Taxing the financial income of multinational enterprises by employing a hybrid formulary and arm's length allocation method.

TABLE OF CONTENTS

I     INTRODUCTION                                         690

II.   THE ISSUE IN CONTEXT: WHY IS FORMULARY ALLOCATION    626
      OF MNES' FINANCIAL INCOME DESIRABLE

      A. How MNEs Utilize Related Party Financial          627
      Transactions to Avoid Tax

      B. The Unitary/Formulary Alternative for Sourcing    628
      MNEs

      C. Using Formulary Allocation to Allocate MNEs       629

      D. Scope of the Forthcoming Analysis                 632

III.  APPLYING A FORMULARY SYSTEM ON MNES' FINANCIAL       632
      INCOME

      A. Formulary Allocation of MNEs' Financial Income    632

      B. Considerations with Regards to the Nature of the  636
      Allocation Formula

      C. Isolating the Financial Income Component          639

      D. Formulary Allocation within a Residence Tax       643
      System

IV.   APPLYING THE UNITARY AND FORMULARY ALLOCATION        645
      METHODS WHILE LACKING MULTILATERAL AGREEMENT

      A. The Hurdles of Implementing Formulary Allocation  645
      in a Nonharmonized International Tax Setting

      B. The Stakes and the Players in the Tax             647
      Coordination

      C. Mutual Recognition as a Limited (but Sufficient)  648
      Form

V.    APPLYING THE UNITARY AND FORMULARY ALLOCATION        656
      METHODS ON MNES WITH COMPLICATED HOLDING STRUCTURES

VI.   IMPLEMENTING FORMULARY ALLOCATION-ASSESMENTS OF THE  664
      PROPOSAL


VII.  THE ALLOCATION PHASE: SOURCING AS TAXING INCOME ON   666
      A NET BASIS

VIII  CONCLUSIONS                                          670

I. INTRODUCTION

International pressures have undoubtedly made the income tax and the corporate income tax more vulnerable. The scope of the abuse is so wide that scholars today often wonder whether the ability of sophisticated taxpayers, namely Multinational Enterprises (MNEs), to avoid taxes through international tax planning renders the income tax irrelevant. (1) In this context, this article advances a proposal (Proposal) to prevent international tax planning from eroding the corporate tax base and to sustain the viability of the corporate tax as a fiscal instrument.

Current legal scholarship seems to be divided into two camps--the optimists, who think that the income tax could be saved if countries adopt a unitary/formulary allocation system, and the pessimists, who think that such a tax allocation mechanism is harmful and difficult to operate. (2) Reality, however, is more complicated and nuanced.

The discussion over formulary allocation is not new and can be traced as far back as the 1930s. (3) In the unitary system, the MNE is considered as a single entity for tax purposes. The pre-tax income should be allocated to different jurisdictions according to a formula that relies on easily observed indicators to approximate the percentage of MNE activity taking place in each jurisdiction. This is very different from the current regime, which determines the tax liability of each MNE entity by monitoring transactions among MNE members and validating that those transactions are made in accordance with the "arm's length standard." This standard requires MNEs to assign the market price to their intra-group transactions and allows tax authorities to adjust intra-group prices. The critics of the unitary alternative argue that it requires high, unrealistic, levels of government cooperation and that it presents opportunities for distortion and abuse.

If the critics of the unitary alternative were correct, then the consequences would be grim. The ability of MNEs to avoid taxes by shifting income to low-tax jurisdictions results in an inequitable and inefficient erosion of the corporate income tax base. (4) In a world of unprecedented fiscal deficits and growing entitlements, tax regimes cannot afford to lose the ability to tax capital income. (5) Accordingly, the premise of this article is that until it is proven that there are alternative sources for government revenues, the income tax and the corporate income tax should be efficiently and equitably enforced. (6)

It is important to stress that the unitary and arm's length methods each have unique strengths and weaknesses. The arm's length standard is a relatively efficient tool for pricing transactions for which there is a market. Consider, for example, various types of oil products. Even though the costs of extracting and refining oil may differ from one MNE to another, there is a relatively transparent market price benchmark for each type of oil product. Relying on market price, therefore, relieves tax authorities from inquiring about each MNE's actual costs of production. Allocation by formula, on the other hand, is more effective when there is no market. This is frequently the case with intangibles, and, as this article argues, for financial transactions. When there is no market, tax authorities have to undertake a costly audit process to determine whether the pricing of related transactions is reasonable. In these inquiries, tax authorities suffer from an inherent disadvantage, given their limited resources and expertise. A formulary allocation that relies on easily observed indicators of economic activity would make it much more difficult for MNEs to avoid taxes by manipulating the prices of their related transactions. Accordingly. MNEs would not be able to take advantage of the fact that there is no market benchmark to drag tax authorities into long fact intensive audit controversies.

This article argues that the arm's length standard and unitary allocation systems are not contrary positions, as they is often portrayed in the literature. Viewing the arm's length and unitary regimes as a binary choice leads one to the erroneous conclusion that neither could be effective. Scholars point to the main weaknesses of each mechanism and conclude that neither the arm's length nor the unitary alternative could sustain the burden of allocating income consistently, equitably, and efficiently. (7)

This paper argues that the arm's length and unitary alternatives should not be viewed as mutually exclusive because the two can be combined and reconciled. The income tax is not beyond saving if tax authorities adopt a system that allocates MNE income by combining both the arm's length and unitary system in a way that capitalizes on the strengths of both methods. This dual system should employ arm's length allocation to allocate transactions where there is a clear market benchmark and employ a formulary system to allocate those components in MNE income that do not have a readily discernable market price.

In two previous papers, the author dealt with how a formulary system should allocate the income of MNEs in two special cases. The first paper concerned the allocation of income derived from intangibles and the second concerned the income allocation of MNEs that are financial institutions (FMNEs). (8) These papers demonstrated that formulary systems can offer an intuitive allocation of income arising from MNEs' affiliate transactions in financial and intangible assets. The lack of market benchmarks for many related intangible and financial transactions, the mobility of these assets, their nonphysical nature, and their tax-sensitivity renders it almost impossible to break down their ownership among different MNE components. Accordingly, as a matter of tax policy, it makes sense to pool the income derived from these assets and allocate it via a formula, and those two papers detailed how such a formula should work.

The scope of this paper is broader, however. Rather than offering a theoretical formula, this article outlines how formulary allocations could be integrated into current tax practices given real world political constraints. This problem is less relevant in the case of intangibles and FMNEs. (9)

To tackle this issue, this article investigates how the taxable income an MNE derives from financial assets should be allocated among the different jurisdictions in which it operates. It will be argued that such a formulary allocation is possible and that there is a way to overcome three fundamental challenges associated with implementing it. First, the article demonstrates that it is possible to separate income derived from financial assets (and obligations), which should be allocated through a formulary approach, and the income derived from other sources, which should be allocated according to an arm's length approach. Demonstrating that tax authorities have the effective ability to separate MNEs' financial income from other sources of income is crucial to the article's argument that tax authorities should adopt a system that deploys both formulary and arm's length methodologies. While this challenge is specific to the current article's Proposal, the other two challenges this article addresses are broader in scope and relate not only to the current Proposal, but also to more general concerns of how unitary and formulary allocation mechanisms could be implemented.

Second, the article demonstrates how different countries adopting a formula can coordinate their tax allocation consequences. The importance of this contribution cannot be underestimated. Unitary and formulary proposals are often dismissed on the grounds that they would require a harmonized global corporate tax base, or, at the very least, the existence of an unrealistic international "World Tax Organization". In contrast, this article argues that a plausible formulary allocation system could be devised with much more limited coordination than what most critics of the unitary system suggest. Showing that only a limited amount of tax coordination is necessary is important because it means that the possibility that a multilateral adoption of a formulary allocation system is more plausible than what most people tend to think. It will be argued that limited coordination could be achieved by modifying current bilateral double taxation treaties and that no new institutional arrangement is needed. This is a bold argument because the double taxation treaty network is typically perceived as a serious drawback to any future reform of international income tax arrangements and not as part of the solution.

Finally, the article addresses another long-standing objection to formulary allocation that is grounded in the difficulty of delineating the boundaries of MNEs. It proposes a system that on the one hand is very difficult to manipulate but on the other hand is not binary. The Proposal undermines the critique that it would be difficult to determine the boundaries of the MNE group when some of its subsidiaries have a complicated ownership structure. Furthermore, the Proposal makes the tax allocation result more intuitively correct because it avoids an "all or nothing" result in those borderline cases where the MNE is not the only dominant shareholder. Instead, it offers a more gradual result which, again, increases the likelihood that tax authorities would be willing to adopt it.

