.

1.69  As noted above, in 1929, the Fiscal Committee of the League of Nations had also commissioned a further round of country reports on the topic and received responses from (p. 30) twenty countries.100 The country reports comprising that survey are reported as revealing great diversity in law and practice from one country to another. However, in the summary report it is noted that there are two areas where the approach is more uniform—these are (i) the separate assessment of a local branch or subsidiary based on its own accounts and without resort to the accounts of the foreign company or a foreign holding company and (ii) the use of a fractional apportionment approach in only a minority of cases (such an approach is akin to a profit split approach and involves computing the income or profits of a branch or subsidiary as a fraction of the entire income or profits of the foreign company or of the holding company).101 It is also mentioned in a number of the country reports that the accounts of a branch or local subsidiary could in some cases be inadequate or misleading. This leads to comments on the ‘empirical methods’ that could be used by the tax authorities in such a case to estimate the true taxable profit. These include the making of assessments based on a ‘comparison with the earnings of local undertakings engaged in a similar business’ or ‘a comparison with the profits of other companies engaged in a similar business’ and in the case of the US it is also mentioned that, in order to prevent evasion or show true income, the fiscal authorities may allocate income as between the foreign parent company and the local subsidiary. The methods used were reported as being diverse and included comparisons of the percentage of net profit to turnover or by taking into account capital invested, turnover, and other relevant data, as well as apportionments that were based on assets, turnover, expenditure, or number of employees.102

1.70  All of this earlier work on profit allocation methods, including the large number of country surveys, meant that, by the time Mitchell B. Carroll was appointed to work on this topic in 1930, the particular areas of interest (and difficulty) were well known and included: (i) the range of legal approaches and particularly the various methods actually used in practice by states to allocate income or profits; and (ii) the relevant conceptual approaches that underlay those approaches—in particular whether countries adopted a stand-alone (‘separate accounting’) approach to the calculation of the profits of a local branch or permanent establishment or whether they would look at the entire (i.e. global) profits of the business and then apportion an appropriate percentage or portion of that to a local permanent establishment (this was the approach often referred to as ‘fractional apportionment’).103 As is (p. 31) evident from the discussion above, another important issue was how countries approached the taxation of affiliated companies—i.e. whether as separate taxpayers in their own right or as a mere extension of the wider business carried on by an overseas holding company. A further issue was how they dealt with inadequate or misleading local accounts to ensure an appropriate allocation of profits. All these issues had been recurrent themes in the earlier discussion.

1.71  The detailed work subsequently carried out by Mitchell B. Carroll, was carried out in two phases. He was appointed following the second meeting of the Fiscal Committee in 1930 to compile an analysis of relevant law and practice in relation to five countries—France, Germany, Great Britain, Spain, and the US. This work had been completed by the time of the third meeting of the Fiscal Committee in 1931 and it was subsequently published as Volume I of the five-volume work, Taxation of Foreign and National Enterprises.104 The initial intention had been to carry out the work with a view to then formulating general principles that would take account of country practices. However, having completed this work, he considered it impossible at that stage to reach final conclusions, presumably due to the small number of countries on which detailed work had been carried out. The Fiscal Committee therefore decided to extend the enquiry to other countries ‘of the greatest economic importance’.105 This led to a large number of further lengthy separate reports (later published as Volumes II and III of the same study).106 These were accompanied by a report, written by Mitchell B. Carroll, setting out the general conclusions that were then drawn on the basis of all the country reports (comprising Volume IV) and, in Volume V, a study of the accounting methods used.107

1.72  As with the preceding material, the main focus of the work was on the issue of branch allocations, largely because this was the principal issue encountered by countries in practice.108 The US, however, was a marked exception because the majority of overseas (p. 32) entities conducted their business in the US through locally incorporated subsidiaries. It was noted that questions of allocation would therefore not generally arise in the US unless profits were shifted from a US company to an overseas affiliate (or vice versa) and for this situation there was already protection in US law from a broad-based provision which, as will be seen below, proved to be the main provision on which the ALP was modelled.109

1.73  Notwithstanding the recurrent earlier efforts of the League of Nations and ICC to survey existing law and practice on the topic, this five-volume study of some 1,300 pages on the allocation of profits was, and remained for several decades, by a long margin the most detailed consideration of the topic undertaken by an intergovernmental body.

2.  First Principles

1.74  In Volume IV, Mitchell B. Carroll set out his conclusions after summarizing country law and practice on the key areas. The detailed work on existing country law and practice was enormously important to Mitchell B. Carroll because he saw the overall goal of the work on allocation as being to bring together in a common approach as much as was possible of the existing principles of the different systems.110 However, he noted that the task would be impossible if clear principles of jurisdiction, based on fiscal domicile and source, were not first established. He emphasized that the concept of fiscal domicile should be uniformly defined (and his suggestion here was that this should be by reference to the centre-of-management test) and that the concept of source should be defined and strictly limited to sources obviously within the jurisdiction of the source state.111 In relation to source taxing rights, he set out rules relating to various categories of income that he considered were observed by the majority of countries and to a certain extent by the model treaties.112

1.75  In relation to the treatment of business income, he identified the key provisions as those contained in the existing draft treaties prepared by the League of Nations which provided that income from industrial, commercial, or agricultural undertakings should be taxable in the state of the relevant permanent establishment and that if there were two or more such permanent establishments each state should tax the ‘income produced’ in its territory.113(p. 33) Having confirmed the general approach to be applied, he then proceeded to deal with, first, the required treatment of subsidiaries and, second, the approach to branches, particularly in relation to where income needs to be allocated between different permanent establishments. The common theme on both matters is the aim to treat the subsidiary or branch as an independent enterprise.

3.  Conclusions on Subsidiaries

1.76  With regard to subsidiaries Carroll’s starting point was to deal with the open issue as to whether subsidiaries should be treated as permanent establishments. He argued strongly that they should not be so regarded:

It is evident from the tenor of Article 5 and its commentary that the term ‘undertaking’ or enterprise includes, when referring to a corporation, merely the corporate entity and its own branches, forming a part of the single corporate entity, and does not include subsidiary corporations organised in the same or other countries which are themselves separate legal entities, or affiliated corporations which are not controlled by the first-mentioned corporation itself, but are owned or controlled by the same interests as those which control the first-mentioned corporation.

1.77  Mitchell B. Carroll viewed this to be in line with the approach of most countries to taxing income from industrial or commercial profits. His summary of the approach adopted by the ‘great majority of countries’ was that they were content to respect the separate legal existence of the subsidiary company and tax it on a stand-alone basis as long as ‘the inter company transactions are carried on under the same circumstances and conditions and on the same terms as they would be between two entirely separate and independent persons, dealing with each other in open market, and in a manner which is graphically described as at “arm’s length” ’.114 He recognized that subsidiary companies may carry on functions which could readily be carried on by a branch of the parent but for him there was a decisive difference:

The fundamental legal difference, however, is that each transaction between the parent company and a subsidiary company, or between two subsidiaries, should be carried out as a legal transaction between independent enterprises, whereas if the corporation carries out its activities through its own branch or branches, it does not generally in practice, nor in law, make contracts with or between such component parts of its own organisation … Economic fact must inevitably give way to the definite principles and provisions of law under which business is conducted.115

(p. 34) 1.78  Accordingly, it was the distinction in law between a subsidiary and a branch that explained why a different tax treatment should be applied.

1.79  Looking back at these historical developments from a modern day perspective, and in particular having regard to the BEPS perspective of seeking to reduce or remove for tax purposes the effect of formal legal differences with little or no impact on the substance of the functions carried on, it seems ironic that this fundamental step, ultimately leading to the quite distinct tax treatment of subsidiaries and branches, and also the need for a detailed set of transfer pricing rules, was so heavily influenced by legal constructs alone.

1.80  Having determined that a subsidiary should not be treated as a permanent establishment, it followed that some rule would be required to prevent the double taxation of business income by virtue of different states taking different views on the measure of business income allocable to subsidiaries located in their states. The approach to developing the rule however was more focused on preventing abuse than preventing double taxation, though the result would be the same. Mitchell B. Carroll clearly recognized that treating the subsidiary as a separate and independent enterprise for tax purposes might lead to abuse of pricing or allocation arrangements and he noted that some tax authorities were already ‘disposed to regard some local subsidiaries as “dummies” or fictitious entities designed to set up artificial legal barriers in the way of the assessment of true profit’.116 Nonetheless, he was confident of dealing with any such abuse by what was referred to as the ‘arm’s length’ approach of adjusting any diverted income. This led to his overall conclusion that ‘if it is shown that inter-company transactions have been carried on in such a manner as to divert profits from a subsidiary, the diverted income should be allocated to the subsidiary on the basis of what it would have earned had it been dealing with an independent enterprise’.117 The study did not propose any particular text or version of the arm’s-length rule but noted that variations of this approach (though with a similar result in each case) were already used in the US, Canada, and Austria. The US rule, contained in section 45 of the United States Revenue Act of 1932, permitted the Commissioner of Internal Revenue to request from the local subsidiary all accounts and information necessary to allocate or apportion the income or deductions between the two companies in such a manner as to prevent evasion (by the shifting of profits) in order to arrive at the true tax liability of the local company. The Canadian law contained a provision which was fundamentally the same, but which was more precise in its language and more limited in scope. The Canadian rule stipulated that:

[W]‌here any corporation carrying on business in Canada purchases any commodity from a parent, subsidiary or associated corporation at a price in excess of the fair market price, or where it sells any commodity to such a corporation at a price less than the fair market price, the Minister may, for the purposes of determining the income of such corporation, determine the fair price at which such a purchase or sale shall be taken into the accounts of such corporation.

1.81  Under the Austrian practice, a local subsidiary company would be respected as legally independent but if there was evidence of diversion of income through artificially fixing invoice (p. 35) prices, the Austrian administration would adjust the price between the subsidiary and parent, with the alteration based on a comparison with information provided by similar enterprises at existing market prices, or, in the absence of such evidence, on the opinion of business experts. The business experts consulted would typically be operating in the same or similar branches of business, excluding direct competitors.118

1.82  Mitchell B. Carroll’s confidence in the effectiveness of the arm’s-length approach in fact led to further support for treating subsidiaries as separate entities for tax purposes:

As the conduct of business between a corporation and its subsidiaries on the basis of dealings with an independent enterprise obviates all problems of allocation, it is recommended that, in principle, subsidiaries be not regarded as permanent establishments of an enterprise but treated as independent legal entities.119

4.  Conclusions on Branches

1.83  Having dealt with the position of subsidiaries, the discussion proceeded to address the position of branches. Here, the overall goal was that branches were to be treated ‘in so far as possible as independent entities’.120 He noted that such a result would be achieved in many cases if the very nature of the activities of the branch concerned made the branch an autonomous establishment, as where, for example, a factory buys in raw materials on world markets and sells its goods in local markets. However, the more the income of the establishment was dependent on its relations with other establishments or subsidiaries in the same group (making the branch a ‘dependent establishment’), the more difficult became the problems of allocation. This would be the case where different functions or operations (such as purchasing, processing, or selling) comprise an integrated business yet are carried on in two or more establishments located in different countries. This obviously led to the issue that if the income was produced by establishments in two or more countries, then the business income must be allocated, which requires the determination of which establishments are productive of income and how much is attributable to each. Mitchell B. Carroll found little evidence of the consideration of this point by governments or the courts but he did identify a range of country practice.121 This included the use of separate accounting, fractional apportionment, and various other empirical approaches. His conclusion was strongly in favour of the need for a separate accounting approach. This was for a number of reasons: that approach was in any event used by most countries including those with most experience in dealing with tax matters; it was preferred by business as represented in the ICC; it would preclude the taxation of (p. 36) unrealized profits; it was supported by the general requirement under fiscal or commercial law for maintaining accounts and overcame problems of language; it would be impossible to secure widespread agreement on uniform factors for the fractional apportionment approach; and because empirical methods—which he considered also included fractional apportionment methods—were unprincipled. The separate accounting approach was also the logical corollary of the adopted principle that branches should be treated as far as possible as independent entities.122 It was also significant that such an approach avoided the impracticability of having to obtain extensive data from sources outside the jurisdiction concerned.123