The analysis presented strongly suggests that adopting a formulary system to source MNEs' financial income could be a viable practical and political option. If tax authorities adopt this option in a limited but effective way. as suggested by this article, it could prevent the erosion of the corporate income tax base and many of the inequities and inefficiencies associated with international tax planning and income shifting.

This conclusion has profound effects on the manner in which international tax law has been examined. For many years, it has been assumed that unitary/formulary allocation systems are fundamentally in opposition to residence taxation. This article demonstrates, however, that while partial formulary allocation weakens residency taxation, it does not eliminate it altogether. More importantly, for many years source taxation has been viewed negatively because it has been perceived as a tax on gross income, which is unable to fully take into account taxpayers' ability to pay. This article explains why a formulary method of allocation renders this criticism wrong. Formulary allocation attributes net pre-tax taxable income to various jurisdictions. Accordingly, the source tax imposed on the revenues allocated to a specific jurisdiction is a tax on income and not on gross income.

From a practical perspective, the Proposal calls for disregarding MNE intra-group financial transactions. Instead, it suggests that the income (and deductible expenses) that MNEs derive from their financial holdings should be pooled and allocated via a formula. Such a reform would have an enormous impact. It would have implications for the ability of MNEs to engage in all sorts of tax planning and not only in related party income shifting transactions. Financial transactions are important in a number of respects. They involve financial assets, which are easy to deploy and costless to store, and operate within financial markets that are tax sensitive and where much tax planning expertise has been generated. (10) Because every MNE operates in the financial markets, financial assets are the spinal column of many tax planning transactions. Thus, providing a comprehensive solution that disregards many intra-group financial transactions to allocate the taxable income derived from financial assets and transactions would not only reduce the ability of MNEs to shift income to low-tax jurisdictions, but it would also limit their ability to engage in a wide array of tax planning transactions.

Part II provides the necessary background for the Proposal, reviewing the difficulties of sourcing affiliated transactions using the arm's length standard and the unitary/formulary alternative. It also explains the special case for providing a coherent allocation treatment to MNEs' affiliated financial transactions. Part III explains how financial income of MNEs should be separated from other forms of income and allocated through a formulary arrangement. Part IV explains how a formulary arrangement should be integrated into current practices. It argues that no massive reformulation of the international tax regime is necessary because formulary allocation arrangements could be integrated with the current bilateral double tax treaty network. Part V assesses different possibilities for redefining the MNE unit. Part VI provides an overall assessment of the Proposal. Part VII advances the normative argument for why formulary treatment of MNE income is a necessary move to reconcile the fundamental tension between taxing income and source taxation. It further explains why this move to a formulary setting also allows one to take into account notions of global wealth redistribution. Part VIII offers some short conclusions.

II. THE ISSUE IN CONTEXT: WHY IS FORMULARY ALLOCATION OF MNES' FINANCIAL INCOME DESIRABLE

This Part briefly summarizes some key arguments made in a previous article (11) in order to provide the reader with the necessary foundation upon which the more general policy arguments of this article are built.

Subpart A explains why there is a special urgency in finding a proper solution to allocate the income corporations derive from financial assets. Subpart B explains the concept of the unitary system of income allocation. Instead of allocating income by trying to isolate the "real economic price" of related transactions, the unitary system allocates income via a formula that tracks the activities of MNEs in different jurisdictions.

Unitary systems and allocation by formula are not inherently related to the problem of sourcing financial income, however. In fact, in the United States practically all states use a unitary formula to allocate corporate income among them. States tend to do so, however, without taking into account income derived from financial assets, which is typically allocated to the state where the company is incorporated. The final step in this analysis is to combine the first two Subparts and explain why a system that uses formulary allocation should be implemented to source the financial income of MNEs. This Part also briefly introduces a unitary system designed to allocate the income of FMNEs. The financial income of FMNEs is derived almost entirely from holding and trading financial assets. Therefore, the allocation formula used under the unitary system serves as a point of reference to a possible formulary allocation method to source MNEs' financial income.

A. How MNEs Utilize Related Party Financial Transactions to Avoid Tax

In recent years, financial transactions have become more difficult to tax audit. Due to recent technological advancements and relative global political stability, there arc currently many active financial markets around the world, including financial havens with low taxes and little regulation. In the evolving financial markets, investors trade a wide variety of financial instruments. As the recent crises demonstrate, understanding what is happening in these markets is often extremely difficult, often confounding financial regulators, central bank commissioners, and other key players. This lack of understanding of financial markets and the nature of the instruments traded in them also makes the tax auditing of financial transactions considerably more difficult.

Recent years have seen a considerable tax competition for financial investments. As a result, countries have significantly reduced their effective withholding taxes, which are taxes laid on payments of interest and dividends made from resident taxpayers to nonresidents. (12)

There are three main ways that MNEs use financial transactions to reduce their tax liabilities. First and foremost, MNEs use related financial transactions for income shifting purposes, increasing their deductible costs of finance in high tax jurisdictions and inflating their income on profitable ventures in low-tax jurisdictions. The problem with financial transactions is that even though tax authorities can price the credit risk of a specific loan contract made between related parties, there is no effective way for tax authorities to determine whether the form of the transaction is genuine. Related party loans are not really subject to the same credit risk as loans made to unrelated parties. In fact, in most cases, related party loans simply reflect movements of capital resources within the same economic unit, which suggests that from an economic perspective the contractual package of the loan is not much more than a piece of paper.

While the first problem is one of source taxation, the other two problems are related to the attempt to tax MNEs on a residency basis. The second problem is the ability of MNEs to defer paying taxes on financial assets accumulated in foreign subsidiaries. The third problem is the way in which MNEs use financial transactions to avoid taxes by engaging in tax planning strategies to inflate foreign tax credits. To be sure, related financial transactions are not the only way to achieve tax planning objectives. Their mobility and fungibility, however, make them the cheapest, most readily available, and most effective planning tool. The ability of financial markets to offer infinite contractual packages to similar cash flows, along with the tax planning expertise generated by financial intermediaries offering services in the financial markets, allows MNEs to use financial transactions for tax planning purposes.

B. The Unitary/formulary Alternative for Sourcing MNEs

Much of the problem with income shifting is the inability of tax authorities to determine the proper market price and proper structure of affiliated transactions. As noted, this problem is particularly true in the case of affiliated financial transactions. The lack of a clear benchmark results not only in a problem of tax avoidance, but also gives rise to huge tax administration and compliance costs. The transfer pricing rules, through which tax authorities regulate the arm's length standard, are a source of tremendous compliance costs because of their complexity. In addition, some commentators claim that the inconsistent manner in which different countries employ the rules is a potential source of double taxation. The problems associated with the arm's length standard are so severe that they compel an examination of alternatives. The most frequently mentioned alternative is the unitary method of allocation.

Under unitary methods, an MNE's total taxable income is aggregated and then sourced by an allocation formula among the jurisdictions in which it operates. Accordingly, the unitary system comprises two steps. The first is an income calculation step. In the unitary system, income is calculated jointly for an entire MNE and not separately for each of its branches and entities. To calculate their income, MNEs' tax filings include a consolidated report of their entire earnings rather than separate reports for each of their entities. This consolidated reporting effectively disregards affiliated transactions. The second allocation step comes after MNEs' incomes are calculated. Here, MNEs' incomes are distributed among the different jurisdictions in which they operate according to an allocation formula. This formula adopts factors that are easily observed and not easily manipulated, such as the amount of sales, payroll, and assets, as indicators of the relative share of MNEs' economic activity taking place in each jurisdiction. Unitary systems do not prescribe identical tax rates among different jurisdictions.

The core idea of the unitary system is that there is no one magic metric that penetrates the opaque process through which MNEs generate their profitability. This opaqueness prescribes that factors of the appropriation formula represent policy choices about how to allocate tax rather than precise economic indicators of how MNEs generate income. Unlike the arm's length standard, which sources according to market benchmarks that approximate what unrelated parties would do, the unitary sourcing regime aspires to tax MNEs' income only in near approximation. Tax authorities' formulary determinations of the proportional contribution of different factors to MNEs' profitability are by definition arbitrary. Furthermore, unitary system sourcing is a crude averaging mechanism, disregarding the different risk propensities and other distinctive circumstances of various MNE investments. For example, under a unitary system it is impossible for an MNE that is profitable overall to recognize loss in a specific jurisdiction no matter how bad its performance there is.

The most conceptually challenging issue in the design of a unitary system is the composition of its allocation formula. The formulary factors should be connected to immobile, easily observed indicators that can provide a proxy for the level of economic activity an MNE has in a given jurisdiction. The relative amount of sales, property, and wages that MNEs have or need to pay in a given jurisdiction are typically mentioned factors in this formula. In a previous paper, the author addressed the difficulty of defining each of these factors and concluded that the most difficult factor to locate is sales--especially in the context of nonretail sale. In the case of financial transactions, the location of the sale could be shifted with low costs to any branch or entity of the FMNE so that adding a sales factor to the allocation formula is basically a license to shift income.

C. Using Formulary Allocation to Allocate MNEs Financial Income

This article advocates that tax authorities should source MNEs using both arm's length and formulary methods. Arm's length methods adequately source most MNE non-financial affiliated transactions, which have market comparables; therefore, tax authorities have a legitimate interest in continuing to use them for these types of transactions. Because of their unique features, however, affiliated financial transactions should instead be sourced through unitary or formulary sourcing conventions.

One factor that supports this position is that the current international income tax regime treatment of affiliated financial transactions is broken beyond repair. In fact, the arm's length standard is so inept at dealing with these transactions that tax authorities already employ quasi-unitary alternatives to tax them. (13) Tax authorities have de facto abandoned traditional arm's length standard methods with regard to the hard-to-source income generating activities of MNEs. (14) Hence, they have a lot to gain and little to lose by shifting to formulary allocation. A unitary proposal would use general, transparent, hard-to-manipulate, and easy-to-observe indicators to determine the relative volume of an MNE's activities in each jurisdiction in which it operates. It would reduce taxpayers' ability to shift income as well as lowering the compliance costs of transfer pricing. Furthermore, formulary allocation methods allocate MNEs' income by relying on hard-to-manipulate indicators tied to specific geographic locations. If MNEs cannot escape the taxes on their financial activities by sheltering them in low-tax jurisdictions or by avoiding repatriating funds, they are free to utilize their assets efficiently without taking post-investment tax considerations into account. Thus, once fiscal ownership is fixed. MNEs would be able to utilize their financial assets in a centralized and tax indifferent manner.

To contextualize this argument, a previous article developed a unitary system for allocating the income of FMNEs. (15) Given the difficulty of determining where financial transactions occur, this unitary system relied on an assets and payroll formula. That article argued that a formulary system was innovative in the way it allocated the percentage of benefits between the property and payroll factors and in the way it defined those factors. For the purpose of this article, however, it is not necessary to elaborate further upon this previous work. This article focuses on how an allocation formula, any allocation formula--and not necessarily the one presented here--could be implemented to allocate the financial income of MNEs.

To avoid complications associated with explaining the principles of the suggested formulary allocation, the model assumed an almost ideal reality in which (1) FMNEs have only financial income (meaning income derived from the ownership and trade of financial assets), (2) there is a multilateral agreement among countries, and (3) all FMNEs have full ownership of their subsidiaries. The article then argued that, in the context of many FMNEs, achieving this "ideal" reality is not as implausible as one may think. Because of financial regulatory requirements, some FMNEs (especially banks and trading companies) tend to operate through branches and are required to separate their activities from the "real" economy. Furthermore, the bulk of FMNEs activities take place in financial centers located in a few developed countries such as London, New York, Chicago, Tokyo, and Frankfurt. These countries have similar interests in preventing FMNEs from avoiding taxes by income shifting so that attaining a multilateral agreement among them (such as the one reached in respect to banking regulation) would be difficult, but not impossible.

I concluded that the unitary alternative is superior to the existing transfer pricing regime with regard to FMNEs. FMNEs are integrated entities operating in integrated international financial markets. These integrated economic settings are in fundamental variance with arm's length standard rationales. This makes the attempt to breakdown FMNE income through transfer pricing rules both costly and insufficient to prevent abuse. The scope of abuse by FMNEs is particularly large because they can engage in income shifting of mobile assets more easily than MNEs operating in other sectors. This makes the proposal for unitary allocation of FMNE income not only attractive but acutely necessary to maintain the integrity of the corporate income tax in the financial sector. Finally, the Proposal allows MNEs and tax authorities to avoid many of the compliance costs, administrative costs, and inefficiencies associated with existing transfer pricing methods.

D. Scope of the Forthcoming Analysis

For the purpose of maintaining the acuity of the unitary proposal and the nature of its allocation formula, the former article (16) made some unrealistic assumptions (e.g., assuming the existence of a comprehensive multilateral agreement). In the following Parts, the current article relaxes each assumption to extend the formula developed for FMNEs so that it can be used to allocate the financial income of MNEs in an imperfect world. It then suggests additional policy measures to overcome the problems introduced and critically assesses their impact on the Proposal. Accordingly, the policy implication of this article's Proposal is much broader and multilayered than those of the previous one. Rather than addressing the important, yet confined, issue of FMNEs, this article opens a much broader discussion of how the financial income of MNEs should be sourced. It identifies the enormous concrete potential of using formulary sourcing methods in practical ways to overcome problems in contemporary international income tax regimes.

Upon evaluating the different measures needed in order to implement a formulary allocation of the financial income of MNEs, the analysis tries to strike a careful balance between concerns about audit accuracy, anti-abuse, and tax administration. There is an inevitable tension between concerns of tax administration, which require simplicity of rules, and concerns of audit accuracy, which require adjustability and particularity. (17) The following Part's analysis is impacted by this tension.

III. APPLYING A FORMULARY SYSTEM ON MNES' FINANCIAL INCOME

A. Formulary Allocation of MNEs' Financial Income

As mentioned, MNEs avoid taxation through related financial transactions that shift income to subsidiaries with low tax rates. As international markets and MNEs continue to integrate at an accelerating pace, this is a serious threat to the integrity of the corporate income tax regime.

By advocating the allocation of the income and deductions that MNEs generate from their financial holdings by a formula, the Proposal attempts to prevent MNEs from using these transactions for shifting income and for other tax planning purposes. Formulary allocation achieves both objectives because it disregards affiliated financial transactions. Nevertheless, the formulary system has limited scope in comparison to the above described unitary treatment for FMNEs. (18) It cannot prevent MNEs from using financially equivalent transactions to shift income or to execute tax planning schemes. Notably, however, tax authorities' inability to source affiliated financial transactions and the central role these transactions play in tax planning make this "limited scope" extremely important.

Given the fungibility of capital, the entire pool of capital resources commanded and obligations incurred by an MNE indissolubly comprise all of its assets and liabilities. This renders futile tax authorities' attempts to deploy transactional analysis to trace and allocate financial income and expenses to specific corporate entities within the MNE group.

It is important to stress the analytical distinction between unitary and formulary sourcing. Although the unitary system requires an allocation formula, the unitary concept primarily refers to determining the income of the taxable group of corporate entities. The formulary concept has a more limited scope, which refers solely to allocating income by formula and not by the arm's length standard. (19) In the context of the Proposal, formulary allocation of financial income would subject MNEs to two different types of income sourcing mechanisms. The first mechanism would source the financial income of MNEs through an allocation formula. The second mechanism would be applied to MNEs' non-financial income and would be sourced under conventional sourcing and transfer pricing rules.