1.84  The country reports, and the summaries in Volumes I and IV of the study prepared by Mitchell B. Carroll, contain discussion of country practices where a separate accounting approach is used for branches but the accounts are unreliable or the profits reported appear to be unreasonably low. In such cases, it is noted that the resort is often to methods that compare the results of the entity concerned with the result that would have arisen in the event of transactions with independent parties.124 Various instances of a similar resort to comparisons to transactions or profits of independent entities in cases where a separate accounting approach is not used are also cited.125 This led Mitchell B. Carroll to the conclusion that the approach of using comparisons with the transactions or profits of independent parties should be of significant help to ensure the autonomy, for tax purposes, of a local branch. Throughout the discussion on branches, there is the same confidence, expressed in the discussion of subsidiaries which has been referred to above, in the effectiveness of the arm’s-length approach where it could be applied due to the availability of third party pricing information: ‘If by any chance the profits are diverted by means of excessive interest rates, royalties, charges for technical advice, engineering or other services, the fairness of these charges can readily be determined and excessive charges can be adjusted.’126 Mitchell B. Carroll recognized there would be many cases where there were no readily available market prices but he remained confident in the efficacy of the arm’s-length methodology, with any such cases ‘left to the judgment of the executives of the company and the interested tax officials’.127

5.  Sector Analysis—Business Income

1.85  Having considered the general principles of the approach to allocation, the discussion then turned to the specific sectors where most difficulties were likely to be encountered. Here, the sector presenting the greatest problems was identified as relating to industrial and (p. 37) mercantile enterprises. Mitchell B. Carroll had already identified the overall goal as being to treat branches in so far as possible as independent entities but at this point he expressed his goal in terms that were designed to encapsulate the necessary conditions to meeting that objective. Specifically, he referred to ‘the purpose of reflecting in the separate accounts of each branch establishment of a foreign enterprise, the income it would derive if it were an independent enterprise carrying on similar activities under similar conditions’. It was this formulation or hypothesizing of the branch as ‘if it were an independent enterprise carrying on similar activities under similar conditions’ that was to prove a blueprint for the fundamental assumptions that must be made for the purposes of the branch attribution rule.128 The approach suggested was in essence the equivalent of the ALP for intra-entity or branch situations.

1.86  The discussion ultimately led to the proposal that for industrial and commercial businesses where profits need to be apportioned, there were two main approaches or models that could be used to deliver that result. The two approaches were: (i) allocation on the basis of remuneration for services rendered to the enterprise; and (ii) allocation on the basis of presuming a sale between the interested establishments at the price which would prevail between independent persons dealing at arm’s length.129 The ‘remuneration for services’ approach was based on the idea that the centre of management was the principal productive establishment of the enterprise, with the result that any branches of the enterprise in other countries would be of secondary importance and rewarded with only a fee or commission, even where engaged in processing, manufacturing, or selling activity. The alternative basis of presuming a sale between the establishments was founded on the supposition that each establishment engaged in producing, processing, or manufacturing materials sells at an appropriate profit to any establishment with which it deals and for this reason was said to be suitable only where the establishment concerned was fairly autonomous.130 The discussion considered the principles of allocation in the case of a number of possible circumstances, e.g. selling or buying establishments, processing establishments, research bureaux, and warehouses, but recognized that there would still be difficult issues of allocation:

There is apparently no theoretically perfect rule for determining exactly how much of the income is attributable to each establishment any more than there is an accurate way for apportioning the compensation of an individual workman to his hands and his feet, and to his brain which has coordinated all his efforts. Income is sometimes classified, for tax purposes, as income from capital, income from work, and income from work and capital combined, the profits of an industrial and commercial enterprise being included in the last mentioned category. It is obvious that the proportion of work to capital varies from business to business and that, in the alchemy of a successful business, the intangible, immeasurable element of brainwork is a very important factor, if not the most vital factor. This is impossible to measure accurately . . .131

(p. 38) 1.87  For this reason, it was important that some flexibility be applied in practice though it was recommended that the selection of the method ‘should depend, in so far as possible, upon recognized methods of operation for the particular types of enterprise and upon practicability’, though any method chosen should also ‘facilitate verification by the tax collector’.132

1.88  Volume V of the study is intended to explore further the different methods that might be used in practice for the required allocations.133 The discussion, which lends further support to the use of a separate accounting approach, re-emphasizes that there is no single method to be preferred and that any method chosen is unlikely to deliver any scientific precision.134 The discussion considers the application of various different approaches such as the commission method, the independent dealer price method, the apportionment of joint profit, etc., but a common strand is that ‘profits can be divided in most instances by resorting to methods used in dealings between independent enterprises’.135

6.  Significance of the 1930–3 Work

1.89  The detailed work on allocation discussed above marked a significant shift in clarifying the fundamental operation of the arm’s length principle and its detailed application to the transactions of both subsidiaries and branches. Though the work had not yet articulated for separate entities (as it had for branches, or permanent establishments) the precise terms of the ALP, the work was clearly a major influence on all the efforts on allocation tests that followed,136 though it cannot be regarded as a complete solution to issues relating to allocation problems.137 It is striking how many of the difficulties that have continued to be sources of transfer pricing conflict were identified in Mitchell B. Carroll’s work.

1.90  This early work on allocating income is also relevant for a number of other reasons. First, the report reflected, in the clearest terms yet expressed in these early deliberations, material concerns about taxpayer abuse, especially in relation to the manipulation or ‘improper restriction’ of transfer prices—for example by invoicing goods at an excessively high price or levying excessive interest or royalty charges to manipulate the level of profits.138 It is also clarified, in the discussion of branches of foreign enterprises, that a number of countries surveyed were already operating domestic approaches to counter such practices by having recourse to a comparison with similar independent enterprises, frequently taking into account the percentage of net profits to gross turnover, and using a variety of other (p. 39) ‘empirical’ methods.139 Similar concerns, and anti-avoidance approaches, are also repeatedly reflected in the comments on subsidiaries of foreign companies.140 The discussion is an early portent of the two functions, or requirements, of the arm’s-length approach, namely as a mechanism for fairly apportioning profits or income and expenses to avoid double taxation and as a mechanism to counter abuse.

1.91  Second, there are some oblique references to a primary conceptual problem that would emerge with the development of the ALP, namely the point that affiliated entities may well organize themselves or their affairs in a way that independent parties would never do, with the result that any independent entity standard may well prove inherently problematic:

[T]‌he present study … reveals the difficulty in determining exactly the part of the profits that an enterprise realises on the territory of each of the states in which it has an establishment. At present, each establishment, each branch, each subsidiary company do not constitute autonomous exploitations having their own interests and with profits which are destined to be added mathematically together; they appear rather to be joint members of a single organisation and to have the purpose of co-operating in various ways to realise a joint profit. It is not possible to say that one establishment has procured a certain profit because that establishment may only serve to suppress a dangerous competition for the establishment situated in another country. In this case, the accounts of the establishment will most frequently show a deficit whereas in reality it will have contributed to increasing the turnover of the other establishments, and consequently, the general profit.141

1.92  Third, the work also seems for the first time to hint at the inevitable complexity that would in time grow with the development of the new approach—there are, for example, references to the need for transfer pricing ‘experts’ and detailed functional analyses, with account taken of capital invested, volume of business, costs, and risks. Also recognized is the need for widely differing approaches in the chosen allocation methodology depending upon the business sector concerned, though at this stage the implications for the scale of the resulting complexity were by no means appreciated.142

H.  The 1933 Draft Model Convention on the Allocation of Profits

1.93  The recommendations of Mitchell B. Carroll were quickly converted into applicable model treaty provisions. Based on his work, a draft model treaty dealing solely with the allocation of profits of industrial and commercial enterprises was agreed at the 1933 meeting of the League of Nations Fiscal Committee.143 The intention was that the draft treaty would (p. 40) supplement the earlier 1928 treaty drafts prepared by the General Meeting of Government Experts as these had left open how the allocation of profits should best be achieved. (Those earlier drafts included provisions under which source taxing rights would apply to the ‘income produced’ by a permanent establishment but it was left to the competent administrations to ‘come to an arrangement as to the basis for apportionment’.) In giving further direction on that open point, it was decided that the model treaty on the allocation of profits should be restricted to only the most essential provisions so that these could more readily be applied by states in their bilateral treaties or in their domestic law. It was also noted that, given the diversity of national laws and the complexity and variety of individual cases, the Model was intended to prescribe only general principles, though with the detailed Carroll report being available as a guide to the application of those principles. The possibility of creating a multilateral treaty was also explored, though this came to nothing.144

1.94  The model treaty contained different provisions applying the separate enterprise principle to branches and companies.

1.  Branches

1.95  Following (and slightly revising) the recommendations of Mitchell B. Carroll in formulating the equivalent of the ALP for branches, the fundamental principle adopted in the draft treaty was that:

If an enterprise with its fiscal domicile in one contracting State has permanent establishments in other contracting States, there shall be attributed to each permanent establishment the net business income which it might be expected to derive if it were an independent enterprise engaged in the same or similar activities under the same or similar conditions.145

1.96  On the critical issue of how that ‘net business income’ should be determined, what was proposed in the draft treaty amounted to a four-step approach designed to arrive at the appropriate measure of what was attributable to the permanent establishment.

1.97  The first step was with the relevant accounts of the branch: ‘Such net income will, in principle, be determined on the basis of the separate accounts pertaining to such establishment.’146 It was recognized that the application of this principle would vary from case to case so, with the exception of banking and financial enterprises for which a special article was included in the treaty, no further comment was made on particular industry sectors on the basis that the use of the ‘separate accounts’ approach would enable all special problems to be solved with the necessary flexibility.147

(p. 41) 1.98  Under a second step, further provisions allowed the tax authorities of the state of the permanent establishment where necessary to ‘rectify’ the accounts produced to correct errors or omissions or to ‘re-establish the prices or remunerations entered in the books at the value which would prevail between independent persons dealing at arm’s length’.148

1.99  Third, in cases where there were no accounts, or where inadequate or unreliable accounts could not be rectified as discussed above, or where the taxpayer agreed, a ‘percentage of turnover’ basis was permitted to fix the relevant business income of the branch concerned. It was clarified that ‘this percentage is fixed in accordance with the nature of the transactions in which the establishment is engaged and by comparison with the results obtained by similar enterprises operating in the country’.149

1.100  Finally, there was also a further backstop provision to the effect that if the above approach could not be applied (which would be the case where it was not possible to compare the activities of the permanent establishment under consideration to those of fully-fledged domestic enterprises), then the net business income of the permanent establishment could be determined (using a formulaic approach) by reference to the total income derived by the enterprise as a whole. However, in this case it was provided that such a determination was to be made by applying to the total income certain coefficients based on a comparison of gross receipts, assets, number of hours worked, or other appropriate factors, provided such factors deliver ‘results approaching as closely as possible to those which would be reflected by a separate accounting’.150 This reflects a marked change from that seen in the earlier 1928 model treaties, with formulary methods by 1933 permissible only as a last resort and even then subject to the requirement of replicating as nearly as possible a separate accounting approach.