In capsule form, this Part suggests that tax authorities disregard all intra-MNE financial transactions when computing their tax liabilities. The net taxable financial income of an MNE should be determined through its financial transactions with unrelated parties. Thus, in the case of MNEs, the goal is to isolate and allocate only their net taxable financial income. The calculation of the net taxable financial income figure differs fundamentally from the calculation of the net taxable income of FMNEs. In the case of MNEs, the net taxable financial income only accounts for the profits and losses of financial transactions and does not allow the deduction of other expenses. This net taxable financial income figure would be allocated to the different jurisdictions in which a specific MNE is operating according to an allocation formula, which, with few exceptions, resembles the property-payroll formula argued for in the case of FMNEs. (20)

The notion behind taxing the financial income of MNEs is that MNEs' financial assets and obligations are owned and borne by the MNE group as a whole. Any attempt to trace the geographic ownership of these assets and obligations is therefore bound to open a wide loophole for income shifting (21). As mentioned, the unitary system's underlying premise is that corporate groups generate income in an opaque process and therefore should be sourced according to proxies that represent policy objectives. The underlying claim of the formulary system is not that an MNE's entire income cannot be adequately sourced but that the income arising from specific assets and activities cannot be traced. Tax authorities' inability to trace coherently the source of financial activities troubles policymakers because it allows MNEs to strip the (financial and non-financial) earnings of subsidiaries in jurisdictions with high tax rates. (22)

Given the fungibility of capital, the net taxable income derived from MNEs' financial positions should be allocated in accordance with the location of their economic activity. As in the case of FMNEs, the overall payroll paid and the tangible assets held by MNEs serve as good proxies for where its economic activity is taking place. Using these proxies allows tax authorities to allocate the net taxable income from mobile financial assets to the location where the economic activities of MNEs take place. (23)

For example, consider Letni, a successful MNE. which develops, manufactures, and distributes computer chips. In a given year, Letni's net taxable income is $10 billion. Out of this, Letni has a negative net financial income of $800 million (which is Letni's ordinary income gains from financial assets minus its deductible losses and expenses from its financial obligations). This net figure should be sourced according to a formulary method similar to the one advocated for FMNEs. The method through which the residual amount of Letni's income ($10.8 billion) should be sourced is beyond the scope of this article. (24)

Under the Proposal, the related financial transactions of MNEs would be disregarded. MNEs could manipulate their tax consequences, however, by deflating or inflating the prices of their non-financial related transactions. These non-financial transactions may be difficult to manipulate because, in most cases, existing transfer pricing methodologies are sufficient to prevent most avenues of abuse. With the exception of affiliated intangibles-related transactions, which the author has addressed elsewhere, many related transactions have sufficient market comparables (at least under a Comparable Profits Method (CPM)). (25) Thus, inflating and deflating prices of related non-financial transactions probably would not be a major source of abuse.

The formulary treatment of the financial income of MNEs should be subject to a de minimis rule. First, this rule should exclude from the Proposal MNEs that do not meet a certain minimal threshold value. Second, it should also exclude MNEs with low gross financial income. Third, the Proposal should be a mandatory default, meaning that it could be modified through Advanced Pricing Agreements (APAs). (26)

These three supplementary rules allow tax authorities to implement discretion as to whether the Proposal's formulary sourcing is unfair or too costly to implement when dealing with nonpublicly traded, small-to-medium sized MNEs. (27)

B. Considerations with Regard to the Nature of the Allocation Formula

As mentioned, the unitary formula offered in the case of FMNEs included only tangible property and payroll factors due to the difficulty of determining the location of "sales" for nonretail financial transactions. It must be noted, however, that tax authorities should consider adding a sales factor to the allocation formula for MNEs because, unlike FMNEs, many MNEs will find the formula difficult to manipulate. An underlying premise of the allocation formula is that MNEs' entire financial assets and obligations support their entire pool of tangible and intangible assets. Some of these intangible assets are consumer-based (e.g., goodwill and brand name) and comprise an important component in the value of many MNEs. Thus, tax authorities should account for the value and geographic location of these intangibles to succeed in the formulary allocation of income according to the location of the economic activities of MNEs. The location of these intangibles is best determined by the sales of an MNE in different jurisdictions. Hence, to reach a more intuitively correct sourcing result, one may consider the introduction of a sales factor to be an important contribution. (28)

This article nevertheless maintains that the sales factor may also entail some significant problems. This is evident in cases where the transportation of services and goods is easy. For example, software companies can easily divert much of their sales to low-tax jurisdictions, and internet retailers in Northern Europe can easily establish their headquarters and servers in Estonia, which has a 0% corporate tax rate. (29) The problem of using the sales factor transcends the problem of internet commerce. Consider Letni, the microchip manufacturer. Although Letni's chips can be found in many PCs, it rarely contracts with end users. Its main consumers are three computer manufacturers: Lled, Abihsot, and Tsacmoc. All four MNEs have subsidiaries in Ireland, a low-tax jurisdiction, where they conduct active manufacturing, research and development (R&D), and distribution business activities. In cases where Letni's net taxable financial income is positive, it has an incentive to locate as much of that income as possible in Ireland. This can be done easily because, with little or no cost, Letni can channel much of its chip sales through its Irish subsidiary, which can then sell them to the other Irish subsidiaries of Lled, Abihsot, and Tsacmoc. This would inflate the sales factor in Ireland, resulting in more financial income allocated to the Irish subsidiary. The computer manufacturers can take these chips and pass them through the MNE entity so that they can be entered into PCs sold in high-tax markets. Tax authorities have no ability to monitor these types of conduit transactions and to trace the true location of the chips. Thus, this article maintains that while the sales factor provides a valid alternative in many cases, it also opens a significant avenue of abuse in many industries. This source of abuse should be of great concern to tax policymakers. As the world economy advances to a post-industrial stage, valuable mobile commodities and services comprise a growing component of modern economic activity. Furthermore, as MNEs broaden the set of geographical locations from which they operate, as well as the scope of their operations, it will be difficult to source any type of business-to-business transactions according to the sales factor. Tax authorities cannot counter the ability of MNEs to incorporate subsidiaries in low-tax jurisdictions through which they may channel sales and purchasing operations.

A comprehensive discussion regarding the virtues and vulnerabilities of a sales factor is beyond the scope of this article. It is important to note, however, that the option of introducing a sales factor to the Proposal's formula should be approached cautiously. Therefore, the sales factor should not be considered as part of the Proposal's mandatory default. Tax authorities should permit MNEs to source their financial income with sale formulas only on an APA-like basis. The APA agreements permitting the use of a sales factor in the allocation formula could (and probably should) be standard and transparent. As a precondition to entering these standard agreements, however, MNEs should prove to tax authorities, on a case-by-case basis, that they have taken measures to assure the sales factor will not be used for tax planning purposes. In the rest of the analysis, the article does not address this issue and assumes that the proposed formulary allocation does not contain a sales factor.

Another important question is how to define the payroll factor in the context of MNEs. Audit accuracy and anti-abuse concerns require that tax authorities define the concept of employees broadly to include independent contractors in the payroll factor. In contrast, tax administration concerns require tax authorities to structure easy-to-observe definitions for "payroll" and to identify employees.

In evaluating this question, it is important to take into account the following considerations. First, MNEs operate in many sectors and have countless positions. Thus, tax authorities would find it administratively impossible to determine which employees contribute most of the value and how to distinguish "true employees" from independent contractors in each business. Second, the total net financial taxable income subject to formulary allocation is likely to represent only a small portion of the income of MNEs. Hence, the motivations of MNEs to manipulate their payroll factor are likely to be significantly lower than those of FMNEs - where the unitary system dictates that all of their income is subject to formulary allocation.

It seems as though tax administration concerns require including only the costs of direct employees in the payroll factor. This definition gives MNEs the chance to engage in payroll planning by substituting direct employees with independent contractors. (30) However, in most cases, MNEs' net financial income may be negative so that the formulary allocation actually allocates deductions to various jurisdictions rather than the right to collect revenues. If that is the case, MNEs may actually have incentives to increase the payroll factor in high-tax countries. Given the high costs of direct employment in developed countries, MNEs may render this manipulation costly, which reduces the probability that they would try to manipulate the payroll factors solely for tax-savings purposes.

It is important to note that, in a previous article, the author reached a different conclusion with respect to FMNEs. (31) Weighing the same considerations of tax administration, accuracy, and tax avoidance, the author argued that the payroll factor should be computed only with reference to the payroll for professionals generating income. (32)

C. Isolating the Financial Income Component

The claim of this Part--that financial income should be sourced independently from other types of MNEs' income--is contingent on whether tax authorities can distinguish between financial and non-financial transactions. The answer to this question is bound to be controversial because MNEs often structure transactions that do not relate to underlying financial assets but which nevertheless contain a financial component. This is a crucial issue. Unless tax authorities are able to determine the financial income of MNEs at reasonable cost, none of the objectives of the Proposal concerning MNEs can be met. If financial transactions are not defined so that they can easily be identified, MNEs have an incentive to structure related and unrelated transactions so that expenses would be recognized as financial expenses and gains would be recognized as non-financial income. Additionally, the Proposal may even increase the burden of compliance and administrative costs.