1.101  The term ‘permanent establishment’ was clarified in broad terms to include a range of establishments such as offices, warehouses, oil wells, factories, and agencies, but, reflecting the conclusion of Mitchell B. Carroll, it was expressly provided the term does not include a subsidiary company.151

1.102  The 1933 Model also introduced the forerunner of what we today refer to as the dependent agent rule (OECD Model Article 5(5)). The early version of the rule deemed a permanent establishment to exist where a foreign enterprise ‘regularly has business relations in a State through an agent established there who is authorised to act on its behalf’. It was further provided this would be the case when the agent concerned was a duly accredited agent who (i) ‘habitually enters into contracts for the enterprise’; was bound by an employment contract, and ‘habitually transacts commercial business’ for the enterprise; or (ii) ‘is (p. 42) habitually in possession, for the purposes of sale, of a depot or stock of goods belonging to the enterprise’.152

2.  Companies

1.103  The emphasis on the independent enterprise standard was not restricted to branches. While the above provisions represent a reflection of the approach previously adopted in draft model treaties, for the first time, a new provision was introduced as Article 5 of the Model:

When an enterprise of one contracting State has a dominant participation in the management or capital of an enterprise of another contracting State, or when both enterprises are owned or controlled by the same interests, and as the result of such situation there exists, in their commercial or financial relations, conditions different from those which would have been made between independent enterprises, any item of profit or loss which should normally have appeared in the accounts of one enterprise, but which has been, in this manner, diverted to the other enterprise, shall be entered in the accounts of such former enterprise, subject to the rights of appeal allowed under the law of the State of such enterprise.153

1.104  This, of course, is the forerunner of the ALP. The Commentary on the provision, which echoes the earlier debate on the status of affiliated companies, may be quoted in full:

Article 5 deals with subsidiaries which will be taxed as independent enterprises provided no profits or losses are transferred as a result of the relations between the affiliated companies. If such transfers are effected, the administration will make the necessary adjustments in the balance-sheets.154

1.105  In conceptual terms, the 1933 version of the ALP is remarkably similar to the version that features currently in the text of Article 9 of the OECD and UN Models. Both versions apply only where there is common control of the parties; both versions express in very broad terms the circumstances in which they will operate (‘conditions’ in their ‘commercial or financial’ relations being in place which differ from those that would have been made between independent enterprises); both versions require independent and comparable reference transactions; and the effect of both versions is the same (an adjustment of profits according to the independent enterprises standard).

1.106  The 1933 version of the ALP is generally regarded as derived from section 45 of the US Revenue Act of 1928,155 which read as follows:

In any case of two or more trades or businesses (whether or not incorporated, whether or not organised in the United States, and whether or not affiliated) owned or controlled directly or indirectly by the same interests, the Commissioner is authorised to distribute, (p. 43) apportion, or allocate gross income or deductions between or among such trades or businesses if he determines that such distribution, apportionment, or allocation is necessary in order to prevent evasion of taxes or clearly to reflect the income of any of such trades or businesses.156

1.107  The terms of section 45 are even broader than those of the 1933 version of the ALP (which itself is cast in broad terms) in that in section 45 it is simply left to the Commissioner to determine both whether an adjustment is required to prevent ‘evasion’ or ‘clearly to reflect … income’ and also how any such adjustment required should be effected. The 1933 version of the ALP is slightly more constrained as a result of the benchmark ‘independent enterprises’ test, though the range in which the rule can operate (‘conditions’ relevant to their ‘commercial or financial relations’) remains very broad. At this stage, there was no clarification of the precise scope and operation of the rule (e.g. dealing with the relationship between the ALP if used in a double tax treaty and the relevant domestic law of the states concerned), presumably because its purpose at the time must have seemed obvious. The terms used in the expression of the ALP (and referred to above), however, might suggest the rule goes beyond correcting price manipulation alone.157

1.108  Mitchell B. Carroll’s thinking had been heavily influenced by experience in the US. The US rule had originally been developed to counter abuse in the financial relations between related parties and responded to the issue that ‘subsidiary corporations, particularly foreign subsidiaries, are sometimes employed to “milk” the parent corporation, or otherwise improperly manipulate the financial accounts of the parent corporation’.158 The concern had been expressed that:

At the present time it is possible—and I am afraid the device is being used increasingly—to incorporate a subsidiary and throw the profits one way or the other. If that subsidiary is a foreign corporation you can throw the profits to it: in other words, by selling products to it at artificially high prices.159

(p. 44) 1.109  In addition to already operating the preferred arm’s-length rule, the discussion in the study Taxation of Foreign and National Enterprises noted a number of ways in which the US seemed to be especially advanced in dealing with allocation issues—for example, in prescribing relatively detailed accounting requirements; in the way the US tax authorities were already tracking by sector the percentages of net income to gross sales for the purposes of policing related party profit shifting; and in operating a definite rule dealing with the allocation of overheads.160 The US was also referred to in connection with some early profit split cases, due to the importance of having introduced the idea that profits can be split up and allocated to different parts of the business process, and also presumably because it was a state in which, unusually, judicial consideration had been given to the allocation of profits.161

I.  The London and Mexico Model Tax Treaties

1.110  For some years, the Fiscal Committee did relatively little further work on the allocation topic after its efforts on the model treaty on the allocation of profits.162 It did, however, continue working on tax evasion and also considered other topics such as the evolution of fiscal systems and the technical principles of tax.163

1.111  The last major work of the Fiscal Committee related to the efforts from June 1939 to update the 1928 model bilateral treaties prepared by the General Meeting of Government Experts on Double Taxation and Fiscal Evasion. An update was undertaken because it was considered that some of the original provisions had been improved in the bilateral treaties entered into in the 1930s and also to take account of certain new trends and new problems that had emerged.164 The work led to various meetings of tax experts in Mexico and London but, because the membership of those meetings differed considerably, ultimately quite different versions of the model drafts were produced, reflecting contrary perspectives on the priority of source taxing rights. The Mexico (1943) and London (1946) meetings each produced three different working drafts of model bilateral treaties, dealing with: Prevention of the Double Taxation of Income and Property; Prevention of (p. 45) the Double Taxation of Estates and Successions; and Reciprocal Administrative Assistance for the Assessment and Collection of Taxes on Income, Property, Estates, and Successions. The material differences in the two versions of the model dealing with the double taxation of income and property related specifically to the treatment of interest, dividends, royalties, annuities, and pensions. In a reversal of the treatment previously followed, the Mexico draft allocated primary taxing rights over all these categories to the source state. This treatment was itself reversed in the later London draft, which restored the primacy of residence taxation for these categories of income. There was also another important difference in that while the London model defined ‘fiscal domicile’ of a company by reference to the country in which the real centre of management is located, the Mexico draft adopted the test of looking to the country under the laws of which it was organized.165 All of the key provisions of the 1933 model treaty on the allocation of profits were included in the relevant London and Mexico drafts, though with some variation as will be discussed below.

1.112  On matters relating to the measure of source taxing rights over business income and allocations, there was much (but not total) commonality between the two drafts, reflecting the influence of the work of Mitchell B. Carroll and the terms of the earlier 1933 model treaty. The London draft used the existence of the by then familiar permanent establishment test as the basis on which source taxing rights over the relevant business income would be created. The Mexico draft, on the other hand, adopted the vaguer test of assigning source taxing rights if the enterprise has ‘carried out’ its business or activities in that country provided such activities did not merely take the form of ‘isolated or occasional transactions’.166 However, somewhat anomalously, the related income allocation rules of the Mexico draft are identical to those in the London draft—i.e. in both cases they apply by reference to the test of whether there was a permanent establishment.167 Both the London and Mexico drafts also provided that if an enterprise has a permanent establishment in each of the states, each state is to tax the part of the income that is produced in its territory.168

1.113  As noted above, the approach to the calculation of taxable profits of an establishment in a foreign country (whether a permanent establishment as in the London draft or an establishment where business activity is carried out as in the Mexico draft) was the same in both cases and reproduced the approach set out in that draft model on the allocation of profits.169 This meant that the objective in relation to any permanent establishment was to identify ‘the net business income which it might be expected to derive if it were an independent enterprise engaged in the same or similar activities, under the same or similar conditions’. To deliver that result, the same four-step approach was included, (p. 46) i.e. (i) the starting point was the separate accounting approach;170 but (ii) the accounts could be rectified in the event of error or omission or to give effect to arm’s-length pricing; (iii) if there were no accounts or if they remained inadequate or the taxpayer and tax authority agreed, then the profits of the establishment could be determined by applying a certain percentage to its gross receipts (with the percentage being fixed in accordance with the nature of the transactions in which the establishment is engaged and by comparison with the results obtained by similar enterprises operating in the country);171 and finally (iv) if the first three steps were inapplicable then a formulaic approach relating to the total profits of the enterprise might be taken provided it delivered results approaching as closely as possible those which would be reflected by the separate accounting method.172 It was also clarified that the use of separate accounting was intended to achieve four purposes, being: (i) to give a strictly territorial scope to the taxation of the branch, ‘not extending beyond the boundaries of the countries concerned’; (ii) to help enforce the equality of treatment of foreigners by placing branches on the same footing as domestic enterprises; (iii) to conform to the usual practice of international business to keep separate accounts for each of their establishments; and (iv) to prevent the concealment of profits or their diversion from one country to another.173

1.114  There were two further differences in the approach of both London and Mexico drafts to the calculation of the profits of an establishment compared to the 1933 treaty. First, the four-step approach to the allocation of the attributable net business income of an establishment (as well as some specific provisions dealing with banking and insurance business) was supplemented by the stipulation that if the rules do not result in a fair allocation of income, the competent authorities may consult to ‘agree upon a method that will prevent double taxation’.174 There was also a second and more important development in the text of the London and Mexico drafts, going beyond the provisions of the model treaty for the allocation of profits. Both London and Mexico models also included provisions requiring ‘corresponding adjustments’ in the event an adjustment of the accounts of an establishment were made based on the arm’s-length (p. 47) standard. This appeared in the form of an addition to the provisions permitting the rectification of unreliable or inadequate accounts and provided that, if such a rectification was made, ‘a corresponding rectification shall be made in the accounts of the establishment in the other contracting states with which the dealings in question have been effected’.175

1.115  With regard to subsidiary companies, the two key aspects of the approach of the 1933 model treaty on the allocation of profits were followed: first, both drafts expressly stated that a subsidiary cannot be regarded as a permanent establishment of the parent company and, second, both drafts contained the early version of the ALP first seen in that 1933 draft, further consolidating the inclusion of the provision in double tax treaties.176 There was no ‘corresponding adjustments’ provision to supplement the inclusion of the ALP rule for inter-company arrangements, which seems anomalous as the case for including it given the inclusion of the early ALP rule would seem at least as strong as for the corresponding separate accounting (branch) rules, where it is included (see above).177

1.116  The Commentary to the London and Mexico drafts spells out the two major consequences that flow from the subsidiary not being regarded as a permanent establishment of the parent company: the country where the subsidiary is located may not tax the parent company (except in relation to any dividends paid by the subsidiary to the parent) and equally the parent company jurisdiction may not tax the profits of the subsidiary (except in relation to dividends or other payments it makes to the parent). The relevant explanation in the Commentary underlines the significance of the separate legal status of the companies as the basis for the approach adopted: ‘These rules follow the principle that a subsidiary constitutes a distinct legal entity and should therefore be taxed separately.’178 Reference is also made in the same discussion in the Commentary to the purpose of the ALP rule, which it describes as indicating the criteria ‘according to which the correctness of the mutual relations between parent and subsidiary companies can be checked so as to avoid abuses resulting in the diversion of profits or losses from one company to the other’.179

J.  The ALP Emergent

1.117  Following the end of the Second World War and the formation of the United Nations, the League of Nations was dissolved in 1946. By this time, the international approach to (p. 48) allocating taxing rights by reference to a bilateral paradigm between residence and source jurisdictions had been largely agreed and a host of individual issues (such as the approach to corporate residence using the central management or incorporation test) had been resolved in a manner familiar to a modern-day reader. The independent enterprise principle had become firmly established, with separate rules and approaches dealing with intra-entity and inter-company situations.

1.118  For intra-entity allocations, discussion of which had dominated this early period of development of the international tax allocation rules, the permanent establishment test had become established as the dominant threshold test in relation to source country taxing rights and, in relation to the attribution of profits, there was also established a clear preference for the separate accounting approach, with a hierarchy of available methods to deliver it. There was also a mechanism to deal with corresponding adjustments.

1.119  To deal with inter-company situations, the earlier open question as to whether subsidiaries should be treated as permanent establishments of the parent company had been decisively resolved in the negative, largely on the grounds of respecting the separate legal form of the subsidiary. The ALP had emerged in a form which is very similar to the rule that operates today, close to a century later. However, there had by this stage been very little discussion on the detailed scope, operation, and purpose of the rule (e.g. on matters such as the reach of the provision and the relationship of the treaty version of the ALP with domestic law of the contracting states).

1.120  With regard to allocation issues generally, the overall approach by the League of Nations reflected a high level of confidence in the solutions adopted to deliver a results-based mechanism for dealing decisively with any diversion of profit. There had been some early hints of the conceptual problems and the practical complexity that would come later, but few, if any, deep concerns had emerged. There had been much discussion of ‘evasion’ (which, from a modern perspective, also included the topic of aggressive tax avoidance) and this included some significant concerns about price manipulation. Although the issue of moveable capital was a key concern, the potential problems arising from interposed or intermediate companies had barely featured.

1.121  From a modern-day perspective, many of the elements in the international tax system that are now seen in some quarters as potentially problematic (e.g. the deferral of taxing rights to residence with limited source country taxing rights, the decision to treat subsidiaries for tax purposes as separate companies in their own right, the relative formality of the test of residence, and the virtually exclusive reliance on the ALP for cross-border allocations for both branches and subsidiaries) had been firmly established. While the merits of the approach adopted in this period to these matters may be debated, there is one point that seems hard to understand, namely the almost complete omission of any discussion of intermediate or interposed related companies and their implications for the conclusions that were reached.180 This seems all the stranger given: (i) the primary focus of concerns (p. 49) about ‘evasion’ in the period related to the movement of capital;181 (ii) the existence of a clear understanding of the potential for tax avoidance in the international arena, including the tax-driven use of foreign companies;182 and (iii) an example referring to the difficulties created by an interposed company featured in the very first League of Nations report on double taxation.183

1.122  In relation to the ALP, it would be some years before the implications of the international tax system created in this period would be fully appreciated.


1  There were some earlier limited examples of states enacting measures to address double taxation, including a Dutch measure dating from 1819 (see J. G. Herndon, ‘The Development of International Reciprocity for the Prevention of Double Income Taxation’, PhD thesis, University of Pennsylvania, Philadelphia, 1932 at p. 11 and generally Chapter X at pp. 146–71; later reproduced as J. G. Herndon, Relief from International Taxation, The Development of International Reciprocity for the Prevention of Double Taxation, Chicago: Callaghan & Co., 1932) and measures to eliminate double taxation by the states of federal unions, such as Germany in 1870 and Switzerland in 1874 (see E. R. Seligman, Double Taxation and International Fiscal Co-operation, New York: Macmillan, 1928, p. 37). The goal of these efforts in the 1920s to address double taxation was to identify a consensus approach that might be adopted by a large number of states.

2  A superb history of the growth of capitalism in the twentieth century is given in: Jeffry A. Frieden, Global Capitalism: Its Fall and Rise in the Twentieth Century, New York: Norton, 2007, from which the trade statistic is taken, p. 9.

3  Frieden (n. 2), p. 16.

4  J. M. Keynes, The Economic Consequences of the Peace, New York: Harcourt Brace and Howe, 1920, p. 12.

5  Great Britain for example incurred the destruction of 10 per cent of its domestic and 24 per cent of its overseas assets, having spent over 25 per cent of its GDP on the war effort between 1915 and 1918—S. Broadberry and M. Harrison, ‘The Economics of World War I: An Overview’, in S. Broadberry and M. Harrison (eds.), The Economics of World War I, Cambridge: Cambridge University Press, 2005, pp. 3–40.

6  See John F. Avery Jones, ‘Sir Josiah Stamp and Double Income Tax’, in John Tiley (ed.), Studies in the History of Tax Law, Vol. 6, London: Hart, 2013, p. 2.

7  The examples are taken from evidence submitted to the 1920 Royal Commission, quoted in Tiley, ed. (n. 6), pp. 2–3.

8  The early work and proceedings of the ICC are captured in a series of ‘Brochures’ (in French and English versions). Individual brochures from this period record discussions or contain reports on individual topics, such as double taxation, customs regulations, and bills of exchange—and contain records of the proceedings of ICC Congresses. The Brochures are held in the ICC archives and have kindly been made available to the authors at the headquarters of the ICC in Paris. The reference made here is to Brochure 25, p. 7. The discussion in the Brochure includes an example of a business being subject to triple taxation.

9  Individual examples involving tax rates of 73.2 per cent and 68 per cent are discussed in Tiley, ed. (n. 6), pp. 3 (at fn. 6) and 16. The OECD would later distinguish between juridical double taxation and economic double taxation. The nature of juridical double taxation is that it involves the same income or capital being taxed in the hands of the same taxpayer by more than one state. Economic double taxation, from which it may be distinguished, involves two different taxpayers being taxed in respect of the same income. OECD efforts to remove double taxation focus in particular on juridical double tax, though the ALP is an exception to this given that it is concerned with economic double taxation. See further OECD Model Double Tax Convention, Introduction, at para. 3 and Commentary to Article 23A and 23B, paras. 1–3.

10  ICC Brochure 25 (n. 8), p. 8.

11  A history of the role of the League of Nations is given in Patricia Clavin, Securing the World Economy: The Reinvention of the League of Nations, 1920–1946, Oxford: Oxford University Press, 2013.

12  The League organized an International Financial Conference in September–October 1920, which first made the call for the League to address the topic of international double taxation. This ultimately led to the creation of the Fiscal Committee, which commenced its work in 1929.

13  The ICC grew out of earlier international congresses of Chambers of Commerce held at Liege (1904), Milan (1906), Prague (1908), Boston (1912), and Paris (1914). The organization conference of the ICC took place at Paris in the spring of 1920 with participation from the five allied countries of Great Britain, France, Italy, Belgium, and the US. See G. L. Ridgeway, The Merchants of Peace, New York: Little Brown & Co., 1959, p. 29. See also, ICC Brochure 20, ‘The Organization of the International Chamber of Commerce’, Feb. 1922, pp. 5–6.

14  This point is expressly recognized in ICC Brochure 25 (n. 8), p. 14. A contemporary account of the developments at the time also suggests that the ICC was the first to call attention to the double taxation issue. See Herndon (n. 1), p. 8.

15  The topic of ‘duplicate taxation’ was the first agenda item for the General Meeting on Finance and subsequently a conference resolution was agreed which called for ‘prompt agreement’ between the governments of the Allied countries to resolve the double taxation problem. See further, Chambre de Commerce Internationale, Congrès Constitutif, Paris, 23–30 June 1920, from the volume, ICC Proceedings Organisation Meeting, Paris, France, 23–30 June 1920.

16  This Committee was Select Committee No. 1 and it was made up of representatives from Great Britain, the Netherlands, and the US. See ICC Brochure 11, p. 61 and ICC Brochure 19, p. 13.

17  See ICC Brochure 11, ‘Double Taxation’. Most of the country reports were published in Brochure 11, where they are included at pp. 7–52 with an explanation at p. 6 of the questionnaire sent to the individual countries. See also, ICC Brochure 12, ‘Special Report by the British National Committee’, which deals with the British tax system.

18  See ICC Brochure 18, ‘Proceedings of the First Congress’ (London, 27 June–1 July), p. 9.

19  Reflecting the profile of the issue in the deliberations of the ICC, the resolution was the first of twenty-seven resolutions adopted at the Congress. See Chamber de Commerce Internationale, Resolutions 1920–31, Vol. 1, London Congress 1921, Resolutions Adopted, pp. 6–7. There had by this time been attempts by some states to enact domestic measures to alleviate double taxation (including the Netherlands in 1893, Belgium in 1906, a limited and temporary measure in the UK in 1916, and the US in 1918), though there was no consistency in the approach adopted. See Mitchell B. Carroll, in IFA (ed.), Cahiers de Droit Fiscal, 1, The Hague: Kluwer, 1939, p. 237.

20  See Herndon (n. 1), pp. 20–2.

21  See ICC Brochure 11, p. 59. These comments were made by W. F. Gephart, representing the American Section. The discussion at the Congress also includes an early example of the difficulties of drafting principles or guidance that can be made acceptable to individual countries in the light of their domestic tax systems, the specific example concerning revisions made to accommodate the position of the US and Great Britain. See ICC Brochure 11, p. 62.

22  See ICC Brochure 23, ICC, Second Congress, Rome (18–25 Mar. 1923), Programme of the Congress and Drafts of Resolutions.

23  ICC, Second Congress, Rome (18–25 Mar. 1923), Brochure 25, ‘Double Taxation’, pp. 8–9. ICC Brochure 25 (published in Mar. 1923) is a summary of the work of the Committee on Double Taxation from the time of the London Congress (June 1921) to the Rome Congress (Mar. 1923). It also includes a commentary on a number of country reports received by the Committee on Double Taxation since the London Congress (see p. 18). The Congress report also recognized the need to develop an agreement between countries as to what constitutes the ‘domicile’ of taxpayers. ICC Brochure 25 (n. 8), p. 9.

24  ICC Brochure 25 (n. 8), p. 10.

25  ICC Brochure 25 (n. 8), pp. 26–7. The principles referred to are those numbered 7–10.

26  By this time, it was clearly recognized that the League of Nations was the appropriate body to deliver the result sought by the ICC. At its meeting of March 1922, the ICC Committee on Double Taxation had already urged the League of Nations to press governments to adopt principles in accordance with those identified at the ICC London Congress. See further, ICC Brochure 25 (n. 8), p. 11.