This question of bifurcating the net taxable financial income of MNEs has a number of components. The first is whether it is possible to trace the financial income of MNEs arising from transactions with unrelated parties. This is not an easy task. Consider once again Letni, the MNE engaged in the production of computer chips. Assume that it enters into a contract with a foreign government to supply its military industry with computer chips. The chips are supposed to be delivered in one year. The foreign government agrees to pay 33% of the total nominal amount in advance, 34% when the chips are delivered, and an additional 33% five years after the chips are delivered. This relatively straightforward transaction contains two embedded loans: the prepaid amount is a loan from the foreign country to Letni, the deferred payment is a loan from Letni to the foreign government. Although complicated, tax authorities can calculate, with sufficient precision, the time value of money in these types of transactions. MNEs track this information for financial reporting purposes and therefore the compliance costs of reporting it should be manageable. The question would become much more complicated if Letni provided a service to the foreign government during the five year period following the delivery. In this case, tax authorities may find it difficult to determine which components constitute loans and which should be categorized as a return for services rendered.

The analysis would be even more complicated in the case of options. Assume in the above example that the foreign country would retain the right to return all the chips within the first two years in return for 34% of the total amount agreed upon and without paying the deferred payment. Since the underlying price and value volatility of Letni's chips may not be readily available, the option will likely be difficult to price.

Fortunately, if the concept of financial income is carefully delineated, this problem could largely be avoided. For the Proposal's formulary allocation, the financial income of MNEs should include only two categories. The first is the net taxable income from all transactions for which underlying assets are financial assets. Financial assets comprise assets representing pecuniary assets (e.g.. loans), equity holdings, assets whose value is a derivative of financial indices (e.g.. foreign exchange swap transactions), and, with few exceptions, (33) all assets that could be traded in financial markets (e.g., stock, bonds, future contracts, options). Subject to the above exceptions, this first category would include all derivative instruments relating to financial assets even if no public market for such derivatives exists (e.g., an OTC foreign exchange swap transaction). The second category of financial transactions would include transactions in which there is an implicit loan where one party gives the other the benefit of using the time value of money. For the purpose of administrative convenience, two de minimis rules should confine the second category of financial income. The first rule would relieve MNEs from keeping track of small embedded loans that happen as a matter of business practice. An additional de minimis rule would exclude differed payments in complex transactions, where it is not possible for tax authorities to determine the time value component of deferred or advanced payments. These two rules restrict the ambit of the Proposal because they include within the second category of financial income only those transactions in which MNEs explicitly use embedded loans.

For example, consider Drof, a company engaged in the development, production, and distribution of cars. It has a network of dealers that offers customers two types of sale transactions. The first involves a sale transaction in which the whole amount is paid at the time of the purchase. Drof allows its large corporate customers to pay up to 30 days after the cars have been delivered. In the second type of sale transaction, Drof helps clients finance the purchase by allowing them to pay with a premium periodically over a five year period. Under both transactions. Drof undertakes to service any malfunction in its cars during the first two years; however, actual expenses from servicing such malfunctions are negligible. Under the Proposal, the first type of sale should not be classified to include an embedded loan even if the payments were deferred. The second category, however, would be classified as an embedded loan.

The inclusion of embedded loans within the scope of financial income may add considerable complexity to the Proposal upsetting the balance between anti-abuse and tax administration rationales. Accordingly, it may be wise for policymakers to restrict the Proposal and define financial income as including only the first category--meaning only income derived from transactions that are composed of financial assets. Simply put. this article argues that it is impossible to draw a perfect line between financial transactions and non-financial transactions. For example, earnings-stripping techniques that use affiliated leasing transactions could replicate the consequences of related lending. Drawing the line at financial assets, however, allows the Proposal to achieve important objectives. It incapacitates the ability of MNEs to utilize their most mobile assets to attain tax planning objectives. Given the special features of financial assets, attaining this anti-avoidance goal is crucial.

Like any bright line rule motivated by a desire for administrative convenience at the expense of economic accuracy, this rule opens a loophole for tax planners. Nevertheless, the potential for MNEs to abuse this rule should not be seen as a major impediment to the Proposal. First, one has to remember that current conventions also allow manipulation of income categories and are widely used by taxpayers to avoid taxes. (34) Hence, the Proposal does not underperform current international income tax regime arrangements. As a policy matter, this type of manipulation is unavoidable once different sources of income are subject to different effective or marginal tax rates. Second, in the case of transactions with unrelated parties, the latitude MNEs have to manipulate financial income is limited by nontax considerations. The natural rivalry between MNEs and unrelated contractual parties operates as a safeguard that reduces the ability of MNEs to manipulate their net taxable financial income.

In summary, this article cannot solve all of the problems involving tax manipulation by extending the Proposal to source all MNE affiliated financial transactions. Therefore, an evaluation of this extension should bear in mind that this is a partial solution to a bigger problem. The question, therefore, is whether this partial solution is "efficient": are the costs of establishing it worth the relief it provides. It is probably safe to assume that tax authorities would define "financial income" restrictively--so as to include only transactions with financial assets. Even so, the Proposal would still be worth pursuing because of the important role played by affiliated financial transactions in the many income shifting and tax planning transactions executed by MNEs.

D. Formulary Allocation within a Residence Tax System

Another issue that emerges from this article's Proposal is how to classify dividend payments between related parties. If the Proposal's formulary allocation categorizes dividend payments made by related parties as financial transactions, it means that they should be disregarded for tax purposes. This, in the case of the United States, means a shift from a tax system that taxes the worldwide income of its resident MNEs to a territorial system. (35) For reasons elaborated upon elsewhere, I do not regard the transition to a territorial system as a negative development. (36) Nevertheless, a shift to a territorial system may be infeasible in the foreseeable future, because of political sentiments that favor worldwide taxation as a way to mitigate the foreign investments of United States residents. Accordingly, even though a wise long term policy would ignore related dividend payments altogether, it is necessary to develop a semi-strong version of the Proposal for formulary allocation for MNEs to improve its political viability.

Under this semi-strong Proposal, dividend payments between related parties would not be recognized as financial transactions subject to the Proposal's formulary allocation. (37) Dividends paid by unrelated parties are considered part of the net taxable financial income of MNEs. (38) Under the semi-strong version of the Proposal, it is necessary to prevent a situation in which subsidiaries elect the classification of financial payments made to parent corporations. Thus, for tax purposes, under the semi-strong Proposal, parent corporations would only be able to finance subsidiaries by capitalizing them. For tax purposes, affiliated subsidiaries will not be able to finance each other through debt investments but only through direct capitalization. As a consequence, every time a financial transaction is made between related parties, it would either be considered a dividend - providing that the subsidiary has the earnings and profits (E&P) to support it - or disregarded for tax purposes. This leaves MNEs with the capability to generate excessive foreign tax credit capacity through planning that involves manipulating subsidiary E&P through the use of hybrid entities and non-financial affiliated transactions. (39)

The ability of subsidiaries with retained earnings to enjoy the benefits of deferral is, however, substantially reduced even under the semi-strong version of the Proposal. This is because, once the subsidiary invests its retained earnings in financial assets, the income proceeds of the assets are calculated as part of net taxable financial income of the MNEs. This is true even if those retained earnings result from business activities that produce active income. Therefore, subsidiaries may only be able to attain deferral benefits on their working capital when earnings are reinvested in active non-financial businesses. In this respect, the semi-strong version of the Proposal could be seen as a stronger version of the Subpart F anti-deferral regime. This stronger version reduces the impact of deferral on all MNE financial income regardless of whether it is classified as passive or active and regardless of the jurisdiction in which it is sheltered.

The semi-strong version of the Proposal is, however, different from the anti-deferral regime in two important respects. First, by disregarding many affiliated financial transactions for tax purposes, it deprives MNEs of one of their most commonly used tax planning tools. Second, unlike Subpart F, it taxes MNE financial income on the source level and does not seek to eliminate deferral by imposing United States residency tax on all of it. This last attribute aligns with the general notion promoted by the article - that the right to tax MNE financial income does not "belong" to the country of residence. The fungibility of capital principle dictates that all of the financial assets and obligations of an MNE are part of the same pool, and thus, its financial income should be aggregated and sourced to all the locations from which it operates. The following table summarizes a comparative evaluation of various parameters between the Proposal's semi-strong version and existing transfer pricing sourcing techniques based on the arm's length standard.

Table One: A Comparative Analysis of the Proposal's Semi-Strong

Version of Allocating MNEs' Financial Income and the Current Arm's

Length Standard Sourcing Techniques

                      CURRENT TRANSFER    SEMI-STRONG VERSION
                      PRICING TECHNIQUES  OF THE PROPOSAL'S
                                          FORMULARY ALLOCATION

Deferral              Very high           Moderate-High

Excessive tax credit  Very high           Moderate-High
repatriations

Income shifting by    Very high           Very low
related transactions

* On the scale from "very high" to "very low," the lower ratings in the
table are the preferential ones.

IV. APPLYING THE UNITARY AND FORMULARY ALLOCATION METHODS WHILE LACKING MULTILATERAL AGREEMENT

A The Hurdles of Implementing Formulary Allocation in a Nonharmonized International Tax Setting

Much of the experience generated in implementing unitary systems has been in the United States, and some other nations with a federal form of government such as Canada. In cases where the unitary system is used, it is employed with relative success, providing corporate groups with benefits of consolidated losses and relief from the administrative burdens of transfer pricing. (40) A number of attributes contribute to the relative smoothness of the unitary allocation system in these regimes, including a general uniformity of the income tax base and tax accounting principles used to determine corporate performance. Additionally, states employing unitary sourcing methods have relatively well-established and obligatory dispute resolution procedures in the federal courts.

In the foreseeable future, there does not seem to be an institutional framework to guide nations to harmonize these complex sets of rules. (41) This is true even in political settings exhibiting relatively high levels of cooperation such as the European Union. (42) Furthermore, it is more than likely that even if sovereigns adopt the Proposal, their methods of income and formulary factors calculations will vary.

Another source of difficulty with the Proposal is its attempt to place a multinational formulary system within an international tax regime based, by and large, on a decentralized network of bilateral tax treaties. (43) These treaties subscribe to a number of soft law principles that adhere to concepts of separate tax accounting and arm's length sourcing. On the other hand, there is a genuine concern that implementing the Proposal unilaterally may result in double taxation and retaliation measures from other tax authorities. (44) Simply put, unless there is a way to overcome such difficulties the implementation of the Proposal by a large, powerful country, such as the United States, may lead to tax expropriation among nations. If tax expropriation occurs, the income tax will impose a heavy burden on multinational trade and investments.

These hurdles could be reduced substantially through double taxation treaties (DTTs). In capsule form, as long as the formulary methods employed by different tax authorities follow the Proposal's principles, the differences among them are not likely to be enormous. To reduce the revenue losses from income shifting, developed countries, which impose higher taxes, have a clear incentive to contract methods to relieve double taxation from the implementation of formulary systems. This type of DTT contracting would allow developed countries to beat tax havens and MNE collaboration to avoid tax through income shifting.

B. The Stakes and the Players in the Tax Coordination Game--An Overview

Understanding the article's argument in this Part first requires identification of the key players in the international income tax arena. Broadly speaking there are four groups: (1) developed high-tax countries, (2) developing (typically) low-tax countries, (3) tax haven jurisdictions, and (4) international investors (which typically are residents of developed countries). For sovereigns, the line distinguishing the different categories is contingent upon murky and highly disputable generalizations. (45) While it is true that all types of sovereigns use reduced tax rates to compete for investments (especially financial foreign investments) the stakes for each type of player differs substantially. Developed countries both export and import financial investments. Developing countries and tax havens primarily import financial assets--in the sense that their residents do not hold the lion's share of the capital assets invested in or channeled through them.

International investors do not typically shelter financial assets in developing countries due to their weak legal systems, political instability, corruption, unreliable financial and monetary regulatory regimes, high deficits, and constant need for revenues. Thus, in the context of financial income shifting, developing countries have a limited role. This, however, may change over time and the Proposal below is designed to allow for the inclusion of developing countries.

The tension caused by financial income shifting is thus primarily a tension among developed countries, on the one hand, and international investors and tax havens on the other. While developed countries compete for capital investments by lowering taxes, they are limited in doing so by their need for revenues to cover governmental expenditures. This Part labels developed countries as high-tax jurisdictions in relative rather than absolute terms. So long as developed countries impose a higher than 0% tax rate on mobile sources of income, tax havens would always compete with them by offering lower tax rates. Tax haven jurisdictions are relieved from the constraints of developed countries because they have a lot more to gain (in terms of employment and income tax laid on wages) by luring financial investments through low or null tax rates. For the purposes of this analysis, it is important to identify two types of tax havens: "sham havens" (46) and boutique havens.

Sham havens are traditional, offshore financial havens, which typically are associated with small islands. Often, these tax havens are previous colonies of pre-World War II European empires. Sham havens enjoy political stability and fairly reliable legal systems. They "host low corporate tax financial intermediaries that may be little more than an address for investment activities directed from elsewhere." (47) They attract investments by allowing commercial activity to be channeled through corporate entities in their jurisdictions, without imposing any regulatory or tax burdens and without requiring any tangible physical presence.

Boutique havens represent a new and important development of tax competition. They include countries such as Ireland and Singapore that offer investors platforms for investment and production in the form of highly developed infrastructures and human capital resources, along with low effective tax rates applicable to almost any type of economic activity. Like traditional offshore sham tax havens, boutique tax havens have a relatively small population (though not as small as sham havens) and therefore also gain more by luring investments through reduced taxes on financial income. Most important to this article, boutique havens allow subsidiaries of MNEs to escape Subpart F income tax liability by tying their accrued unrepatriated earnings to active business activities. (48)

C. Mutual Recognition as a Limited, but Sufficient, Form of Tax Coordination

MNEs, as the dominant type of international investors, use the opportunities offered by tax havens to avoid their taxes in high-tax jurisdictions. The collaboration between investors and tax havens reveals an important insight. The Proposal does not need to establish multilateral tax harmonization or multilateral mechanisms to avoid double taxation in order to provide an effective alternative to the current transfer pricing regime. To function effectively, all that is needed is a set of bilateral agreements among developed countries that reduces potential double taxation from conflicting formulary systems. This eliminates the possibility of lucrative collaboration between tax havens and MNEs. There are numerous methods through which high-tax, developed countries can mitigate this source of double taxation. (49) All methods are likely to have two fundamental steps in common: (unilateral) formulary calculation and (bilateral) mutual modification.

An example may help illustrate how this would work. Going back to the Letni example, assume it has four subsidiaries in the United States, the United Kingdom (both high-tax developed countries), Bermuda, and Ireland (sham and boutique tax havens, respectively). All figures in this example are in millions of United States dollars. In a given fiscal year Letni had a positive financial income of $800. By engaging in intra-MNE financial transactions, Letni shifted its financial income so that the United States and the United Kingdom subsidiaries were each allocated $100 of net taxable financial income, and the Bermuda and Ireland subsidiaries were each allocated $300 of net taxable income.

The first step would be each high-tax developed country applying its own formulary rule to source Letni's financial income, which should be attributed to their jurisdictions. Assume that both the United States and the United Kingdom employed a formulary method, which could be similar or different from the one advocated by the Proposal. Because of the lack of harmony between the rules of the two countries, there were differences in the amount of income allocated to each jurisdiction. While formulary systems of both the United Kingdom and United States allocated $100 to the Irish and Bermudian subsidiaries jointly, the United States and United Kingdom each allocated $450 to their own jurisdiction and $250 to the other. Given this scenario, the United States and the United Kingdom had a clear ability to mitigate double taxation by entering into a DTT requiring each of them to consider the formulary allocation of the other.

The second step would require both the United States and the United Kingdom to modify their sourcing results according to the sourcing consequences of the other country's sourcing formula. The most plausible way for each country to accept the allocation of the other is probably to accept the relative share in Letni's income that the other country assigned itself. Thus, in the above example, the United States modifies Letni's income by replacing the $250 it allocated to the United Kingdom (under the United States formula) with $450, which is the amount that the United Kingdom allocated to itself under its own formulary calculations. All other figures remain the same as under the original calculation of the United States formula, including the portions of Letni's income the United States sourced to itself and to other countries with which it did not establish DTT relief. This artificially inflates Letni's financial income to $1000 instead of $800 because the United States now allocates $450 to the United Kingdom, $450 to itself, and $100 to Ireland. This inflated figure of $1000 is necessary for the United States tax authorities to determine the share of each jurisdiction, though it does not affect Letni's actual financial income, which remains $800. Under the modified United States formula, both the United States and the United Kingdom are entitled to $450 out of Letni's inflated $1000 income figure. Thus, under this treaty modified United States formula, the United Kingdom and the United States are entitled to 45% of Letni's actual financial income. Because the actual financial income of Letni is $800, the 45% share that the United States attributes to itself is only $360.