27  The US delegate to the Rome Congress expressed certain US reservations on the Double Taxation report and asked for time for further study of the topic of double taxation. This was granted with the result that the report was not formally adopted. It was also reported that there were concerns that the proposals were aimed only at alleviating double taxation, not abolishing it. This position reflected the very strong opposition of the British delegation to any form of source-based taxation. See ICC Brochure 32, ‘Proceedings of the Second Congress’ (Rome, 18–24 Mar. 1923), p. 136; ICC Brochure 34 ‘Double Taxation’ (A Survey of the Work of the ICC since the Rome Congress), p. 6. See also, Michael J. Graetz and Michael M. O’Hear, ‘The “Original Intent” of US International Taxation’, Duke Law Journal 46 (1997), 1071–2.

28  See further, the report prepared for the Third Congress in Brussels in June 1925, ICC Brochure 34 (n. 27), p. 6.

29  ICC Brochure 34 (n. 27), pp. 7–10.

30  See resolutions 1 and 2 regarding the double taxation of income in ICC Brochure 34 (n. 27), pp. 7–8. The wording quoted is not clear but might be taken to suggest, by implication of a stand-alone approach to rewarding branch activity, that the earlier profit split approach had by now been dropped.

31  ICC Brochure 34 (n. 27), resolution 3, p. 2. A call was also made for the development of a uniform definition of fiscal residence (resolution 4).

32  The resolutions were adopted by the Council of the ICC on 7 Mar. 1924, though with an Italian reservation to the effect that existing tax systems were too complex for a single solution and that what was proposed was acceptable for personal (income) taxes but not for impersonal taxes (i.e. taxes imposed on designated forms of income, without regard to the characteristics of the owner) where the origin principle (i.e. taxation at source) ought to be more equitable (ICC Brochure 34 (n. 27), p. 9). The discussion between personal and impersonal taxes is discussed further below.

33  There was a continuing general interaction between the League of Nations and the ICC at the time. See, for example, Ridgeway (n. 13), pp. 83–106. In relation to the double tax issues, ICC Brochure 25 (n. 8), p. 16, notes that the ICC ‘has kept continually in contact with the financial commission of the League of Nations’. There are very many examples of the ongoing cooperation between the ICC and the League of Nations on double taxation issues. For example, the League of Nations used the ICC as the mechanism to obtain the views of taxpayers on the topic of double taxation at the time it was also obtaining the views of tax authorities through the work of its Technical Experts; the Technical Experts themselves requested the ICC to send a delegation to their meeting in April 1924 to give their views on double tax and also give some background on the work of the ICC Committee on Double Taxation; the League of Nations reports on taxation mention the input of the ICC a number of times; the ICC was represented at virtually all of the meetings of the League of Nations in this early period. See further, ICC Brochure 34 (n. 27), p. 10 for a general explanation of this interaction and the League of Nations 1925 Report, Double Taxation and Tax Evasion—Report and Resolutions submitted by the Technical Experts to the Financial Committee, Document F212, Geneva: League of Nations, Feb. 1925, pp. 7–8.

34  See Herndon (n. 1), p. 34.

35  The League of Nations had been requested to develop an ‘international understanding’ on the topic of double taxation by a resolution of the International Financial Conference held in Brussels in 1920. See Herndon (n. 1), pp. 41–2.

36  The topic of tax evasion was initially referred to the League of Nations by the April 1922 meeting in Genoa of the International Economic Conference. See Double Taxation and Tax Evasion, Report Presented by the Committee of Technical Experts on Double Taxation and Tax Evasion, Geneva: League of Nations, 12 Apr. 1927, C.216.M.85, at p. 5. From 1925, the topics of double taxation and tax evasion were discussed in tandem: the next three reports compiled by the League of Nations and appearing in 1925, 1927, and 1928 were each entitled Double Taxation and Tax Evasion and each contained extensive discussion of both topics. The League of Nations Fiscal Committee also worked on the topic, though it acknowledged that work on combatting double taxation, rather than combatting tax evasion, had been its chief focus. See League of Nations Fiscal Committee, Work of the Fiscal Committee during its Sixth Session, Geneva: League of Nations, 1936, C.450.M.266.1936.II.A, at p. 2.

37  Report on Double Taxation submitted to the Financial Committee—Economic and Financial Commission Report by the Experts on Double Taxation—Document E.F.S.73. F.19, Geneva: League of Nations, 5 Apr. 1923 (hereafter, 1923 Report on Double Taxation). There is a contemporary account of the Report in Herndon (n. 1), pp. 41–55. More modern perspectives are given by Graetz and O’Hear (n. 27), pp. 1021, 1074–80 and Tiley, ed. (n. 6), pp. 16–26. The Report does not make for easy reading and at times seems contradictory (Part I seems to emphasize the importance of the residence principle yet Part II develops an economic allegiance principle which leads to the assignment of taxing rights by a source approach). This seems largely due to the manner in which the Report was compiled by the authors. See Sunita Jogarajan, ‘Stamp, Seligman and the Drafting of the 1923 Experts’ Report on Double Taxation’, World Tax Journal 5(3) (2013).

38  For an interesting discussion of the ‘economic allegiance’ approach and its relationship with the benefits theory of taxing jurisdiction, see Reuven S. Avi-Yonah, ‘All of a Piece Throughout: The Four Ages of U.S. International Taxation’, Va. Tax Rev. 25(2) (2005), 317–23.

39  The Report notes, however, that the different stages of development of country tax systems means that countries have different perspectives on first principles. In particular, the Report recognizes that the approach of the more developed countries, such as the US, Germany, Great Britain, and Holland, was to apply progressive income taxes on total (global) income and for this approach an exclusively residence basis of tax was required. On the other hand, countries with less developed fiscal systems, which taxed specific streams of income or specific objects of wealth, would be more likely to apply a source or origin approach. See 1923 Report on Double Taxation (n. 37), pp. 26, 45–6.

40  The Report does recognize the imprecision of the approach adopted: ‘To allocate the exact proportion of economic allegiance to origin or domicile in each particular category is well-nigh impossible. Such an attempt would savour too much of the arbitrary’ (1923 Report on Double Taxation (n. 37), p. 39). The ‘classification and assignment’ approach adopted had previously been used in the first multilateral treaty which had been entered into by Austria, Hungary, Italy, Poland, Yugoslavia, and Rome in 1921.

41  1923 Report on Double Taxation (n. 37), pp. 45 and 49.

42  1923 Report on Double Taxation (n. 37), p. 40.

43  The primacy of the residence approach also reflected the assumption that residence (‘creditor’) countries were the providers of capital and know-how and so made business possible, which also explains the wish to limit taxation in source (‘debtor’) countries. See Bret Wells and Cym H. Lowell, ‘Tax Base Erosion and Homeless Income: Collection at Source Is the Linchpin’, Tax Law Review (2012), 547. A further factor in the approach was the difficulty in identifying profits that were properly allocable to the source state. See 1923 Report on Double Taxation (n. 37), p. 46.

44  The economists were clearly aware of the difficulties of arriving at a single basis of taxation for all countries. Earlier discussions between them had considered the example of Australia, which had plentiful natural resources, and Great Britain, as a financial centre, as examples of countries that would not relinquish taxation on the basis of source and residence respectively. See Jogarajan (n. 37), 7.

45  1923 Report on Double Taxation (n. 37), pp. 7–16.

46  The rejection of source-based taxation was based on its perceived theoretical weakness, namely that it is not based on the taxpayer’s full ability to pay or ‘faculty’. The report refers to ‘a development away from localised ideas and from the earlier stages of economic thought typified by strict adherence to the principle of origin’ (1923 Report on Double Taxation (n. 37), pp. 18, 51). A similar preference in favour of no origin taxes was reached by the ICC Committee on Double Taxation in drawing up its revised resolutions in March 1924 but in that instance also there was a recognition that there were already far too many instances of source taxation being levied by states for an exclusively residence basis to be a realistic possibility, as exemplified by an Italian reservation on the 1924 ICC proposal disputing the validity of a residence basis approach. For a contemporary defence of source country taxation, see Michael J. Graetz, ‘Taxing International Income—Inadequate Principles, Outdated Concepts, and Unsatisfactory Policy’, Tax Law Review 54 (2001), 298.

47  See Extract of the Report of the Financial Committee to the Council of the League of Nations, which is included in the note entitled Double Taxation and Tax Evasion which is attached to the Technical Experts Report, Double Taxation and Tax Evasion—Report and Resolutions Submitted by the Technical Experts to the Financial Committee, Geneva: League of Nations, 1925, Document F.212, at p. 1 (hereafter, 1925 Technical Experts Report).

48  The preliminary note to the 1925 report notes that the Financial Committee of the League of Nations had considered that, to arrive at a real solution to the problem of double taxation, it was essential to obtain the opinion of representatives of certain governments and it would be better still if they met to work on an agreement to enable common action to be taken. The report sets out their task as the representatives saw it. ‘We have regarded our task as being that of technical experts endeavouring to prepare the best possible system for remedying the evils of double taxation and tax evasion’ (1925 Technical Experts Report (n. 47), pp. 2 and 6).

49  A more detailed explanation and discussion of impersonal and personal taxes is given in W. H. Coates, ‘Double Taxation and Tax Evasion’, Journal of the Royal Statistical Society 88(3) (May 1925), 405ff. The article is critical of the approach taken by the Technical Experts, on the grounds that the distinction between personal and impersonal taxes rests on a misapprehension and is in any event bogus given the economically similar effects of the taxes.

50  1925 Technical Experts Report (n. 47), pp. 14–15.

51  1925 Technical Experts Report (n. 47), p. 31.

52  1925 Technical Experts Report (n. 47), p. 16.

53  1925 Technical Experts Report (n. 47), p. 16.

54  The treaty between Austria and Czechoslovakia is also referred to as providing methods for the flat-rate computation of the profits of firms in the case of purchasing and selling activities. 1925 Technical Experts Report (n. 47), p. 16.

55  The resolution also provided for the relevant tax authorities to be able to request the taxpayer to hand in balance sheets and other financial data and relevant documents to facilitate the apportionment (1925 Technical Experts Report (n. 47), p. 31).

56  For example, see the discussion of Germany, France, and Spain in Taxation of Foreign and National Enterprises, Vol. I, Studies of the Tax Systems and the Methods of Allocation of the Profits of Enterprises Operating in More than One Country in France, Germany, Spain, the United Kingdom and the United States of America, Geneva: League of Nations, 1932, C.73.M.38.1932.II.A, at pp. 22, 37 (hereafter, Carroll Vol. I). A few German and Italian treaties adopted the same rule from 1925 to as late as 1940. J. F. Avery Jones et al., ‘The Origins of Concepts and Expressions Used in the OECD Model and their Adoption by States’, IBFD Bulletin, June 2006, 220, file note 240, p. 237.

57  1925 Technical Experts Report (n. 47), pp. 19, 33. The credit model was by this time followed in the US following the introduction of the US foreign tax credit in 1918 and its further refinement in 1921, and under which Americans could claim a reduction, or credit, against their US tax bill of the amount of taxes paid to other countries.

58  1925 Technical Experts Report (n. 47), pp. 19 and 32.

59  1925 Technical Experts Report (n. 47), pp. 21 and 34. The approach adopted was consistent with the view supporting the primacy of a residence approach to taxing rights—i.e. that it was the provision of capital and know how that fuelled entrepreneurial activity and so the essential test was intellectual control by the directors, not the more routine activities on the ground.