Under the DTT between the United States and the United Kingdom, the United Kingdom is required to go through the same process, which means that, in this example, it too will allocate only $360 to its jurisdiction under its modified formula. To mitigate double taxation, both the United Kingdom and the United States had to agree to reduce the tax allocated to them under their own formulas. Instead of taxing Letni for $450 as determined under their own formulas, they had to reduce their audit to $360 to accommodate for one another's sourcing results. Both retained much more income than they would have under the transfer pricing regime, however. In a transfer pricing regime, Letni used its ability to shift its financial income, so most of it was recognized in Ireland and Bermuda, and the United States and the United Kingdom were both allocated only $100.

The following table summarizes the different allocation results of Letni's financial income under the current transfer pricing regime, the United Kingdom unitary system, the United States unitary system, and a hypothetical DTT between the United Kingdom and the United States to remove double taxation.

Table Two: Letni's financial income allocation under three sourcing
regimes (figures in millions of United States Dollars) (50)

            U.S.     UNITED   BERMUDA    IRELAND
                     KINGDOM

Allocation  12.5%    12.5%    37.5%      37.5%
under       100      100      300        300
transfer
pricing

United      56.25%   31.25%   0          125%
States      450      250                 100
formulary
allocation

United      45%      45%      0          10%
States      (450) *  (450) *             (100) *
allocation  360      360                 80

under
United
States
United
Kingdom
DTT.

United      31.25%   56.25%   6.25%      6.25%
Kingdom     250      450      50         50
formulary
allocation

United      45%      45%      5% (50) *   5% (50) *
Kingdom     (450) *  (450) *  40         40
allocation  360      360
under
United
States
United
Kingdom
DTT.

* Figures in parentheses represent inflated income figures, which are
only used to determine the percentage of income that each jurisdiction
should allocate under the DTT.

For this type of DTT arrangement to work, it is not necessary for the United States and the United Kingdom to agree on Letni's total net taxable income figure or to employ a similar allocation formula. All that is needed is for each regime to recognize the integrity of its treaty partner's formulary system and to take into account the percentage of Letni's financial income identified by the other formulary system when making sourcing determinations. One could make the above calculation while using income percentages instead of actual income figures. (51)

To prevent a situation in which parties to the DTT continuously inflate their share, the DTT should refer to specific formulary or unitary systems. Under Ihis arrangement, any (fundamental) change of the formulary or unitary allocation mechanism by one party would trigger the right of the other country to absolve the mutual recognition agreement. This reduces the chance that parties to the DTT will undermine the integrity of their formulary allocation arrangements to beggar-thy-neighbor's revenues.

The above mentioned method of relieving double taxation demonstrates the potential of DTTs to coordinate unitary and formulary systems. Sovereigns may use other types of DTT contractual measures to mitigate double taxation incidents that arise from sourcing through formulary systems. The chances for the success of such DTT arrangements increase if developed countries implement formulary systems that subscribe to the same set of principles. Attaining such a principled agreement is necessary to control the differences among sovereign formulary sourcing methods. Otherwise, it may be difficult to facilitate the trust among tax authorities necessary for such mutual recognition arrangements. All developed high-tax sovereigns are repeat players, however, and are likely to benefit from such new DTT arrangements. If the principled platform of the Proposal were adopted by the United States and a number of other major economies, this innovative DTT double taxation relief measure would be very effective. It would allow different formulary systems to operate without significant incidents of double taxation--despite a lack of central administration and harmonized tax rules.

Critics may argue that Letni would still be subject to double taxation because Bermuda and Ireland are not likely to waive their revenue claims through similar DTT arrangements. This criticism is materially mistaken. Letni will not be subject to double taxation because its financial income is not taxed by those tax haven jurisdictions. True, Bermuda and Ireland are not likely to adopt formulary allocation sourcing methods that align with the Proposal. However, the fiscal interests of high-tax developed countries require that the tax regimes of such tax havens would not be recognized or validated. The confrontation between high-tax developed countries and international investors cooperating with tax havens is a zero-sum game. High-tax developed countries should thus only be concerned with mitigating "real" double taxation--where income is actually taxed twice. This is very different from preventing double income recognition, when one jurisdiction in which income is recognized is a tax haven that does not tax that income. To equitably tax MNEs and other international investors, high-tax countries have the incentive to outplay tax havens.

Although the Proposal may increase the effective tax rate on international investors, this increase does not spring from double taxation. Under the Proposal, MNEs still enjoy the flexibility to arrange their financial structures as they see fit in a liberal capitalist market. MNEs, however, are not granted with a vested right to be sourced under the arm's length standard or with a right to shift income. By placing financial assets in tax havens, Letni can formulate its financial structure to attain nontax business objectives of any type. It nevertheless cannot contractually designate a significant portion of its financial income to go to tax havens in which it holds relatively low levels of economic activity.

It is important to stress that this article remains mute about the appropriate tax rate for the financial income of MNEs. Although minimal tax rate coordination is probably desirable, every sovereign can change the tax rate it imposes on financial income allocated to its jurisdiction. Thus, the United States could choose to tax the financial income of MNEs at a subsidized rate of 17.5% to attract their activities. Since the ordinary corporate tax rate is 35%, the United States could limit the ability of MNEs to deduct financial income losses from their ordinary income as a condition of the subsidy. (52) This tax rate flexibility allows countries to compete for foreign investments by offering low source tax rates on economic activities taking place in their jurisdiction but prevents them from offering discounts on shifted income. (53)

This tax rate flexibility is an important advancement over rival systems, because it enables sovereigns to effectively cooperate to implement formulary allocation without first harmonizing their tax regimes or establishing binding multilateral institutional, treaty, or dispute resolution frameworks. (54) To be sure, if such extensive multilateral coordination schemes were possible, it would be desirable, because it would allow smoother functioning of the international income tax regime. The DTT network is arguably responsible for many of the international income tax regime's distortions. Its decentralized and flexible features come at the grave cost of a body of law that is limited in scope, cumbersome, obsolete, incoherent, formalistic, and difficult to negotiate and amend. (55) Its tendency to provide different tax treatment for different income categories encourages taxpayers to shift earnings into more favorable income categories. (56)

While this article recognizes that DTTs are part of the international income tax regime's problem, it also recognizes that they can be part of the solution. One must acknowledge the potential of the DTT as an existing political tool through which developed countries may coordinate their different formulary allocation sourcing regimes.

History proves that sovereigns' attempts to attain cooperative equilibrium on tax matters encounter enormous difficulties. (57) While a comprehensive game theory analysis of how developed countries may best use DTTs goes beyond the scope of this article, it is important to stress the main considerations that such an analysis should take into account. First, the ability to use the DTTs as an effective coordination method decreases as the number of potential participants necessary to form a cooperative critical mass increases. Second, the ability to form a critical mass increases when the short term benefits of defecting from the cooperative arrangement are relatively small compared to the potential long term costs. These straightforward considerations interact interestingly with respect to the current Proposal.

First, avoiding double taxation on the financial income of MNEs requires the cooperation of many developed countries, which means that it would be difficult to promote this agenda on a bilateral negotiation basis alone. Some type of multilateral negotiation should take place--perhaps in the OECD--which has recently shown some willingness to consider shifting to formulary allocation in the case of financial income. (58) Second, if one adopts a payroll and tangible property formula, the benefits defecting countries can expect to receive are relatively low, because of the limited ability of MNEs to move their workforce and tangible properties. (59)

The DTT network may allow policymakers to start advancing the international income tax regime to a state in which income from mobile assets is sourced through formulary mechanisms without first attaining high levels of coordination. Put differently, while the DTT network and the arm's length standard developed simultaneously, they are not analytically related. Despite all the institutional difficulties involved in entering and coherently implementing DTTs, better DTT law could still take the international income tax regime a long way forward.