60  1925 Technical Experts Report (n. 47), pp. 27–8. The discussion first clarifies that tax evasion in concept is not synonymous with the exportation of capital as capital may be exported for various reasons, though the report proceeds to define tax evasion as effected particularly by means of the flight of capital which allows the interested persons to escape taxation which is legally due. See p. 22.

61  This point is again emphasized in the following report of 1927. See Double Taxation and Tax Evasion, Report Presented by the Committee of Technical Experts on Double Taxation and Tax Evasion, Geneva: League of Nations, 1927, C.216.M.85, p. 23 (hereafter, 1927 Technical Experts Report).

62  In addition to the initial seven countries of Belgium, Czechoslovakia, France, Great Britain, Italy, Netherlands, and Switzerland were now added Argentina, Germany, Japan, Poland, the US, and Venezuela. A delegation of the ICC assisted at these meetings.

63  1927 Technical Experts Report (n. 61).

64  The permanent establishment concept was developed in the mid-nineteenth century in the German states to prevent double taxation among the Prussian municipalities, though the term was not used in German tax law until 1885. See Arvid Skaar, Permanent Establishment, Series on International Taxation 13, Kluwer: Deventer, 1991, pp. 72, 74.

65  The term ‘permanent establishment’ is derived from the German ‘Betriebstätte’, which was first used in Prussian tax law in 1885 and first appears in a treaty between Austria/Hungary and Prussia in 1899. See Avery Jones et al. (n. 56), pp. 233–4. The uncertainty of the distinction between impersonal and personal taxes is discussed in Coates (n. 49), pp. 408–9 and 425.

66  1927 Technical Experts Report (n. 61), pp. 10–11. The Commentary accompanying the draft treaty also clarified that ‘The word “undertaking” must be understood in its widest sense so as to cover all undertakings … without making any distinction between natural and legal persons’. See Article 10 of Draft I, p. 15. The term ‘undertaking’ was used interchangeably with the word ‘enterprise’ in the League of Nations Models (Avery Jones et al. (n. 56), p. 225).

67  1927 Technical Experts Report (n. 61), pp. 11 and 15.

68  As discussed further below, the consideration of this matter continued over the first three meetings of the League of Nations Fiscal Committee, from 1929 to 1931. The exemption for such independent agents was important as a number of countries taxed foreign enterprises on sales made through local brokers and commission agents. See Avery Jones et al. (n. 56), p. 237.

69  1927 Technical Experts Report (n. 61), Article 10 of Draft I, p. 11.

70  The deduction would operate only to the extent that the residence state did not itself levy impersonal taxes on taxpayers domiciled in that state. The amount of the deduction would be the lower of the residence tax attributable to the foreign income or the amount of the source state tax. There was also an optional mechanism to limit the overall amount of deductions by reference to the total personal tax in the state of residence. See 1927 Technical Experts Report (n. 61), Draft I, Article 10, p. 11.

71  See 1927 Technical Experts Report (n. 61), Draft I, Articles 13 and 14, p. 12.

72  See 1927 Technical Experts Report (n. 61), p. 23.

73  The single taxation principle would explain the work in parallel of the League of Nations on tax evasion as well as on addressing double taxation but the existence of the principle has proved highly contentious. See further, Reuven S. Avi-Yonah, ‘Who Invented the Single Tax Principle?: An Essay on the History of U.S. Treaty Policy’, N.Y.L. Sch. L. Rev. 59 (2014–15), 309, at 310–12. The single taxation principle also seems to have been accepted by the ICC. See ICC Brochure 25 (n. 8), p. 8.

74  See 1927 Technical Experts Report (n. 61), pp. 31–3.

75  The Meeting was called as a result of the League of Nations resolution of 17 June 1927. See Double Taxation and Tax Evasion: Report Presented by the General Meeting of Government Experts on Double Taxation and Tax Evasion, Geneva: League of Nations, 1928, Document C.562.M.178, p. 5 (hereafter, 1928 Report of Government Experts). It was a lengthy meeting, lasting from 22 to 31 October 1928.

76  See 1928 Report of Government Experts (n. 75), p. 5. A multilateral treaty had been proposed by the US delegate, T. S. Adams, and in 1929 he accurately predicted the disadvantages of a bilateral treaty approach being adopted, namely undue complexity and a myriad of different standards from one treaty to another. See T. S. Adams, International and Interstate Aspects of Double Taxation 22 Nat’l Tax Ass’n Proc. 193 at 195 (1929).

77  The original Technical Experts draft from 1927 was renumbered as I-A and then in effect translated into the two additional versions, being I-B and I-C.

78  See 1928 Report of Government Experts (n. 75), discussion of Treaty I-A, pp. 10 and 15.

79  Germany and Italy were treating subsidiaries as permanent establishments in their domestic law and preserving this treatment in double tax treaties until as late as 1940. See Avery Jones et al. (n. 56), p. 237. It may have remained an issue until 1958 given the inclusion of the provisions of Article 5(7) in the first report of the OEEC (see further Chapter 2).

80  See Minutes of the Seventeenth Session of the General Meeting of Government Experts on Double Taxation and Fiscal Evasion, Geneva: League of Nations, 22–31 Oct. 1928, as discussed in Herndon (n. 1), pp. 197–9. There is also a detailed account of the discussions in Wells and Lowell (n. 43), pp. 556–7.

81  The new text, I-B, drafted by the US and British representatives for states taxing by reference to domicile, provides that the state of domicile will levy its tax on all kinds of income. The draft still permits source taxation, though this is now limited to certain designated instances (with the wider concept of personal taxes now removed), including in the case of industrial, commercial, or agricultural income, where the ‘permanent establishment’, and ‘income produced’ tests remain applicable. Where such taxation was levied, the draft called for corresponding tax credits to be given in the residence territory. The main difference to the earlier Technical Experts’ text is that this version assigns income from transferable securities by priority to the state of domicile. The second new text, I-C, which is intended for countries with different tax systems, differs from the Technical Experts’ version chiefly in its treatment of the taxation of income from movable capital on which taxing rights are in principle allocated to the state of domicile, though there are also provisions to address double taxation in the event of source taxation being imposed. See 1928 Report of Government Experts (n. 75), discussion of Treaty I-A, pp. 16–21.

82  See Graetz and O’Hear (n. 27), pp. 1082–6.

83  Bret Wells and Cym H. Lowell, ‘Income Tax Treaty Policy in the 21st Century: Residence vs. Source’, Columbia Journal of Taxation 5 (2013), 26. See also Graetz and O’Hear (n. 27), pp. 1023 and 1086. The influence of the model treaties on actual treaties was constantly monitored by the League of Nations and the merits of the approach were summarized in 1935 by the Fiscal Committee in commenting on their value in facilitating the negotiation of tax treaties: ‘The existence of model draft treaties of this kind has proved of real use in such circumstances in helping to solve many of the technical difficulties which arise in such negotiations. This procedure has the dual merit that, on the one hand, in so far as the model constitutes the basis of bilateral agreements, it creates automatically a uniformity of practice and legislation, while, on the other hand, inasmuch as it may be modified in any bilateral agreement reached, it is sufficiently elastic to be adapted to the different conditions obtaining in different countries or pairs of countries’ (League of Nations Fiscal Committee, Report to the Council on the Fifth Session of the Committee, Geneva: League of Nations, 1935, C.252.M.124.1935.II.A, p. 4).

84  1928 Report of Government Experts (n. 75), p. 26.

85  1928 Report of Government Experts (n. 75), discussion of Treaty I-A, pp. 34–6.

86  The issues previously identified by the General Meeting of Government Experts were, in addition to the allocation of profits of industrial and commercial enterprises, the taxation of income from patents and from authors’ rights; measures for the avoidance of double taxation of trusts and companies owning large amounts of transferable securities. See 1928 Report of Government Experts (n. 75), pp. 34–6.

87  It is suggested in Wells and Lowell (n. 83), p 27, that an impetus for this work on apportionment came from concerns about the terms of the Hungary–Poland double tax treaty in May 1928 given that that treaty adapted version I-A of the 1928 model to deliver an emphatic preference to source-based taxation and went further in providing that the combined income of a non-resident company and its subsidiaries would be apportioned using a profit split methodology based on gross revenues.

88  The matter was raised as a result of a resolution from the ICC Congress of July 1929: ‘The fact of an undertaking having business relations with a foreign country through a local company holding some or all of the shares does not imply that the undertaking has a permanent establishment in that country’ (League of Nations Fiscal Committee, Report to the Council on the Work of the First Session of the Committee, Geneva: League of Nations, 17–26 Oct. 1929, C516.M.175.1929.II. p. 5 (hereafter, 1929 Report to the Council)).

89  The ultimate intention of the Fiscal Committee was to produce provisions for a multilateral treaty to deal with the problems of double taxation and in that regard mention was made specifically of the intention to produce a multilateral instrument to regulate the taxation of industrial and commercial enterprises that conduct business in more than one country. See 1929 Report to the Council (n. 88), p. 6.

90  The consideration of this matter continued for the first three meetings of the Fiscal Committee, from 1929 to 1931, and included an early version of the dependent agent test, which was later, in 1933, adopted in the League’s Model Convention on the Allocation of Profits. Given the high level of focus on arrangements involving commissionaires some eighty-five years later in the BEPS project, it is also of note that commissionaires were considered in the first meeting of the Fiscal Committee and readily accepted as falling within the definition of autonomous agents that were excluded from being deemed permanent establishments. See 1929 Report to the Council (n. 88), pp. 3–4.

91  League of Nations Fiscal Committee, Report to the Council on the Work of the Second Session of the Committee, Geneva: League of Nations, 22–31 May 1930, C.340.M.140.1930.II, p. 14 (hereafter, 1930 Report to the Council). Various country source taxation rules are also discussed in Mitchell B. Carroll, ‘Allocation of Business Income: The Draft Convention of the League of Nations’, Columbia Law Review 34 (1934), 474–6.

92  1930 Report to the Council (n. 91), p. 5.

93  A grant of $90,000 was secured by T. S. Adams for the purposes of assisting the Fiscal Committee in its work on double taxation. The plan was for the League of Nations Secretariat to hire some specialist staff to research existing laws, country practices, and accounting methods. The scope of the work was to include an analysis of the use of the methods of separate accounting and fractional apportionment. See 1930 Report to the Council (n. 91), pp. 7–8.

94  Four areas for inclusion in such a multilateral treaty are suggested and these relate to (i) certain categories of income to be taxed in the state of domicile of the recipient (annuities, authors rights, interest and income of workers who live on one side of a frontier but work on the other); (ii) salaries of officials working abroad, etc.; (iii) land and houses—to be taxable where situated; and (iv) allocation of profits of companies.

95  1930 Second Session Report to the Council (n. 91), pp. 8–9.

96  League of Nations Fiscal Committee, Report to the Council on the Work of the Third Session of the Committee, Geneva: League of Nations, 29 May–6 June 1930, C.415.M.171.1931.II.A, pp. 3–5 and 8–15 (hereafter, 1930 Third Session Report to the Council). The report of the Third Session includes the original and an amended version of this draft. The categories of income covered are those suggested at the previous meeting (namely, (i) certain categories of income to be taxed in the state of domicile of the recipient (annuities, authors rights, interest and income of workers who live on one side of a frontier but work on the other); (ii) salaries of officials working abroad; (iii) land and houses—to be taxable where situated; and (iv) allocation of profits of companies), together with interest derived from mortgages.

97  1930 Third Session Report to the Council (n. 96), pp. 3 and 7–8. The position of the British is not expressly referred to in the explanation but Great Britain was the most vocal state in arguing against source taxing rights.

98  The initial draft of the proposal prepared by a subcommittee had sought to limit the application of the rule to cases where the enterprise was being carried on by a ‘company (or other association having a legal existence of its own)’ but this reference was removed in the version of the draft agreed by the full Fiscal Committee. See 1930 Third Session Report to the Council (n. 96), pp. 3, 11, and 13.

99  1930 Third Session Report to the Council (n. 96), pp. 9–14.

100  1930 Report to the Council (n. 91), p. 5.

101  The summary of the country reports was prepared by T. S. Adams, the US delegate, and was included in the report of the second meeting of the Fiscal Committee. See 1930 Report to the Council (n. 91), pp. 5 and 10–17. It is argued by Langbein that the importance of formulary methods at that time was not fully recognized. See S. I. Langbein, ‘The Unitary Method and the Myth of Arm’s Length’, Tax Notes 30 (1986), 625, at 632–3.

102  1930 Report to the Council (n. 91), pp. 10–17. Various country approaches to dealing with allocation issues were also discussed in the earlier 1925 Technical Experts Report (n. 47), pp. 15–16.

103  A detailed explanation of separate accounting is contained in the study Taxation of Foreign and National Enterprises (see n. 104): ‘The term separate accounting does not refer to a single well-defined method, but rather to a whole group of methods with a common aim. Indeed, separate accounting is only a point of view or an avenue of approach to the general problem of allocating business profits on a geographical basis. The accounting methods by which the allocation of profit may be accomplished must be as diverse as the conditions within the different industries are varied. Separate accounting rests on the assumption that the different branches of an enterprise are separate business units which may be treated for accounting purposes practically as independent concerns’ (Ralph C. Jones, Taxation of Foreign and National Enterprises, Vol. V, Allocation Accounting for the Taxable Income of Industrial Enterprises, C. 425 (c) M.217 (c)1933.II.A, Geneva: League of Nations, 1933, ch. 1, para. 13). A later explanation of fractional apportionment is that ‘Under this method, the earnings of each establishment are computed as a proportion of the entire profits of the enterprise to which the establishment belongs, on the basis of the general balance sheet and profit and loss account of the enterprise. Such fractional apportionment may be unlimited or limited. In the first case, it takes as its starting point the total income derived by the enterprise as a whole from all sources. In the second case, reference is made only to that part of the total profits of the enterprise which is derived from transactions in which a part has been taken by the establishment whose share in the total profits is to be determined’ (League of Nations Fiscal Committee, London and Mexico Model Tax Conventions: Commentary and Text, Geneva: League of Nations, 1946, C.88.M.88.1946.II.A., p. 20) (hereafter, Report on London and Mexico Models). At the time the work was commissioned, both separate accounting and fractional apportionment were accepted methods of apportioning profits. See Michael Kobetsky, International Taxation of Permanent Establishments, Cambridge: Cambridge University Press, 2011, p. 132.

104  Carroll Vol. I (n. 56).

105  1930 Third Session Report to the Council (n. 96), p. 6.

106  All of the country reports, and all later country reports, are organized in the same manner, containing a Part I on general description of the country’s income tax system; a Part II on methods of taxing foreign and national enterprises; and Part III on methods of allocating taxable income. In total, the three-year survey covered thirty-five states (Carroll (n. 91), p. 473).

107  See Mitchell B. Carroll, Taxation of Foreign and National Enterprises, Vol. IV, Methods of Allocating Taxable Income, Geneva: League of Nations, 1933, C.425(b).M.217(b).II.A (hereafter, Carroll Vol. IV), and Ralph C. Jones, Taxation of Foreign and National Enterprises, Vol. V, 1933, C.425 (c) M. 217 (c) Geneva: League of Nations, 1933, II.A.21 (hereafter, Jones Vol. V). The work was completed in time to be considered at the fourth session of the Fiscal Committee in June 1933.

108  While the overwhelming bulk of the discussion is given over to branch allocation methodologies, there is also some discussion of instances where separate companies may together be regarded as an economic unit for tax purposes. See, for example, the comments on Spain in Carroll Vol. I (n. 56), pp. 37–8 and 50.

109  See Carroll Vol. I (n. 56), p. 251, which notes that US companies operating abroad also generally used subsidiary companies. However, some branches of foreign enterprises clearly did exist, and raised allocation issues, in the US (Carroll Vol. IV (n. 107), pp. 48–9).

110  It was clear that Mitchell B. Carroll saw the exercise as bringing together as much as possible of the existing approaches: ‘Obviously, a regime for the allocation of income as between countries must take into consideration these divergent or conflicting principles and provide for the allocation or apportionment of items of income or expense to the various jurisdictions concerned therewith. It is impossible to avoid the task of synthesising as far as possible the underlying principles of the different systems and the methods of taxing the principal items of income’ (Carroll Vol. IV (n. 107), p. 14).

111  Carroll Vol. IV (n. 107), p. 169. His comments on p. 170 of Vol. IV suggest that he had in mind as the appropriate definition of fiscal domicile the ‘real centre of management test’, as adopted by the General Meeting of Government Experts in 1928.

112  For example, his suggestions included taxing income from real estate in the country in which the real estate is located, taxing compensation for services in the country in which the services were rendered, etc. (Carroll Vol. IV (n. 107), p. 173).

113  Mitchell B. Carroll viewed this test to be in line with the approach of most countries to income from industrial or commercial profits (Carroll Vol. IV (n. 107), pp. 172, 174). This followed from his finding in relation to the legal principles for allocating business profits that, ‘Although the wording of the basic provisions concerning liability varies greatly from country to country, broadly speaking there are three general precepts with regard to a foreign enterprise: (i) that it is taxable on profits from carrying on a trade or business within the territory of the State, (ii) that it is taxable in respect of income attributable to a permanent establishment in such territory, and (iii) that it is liable in respect of income from sources within such territory. The first prevails quite generally in countries of the British Commonwealth of Nations, the second in continental European countries and their dependencies, and the third in the United States of America. Interpretations vary as to what constitutes trading within the country, or a permanent establishment, or whether income from certain transactions is from sources within the country’ (Carroll Vol. IV (n. 107), p. 18).

114  Carroll Vol. IV (n. 107), pp. 109, 176. The wording referring to the ‘same circumstance and conditions’ would later be adapted to formulate the independent enterprise principle for branches as is discussed further below. It is argued by Langbein that Carroll drew the wrong conclusions on prevailing country practice. See S. I. Langbein, ‘The Unitary Method and the Myth of Arm’s Length’, Tax Notes 30(7) (1986), 625–81.

115  Carroll Vol. IV (n. 107), pp. 176–7. Subsequently, he expressed the point in even stronger terms: ‘Under the United States Constitution, respect for the corporation as a legal entity separate from its shareholders is the basic principle on which our economic structure rests, not only at home but also abroad’ (Mitchell B. Carroll, ‘Income Tax Conventions as an Aid to International Trade and Investment’, International and Comparative Law Bulletin 6(3) (July 1962), 16, at 24).

116  Carroll Vol. IV (n. 107), p. 60.

117  Carroll Vol. IV (n. 107), p. 177.

118  Carroll Vol. IV (n. 107), p. 110. The approach adopted in practice by Austria seems to have applied in a number of other states. See, for example, the discussion on the assessment of tax of companies in Italy in Taxation of Foreign and National Enterprises, Vol. II, C.425 (a).M.217(a).1933.II.A, country report on Italy, Part I. A variant of this approach is the concept of ‘abnormal management decision’, as applied in France. See Avery Jones et al. (n. 56), p. 243.

119  Carroll Vol. IV (n. 107), p. 177. This approach of making readjustments for tax purposes by reference to the pricing of independent enterprises was his preferred option of the three alternatives that were generally used by states to counter situations where the true profits of the subsidiary could not be readily identified or had been diverted to the foreign parent. The other two approaches were the approach of taxing the parent in the name of the subsidiary as agent (as in the UK) and the approach of assessing on the basis that the subsidiary and parent together form an economic unity (as in Spain). Carroll Vol. IV (n. 107), pp. 109–10.

120  Carroll Vol. IV (n. 107), p. 177.

121  Carroll Vol. IV (n. 107), pp. 37, 45–6.

122  Carroll Vol. IV (n. 107), pp. 47, 88, and especially 188–90.

123  Carroll Vol. IV (n. 107), p. 189. The same issue had been referred to in the 1923 economists report (see 1923 Report on Double Taxation (n. 37), p. 46). The difficulties of data from outside the source jurisdiction being used by both the source and the residence tax authorities as the basis for taxation are referred to by the ICC (ICC Brochure 25 (n. 8), p. 10) as well as in Carroll (n. 91), 486. Thinking on the desirability of obtaining information from outside the source jurisdiction has changed markedly in recent years, a development particularly reflected in the outcome of the OECD’s BEPS project.

124  See Carroll Vol. IV (n. 107), p. 47.

125  See Carroll Vol. IV (n. 107), pp. 54–5.

126  Carroll Vol. IV (n. 107), p. 178. The discussion also contemplates a role for the arm’s-length approach in resolving difficult issues of allocation on the basis it is a ‘fair’ approach to the state and the taxpayer. Carroll Vol. IV (n. 107), pp. 183–5.

127  Carroll (n. 91), 482 and 486.

128  Carroll Vol. IV (n. 107), p. 202. As will be seen in the discussion that follows, the formulation was taken up and slightly amended in the League of Nations 1933 Model Treaty on the Allocation of Profits. A substantially similar approach is taken by the existing Article 7(2) of the OECD Model Convention.

129  Carroll Vol. IV (n. 107), p. 202. The two approaches are also discussed in Carroll (n. 91), pp. 481–4.

130  The two methods reflect different approaches that may, depending upon the circumstances, still be relevant today. The former approach would result in more of the profit being allocated to the centre than would normally be the case where the second method is used.

131  Carroll Vol. IV (n. 107), pp. 178–83 and 191.

132  Carroll Vol. IV (n. 107), pp. 202–3.

133  Unlike the previous four volumes, this volume is neither written nor supervised by Mitchell B. Carroll. The author is Ralph C. Jones, Professor of Accounting, Yale University. Jones Vol. V (n. 107).

134  Jones Vol. V (n. 107), Chapter II, para. 19.

135  Jones Vol. V (n. 107), Chapter III, para. 42. There are countless references to the pricing of third party transactions. See, for example, Jones Vol. V, paras. 48, 50, 51, 57, 97.

136  The work also contains what seems to be the first reference to ‘transfer pricing’. See Carroll Vol. IV (n. 107), p. 47 and Jones Vol. V (n. 107), para. 108.

137  By 1946 the Fiscal Committee suggested a further study be undertaken to determine a comprehensive set of rules for allocating the business income of international enterprises carrying on business in more than one country. League of Nations Fiscal Committee, Report on the Work of the Tenth Session of the Committee, Geneva: League of Nations, 1946, C.37.M.37.II.A, p. 10.

138  Carroll Vol. I (n. 56), pp. 21 and 51. See also similar concerns where branch operations are incorporated in a foreign jurisdiction. Carroll Vol. IV (n. 107), p. 157.

139  This was the case in four out of the five countries surveyed in the first phase of work. Carroll Vol. I (n. 56), p. 21.

140  Carroll Vol. I (n. 56), p 22. See also Carroll Vol. IV (n. 107), p. 119 where pricing abuse is noted as frequent and dependent on the relative rate of customs duty as compared with the rate of income tax.

141  Carroll Vol. I (n. 56), p. 51. The same theme is referred to elsewhere, e.g. Carroll Vol. IV (n. 107), p. 12.

142  See Carroll Vol. I (n. 56), pp. 35, 37–46, 53 and Carroll Vol. IV (n. 107), pp. 192, 197. The discussion in the fifth volume of the study also brings out the required diversity of allocation methods, according to the circumstances of the taxpayer and business sector. See Jones Vol. V (n. 107), Chapter I, paras. 13–14 and Chapter II, para. 19.

143  The draft was the subject of some minor changes (relating to the amendment to its provisions on banks and the option of including a provision involving formulary apportionment dealing with insurance business) made at the subsequent meeting of the Fiscal Committee in 1935, and again at the 1936 meeting (which reflected the confirmation that the same basic approach should be applied to the allocation of assets for the purposes of taxes on property and capital).

144  League of Nations Fiscal Committee, Report to the Council on the Fourth Session of the Committee, Geneva: League of Nations, 1933, C.399.M.204.II.A, p. 2 (hereafter, 1933 Fourth Session Report to Council); League of Nations Fiscal Committee, Report to the Council on the Fifth Session of the Committee, Geneva: League of Nations, 12–17 June 1935, C.252.M.124.1935.II.A.9, p. 3.

145  1933 Fourth Session Report to Council (n. 144), Annex, Draft Convention, Article 3, p. 4. The wording suggested initially by Mitchell B. Carroll was that the branch or permanent establishment should be regarded as ‘if it were an independent enterprise carrying on similar activities under similar conditions’ (Carroll Vol. IV (n. 107), p. 202).

146  1933 Fourth Session Report to Council (n. 144), Annex, Draft Convention, Article 3, p. 4.

147  1933 Fourth Session Report to Council (n. 144), p. 2. A special rule for banking and financial enterprises (in Article 4 of the Draft Convention) was considered to be necessary as the generally accepted rules for the allocation of income of this class of enterprise was at variance with the general approach proposed in the Draft Convention. The specific Article proposed contained a number of detailed provisions relating to deposits, dealing in financial products and securities, and services such as underwriting.

148  1933 Fourth Session Report to Council (n. 144), Annex, Draft Convention, Article 3, p. 4.

149  1933 Fourth Session Report to Council (n. 144), Annex, Draft Convention, Article 3, p. 4.

150  1933 Fourth Session Report to Council (n. 144), Annex, Draft Convention, Article 3, p. 4.

151  1933 Fourth Session Report to Council (n. 144), Annex, Draft Convention, Protocol, p. 6. This provision was included in a number of early treaties from 1931. See Avery Jones et al. (n. 56), p. 238, where it is also noted that the present wording of Article 5(7) of the OECD model, providing broadly that the control relationship between a parent and subsidiary does not cause a permanent establishment to exist, dates from the OEEC First Report of 1958.

152  1933 Fourth Session Report to Council (n. 144), Annex, Draft Convention, Protocol, at p. 6. The rule was derived from the discussion of the terms ‘autonomous agent’ and ‘permanent establishment’ four years earlier. 1929 Report to the Council (n. 88), p. 4.

153  1933 Fourth Session Report to Council (n. 144), Annex, Draft Convention, Article 5, p. 5. The provision was moved and renumbered as Article 6 of the 1935 version of the Model.

154  1933 Fourth Session Report to Council (n. 144), Annex, Draft Convention, Commentary to Article 5, p. 7.

155  The immediate derivation however is from Article IV of the US–France tax treaty of 1930, which was ratified in 1932. That treaty (in relation to which Carroll had been one of the US negotiators) had already, and for the first time, reflected the principles of section 45 in an international tax treaty and so provided a ready model for Article 5 of the Model of 1933. See Carroll, ‘Income Tax Conventions’ (n. 115), 20.

156  Section 45, which became section 482 of the Internal Revenue Code in 1954, was based on section 240(d) of the Revenue Act of 1921 (renumbered to section 240(f) of the Revenue Act of 1926), which had been enacted to prevent income shifting to foreign affiliates. The 1921 hearings of the Committee on Finance of the US Senate, which voiced concerns on the manipulation of inter-company pricing, are discussed in Graetz and O’Hear (n. 27), pp. 1060–1. Section 240(d) permitted the consolidation of the accounts of related trades or businesses under common control ‘in any proper case, for the purpose of making an accurate distribution or apportionment of gains, profits, income, deductions, or capital between or among such related trades or businesses’. Section 240(d) was in turn based on Regulations No. 41 Relative to the War Excess Profits Tax, Articles 77 and 78. These Regulations were enacted pursuant to the War Revenue Act of 1917. Article 77 was concerned with the reporting requirements imposed on affiliated corporations in relation to their ‘intercorporate relations’ to ensure the US Internal Revenue Service were able to ‘compute the amount of the tax properly due from each corporation on the basis of an equitable and lawful accounting’. The provision, which required reporting of all intercorporate relationships with affiliated companies, applied where: broadly, the corporations were under common control or either (i) where products or services were sold ‘at prices above or below the current market, thus effecting an artificial distribution of profits’ or (ii) where one corporation ‘in any way so arranges its financial relationships with another corporation as to assign to it a disproportionate share of net income or invested capital’. Article 78 gave the Internal Revenue Service the power to require from such affiliated companies consolidated returns ‘whenever necessary to more equitably determine the invested capital or taxable income’. Section 45 is discussed in Carroll Vol. I (n. 56), pp. 251–2.

157  The later comments of Mitchell B. Carroll suggest he saw the ALP as a broad-based rule ‘to prevent diversion of income’. Carroll, ‘Income Tax Conventions’ (n. 115), p. 20.

158  H.Rep. 67–350, at 14 (1921).

159  Internal Revenue: Hearings Before the Committee on Finance of the United States Senate on H.R. 8245, 67th Cong. 256 (1921), at 80, quoted in Graetz and O’Hear (n. 27), pp. 1060–1.

160  Carroll Vol. IV (n. 107), pp. 45, 57, and 100.

161  It was noted that the proposition argued before the Supreme Court of the United States in the case of Hans Rees’ Sons Inc. v. State of North Carolina (283 US 123 (51 S.Ct 385, 75 L.Ed. 879)) was that the income realized on the sale of goods may be split up into profits pertaining to the various stages in the acquisition or manufacture of goods which precede the relevant sales. See Carroll Vol. IV (n. 107), p. 34.

162  In 1938 at the eighth meeting of the Fiscal Committee, it was decided to stop any further separate work on business income but to use the results already obtained since its sixth meeting in 1936 as a contribution to the new work it was doing on the principles of taxation. See League of Nations Fiscal Committee, Report to the Council on the Work of the Eighth Session of the Committee, Geneva: League of Nations, 1938, C.384.M.229.1938.II.A. That work commenced with a review of the principles relating to taxes on income and led to the formulation of the practical principles which were considered essential in the administration of taxes on income, though the discussion was pitched at a fairly general level. For example, it dealt with matters like the structure of state income taxation, assessment, appeals, and character of the law. See League of Nations Fiscal Committee, Report to the Council on the Work of the Ninth Session of the Committee, Geneva: League of Nations, 1939, C.181.M.110.1939.II.A, pp. 2–6.

163  The work on tax evasion started from 1936. The work on the technical principles of taxation and fiscal policies was carried out from 1938 and was largely concerned with the orientation of fiscal policy in relation to economic fluctuations (the work was carried out in the light of the global crash from 1929).

164  League of Nations Fiscal Committee, Report on the Work of the Tenth Session of the Committee, Geneva: League of Nations, 1946, C.37.M.37.1946.II.A, pp. 6–8.

165  The use of this test in the Mexico draft was attributed to the fact that ‘it agreed better with American legal systems’ (Report on London and Mexico Models (n. 103), p. 11). The London definition was the same as that used in the earlier work of the Fiscal Committee and by then it appeared in most tax treaties concluded between European countries.

166  At the London meetings, however, it was observed that the permanent establishment concept was already used in almost all treaties dealing with business income. Report on London and Mexico Models (n. 103), p. 14.

167  Report on the London and Mexico Models (n. 103), pp. 77–9.

168  Report on the London and Mexico Models (n. 103), pp 60–1.

169  The relevant provisions are contained in the protocols to the Mexico and London drafts. Report on the London and Mexico Models (n. 103), pp. 77–81.

170  Report on the London and Mexico Models (n. 103), p.18.

171  This third step was expanded in the London and Mexico drafts in that those drafts also included reference to the situation where the activities of the permanent establishment was in the nature of those of a genuinely independent commission agent or broker. In such a case, the drafts directed that the income of the permanent establishment may be determined on the basis of the customary commission received for such services. Report on the London and Mexico Models (n. 103), pp. 79–80.

172  The Commentary clarifies that this approach is to be the limited fractional apportionment method, which considers only the profits of the enterprise as a whole from transactions in which a part has been taken by the permanent establishment (this contrasts to the method of unlimited fractional apportionment which takes as its starting point the total income derived by the enterprise as a whole from all sources) and also that the relevant total profits would then be apportioned to the permanent establishment by the ratio that exists between certain factors such as plant and equipment, circulating capital, payrolls, operating costs, physical output, and turnover, with the selection and weighting of the different factors varying according to the type of business concerned. Report on the London and Mexico Models (n. 103), pp. 20–1.

173  Report on the London and Mexico Models (n. 103), pp. 18–19.

174  Report on the London and Mexico Models (n. 103), Protocol, Article VI, para. 5 in both drafts. This provision seems somewhat curiously worded in that it seems to let stand a result that is not a fair allocation of income but presumably it was intended to allow taxing rights to be redrawn by the agreement of the competent authorities.

175  Report on the London and Mexico Models (n. 103), Protocol, Article VI, para. 1 in both drafts (see pp. 78–9).

176  In both drafts the provision stipulating that a subsidiary company is not to be regarded as a permanent establishment of the parent company was contained in the Protocol to the treaty at Article V, para. 8 and, again in both drafts, the early version of the ALP appeared in Article VII of the Protocol. It would later appear as Article 9 of the OECD Model of 1963.

177  This ‘corresponding adjustments’ provision would not appear in a model tax treaty until as late as 1977. See Chapter 2.

178  Report on the London and Mexico Models (n. 103), p. 17.

179  Report on the London and Mexico Models (n. 103), p. 17.

180  A possible explanation is that prevailing thinking was then largely limited to an essentially bilateral perspective, for example in relation to source and residence countries. Interestingly, in early US tax treaties the scope of provisions on associated enterprises was limited to income shifting between parent company and subsidiary, but not between sister companies. See Jens Wittendorf, Transfer Pricing and the Arm’s Length Principle in International Tax Law, the Netherlands: Kluwer Law International, 2010, at p. 77.

181  See, for example, League of Nations Fiscal Committee, Work of the Fiscal Committee during its Sixth Session, Geneva: League of Nations, 1936, C.450.M.266.1936.II.A, at pp. 1–2.

182  See, for example, Graetz and O’Hear (n. 27), p. 1099; Reuven S. Avi-Yonah, ‘The Rise and Fall of Arm’s Length: A Study in the Evolution of U.S. International Taxation’, Law and Economics Working Papers, Public Law Working Papers 92 (2007), 4.

183  1923 Report on Double Taxation (n. 37), p. 45.