This insight has far reaching impacts. Existing literature typically refers to tax rule harmonization and uniformity as the main alternative to the prisoners' dilemma dynamics of tax competition in the international income tax regime. (60) At the same time, it is widely agreed that this uniformity is currently unattainable. Tax law diversity among different nations is deeply rooted, because tax laws reflect different values and economic histories, and are shaped by national sovereignty. (61) This article, however, offers high-tax countries a plausible option to exit the competitive dynamic without first harmonizing or restructuring the eighty-five years old institutional framework of the international income tax regime. To date, this alternative has been overlooked.

V. APPLYING THE PROPOSAL'S UNITARY AND FORMULARY ALLOCATION METHODS TO MNES WITH COMPLICATED HOLDING STRUCTURES

The operation of the Proposal's formulary and unitary allocation techniques requires a shift of emphasis from scrutiny of affiliated transactions to identification and distinguishing of MNEs. This process requires identifying corporate entities, which are part of an MNE (or FMNE) and, therefore, should be sourced according to the Proposal's formulary sourcing methods. The difficulty with this determination is that MNEs often hold less than 100% of the equity rights and voting powers in subsidiaries.

Determining MNE boundaries is not just a question of classifying a specific subsidiary as part of an MNE. It also entails the complicated question of what type of economic presence justifies source taxation. This question is primarily relevant to branches. Not every type of foreign taxpayer activity in the source jurisdiction should trigger the source tax. Rather, only branches that amount to a "permanent establishment" should trigger the source tax. The definition of "permanent establishment" is debated constantly on both academic and policymaking levels. This article refrains from addressing it directly, arguing that the issue of permanent establishments' sufficient economic presence could be determined in accordance with existing international income tax regime conventions. (62) Although the notion of permanent establishment is significantly under-theorized, a comprehensive discussion of it transcends the scope of this article.

The formulation of rules outlining the scope of unitary MNEs is crucial for the Proposal to operate smoothly. The previous Part explained how countries could accommodate differences in allocation formulas through bilateral agreements. It would, however, be much more difficult for countries to use DTTs in order to contract out the scope of MNEs. Thus, this Part discusses the ownership or control requirements tax authorities should use to analyze whether an entity is part of a unitary MNE. It formulates a set of innovative rules to define this unitary threshold.

One should not underestimate the importance of these rules. Unless there is a principled multilateral agreement regarding the scope of an MNE entity, countries may easily find themselves on a slippery slope of over-expansive formulary systems and retaliation measures. This article maintains that, on this issue, there is need to reach some kind of weak multilateral agreement. This requires that rules addressing this issue be clear and consistent, so as to allow coherent implementation by different tax authorities. The emphasis on consistent rules is bound to come at the expense of preventing MNEs from tax avoidance, since clarity enables planning around the rules. Because this article perceives sovereigns' inclination to extend their fiscal jurisdictions as the primary evil on this specific issue, it considers MNE tax avoidance to be a price worth paying. That being said, this article does not suggest that preventing sovereigns from extending their tax jurisdiction is the only objective worth pursuing with the rules that determine the scope of MNEs. The rules should be clear and simple so countries do not manipulate them, but, at the same time, tax policymakers should try to prevent these features from becoming anchors for excessive tax planning.

The greatest danger is that MNEs will be able to freely select which of the entities they (de facto) control to be entered into the Proposal's formulary arrangement. If this were the case, MNEs would still be able to shelter low-tax income in different types of tax havens, while allocating to low-tax jurisdictions financial income that otherwise would be subject to a high-tax rate. Simple rules are also insensitive to borderline cases. Part of the political feasibility of the Proposal depends upon its ability to source MNE income appropriately. If borderline cases were taxed in extremely different ways, the political feasibility of the Proposal would be eroded. Existing unitary systems operate under an all or nothing rule, which means that subsidiary income is either completely included or excluded from the unitary formula. The Proposal seeks to avoid this type of arbitrariness by providing an identification system that aligns with the economic reality of MNE structures. Thus, the identification rule resonates with the way that partially owned subsidiaries are taxed. In other words, if an MNE controls a subsidiary, but does not fully own it, only part of the financial income of this subsidiary should be formulary allocated.

Tax authorities can determine the scope of MNEs through criteria of ownership (direct and indirect) or effective management control. Both sets of criteria are admittedly difficult to implement consistently on corporate structures involving partially owned subsidiaries and/or diversified categories of equity rights. Any formulary or unitary sourcing methodology needs to determine a certain threshold upon which a corporate entity would be recognized as part of an MNE group. These thresholds are bound to be arbitrary to a certain degree. When formulating such rules, however, tax authorities should give heed to their administrability in light of prevalent MNE business structures.

In order to avoid administrative difficulties, it is best to determine the scope of MNEs using the benchmarks of equity ownership and voting rights. This requires formulation of an inclusive per se equity ownership or voting right percentage threshold, which prescribes that, above a certain holding level, a subsidiary is part of an MNE. This rule should be complemented by a number of safe harbors exempting certain ownership structures from unitary filing. The most simple bright line rule would dictate that every subsidiary in which 50% of the equity or voting rights are held by an MNE is part of that MNE. Any other type of equity held by MNEs would be treated as a capital asset. Since the article argues that capital assets should be treated for tax purposes as if they were held directly by MNEs and not by any of their subsidiaries, dividends paid by less than 50% owned subsidiaries should be included in the net taxable financial income of MNEs. This means that, under the semi-strong version of the Proposal, the net taxable income earned by such subsidiaries may not be treated as belonging to any specific corporate entity but to the MNE as a whole. Dividends from less than 50% owned subsidiaries would be exempt from residency taxation but subject to the Proposal's formulary allocation. This is true even if the subsidiary itself does not have any financial income. The underlying idea is that since the equity shares of this subsidiary are mobile assets they could be legally owned by each of the entities in the MNE group. These shares are financial assets and their dividend proceeds are part of the financial income of the MNE group.

Although even such an elementary and unrefined bright line rule is sufficient, one could create a slightly more sophisticated Identification Rule (the Identification Rule). The Identification Rule provides a more complex set of rules that could be applied consistently and less arbitrarily in borderline cases.

In tax law, complexity is not antithetical to clarity. Policymakers write tax laws for tax experts. Thus, complex legislation resulting in outcomes that are difficult to manipulate and unambiguous may be opaque to the general public, even though it is clear to the professional audience to which it is directed. The Identification Rule aims to attain this type of complex-yet-clear rule.

According to the Identification Rule, every subsidiary in which an MNE holds 35% of the voting power or equity would be considered to be part of that MNE for formulary purposes. The 35% benchmark is an arbitrary one. It serves as a proxy for a control block in the subsidiary. (63) It may be decreased or increased. (64) The subsidiary should not be considered part of the MNE if there is an additional shareholder whose voting power and equity are each equal to or greater than those held by the MNE. The computation of equity shares includes all convertible equity instruments (convertible preferred stock and convertible debt), which may carry a potential voting power in the subsidiary and/or a right to its residual profits.

The above rule could open a wide array of abuse because MNEs have incentives to structure transactions that allocate financial instruments containing an inferior (out of the money) convertibility option to manipulate formulary allocation. This would allow MNEs to elect whether the subsidiary's financial income is to be included in the Proposal's formula. Thus, for subsidiaries with positive financial income in low-tax jurisdictions (or subsidiaries with negative financial income in high-tax jurisdictions) MNEs have incentives to structure deals where counterparties have inferior convertible instruments. This would allow MNEs to avoid the Proposal's formulary allocation, because it would reduce their share in those subsidiaries for formulary computation purposes, without reducing their real economic share in those subsidiaries' equity. By doing this, MNEs would be able to benefit from sheltering income and deductions in favorable tax jurisdictions.

For subsidiaries with positive financial income located in high-tax jurisdictions (or subsidiaries with negative financial income in low-tax jurisdictions) MNEs prefer to inflate their convertible assets by holding inferior (out of the money) options in those subsidiaries. By doing so, they assure that for formulary purposes, such subsidiaries are considered part of the MNE, and that part of the net taxable financial income is allocated to entities in other jurisdictions through the averaging mechanism of the Proposal's formula. Because from an economic perspective the options are almost worthless, this inclusion has no real impact on the holding or finance structure of the subsidiaries.

This article has only partial remedies for these abuses. In the first scenario, where an MNE desires to avoid formulary treatment, the article suggests that other shareholders' convertible instruments should not be taken into account when determining the ownership share of MNEs. In the second scenario, where MNEs have incentive to inflate their own convertible holdings to assure formulary inclusion of subsidiaries, the abuse prevention is admittedly more complicated. Tax authorities lack a solid doctrinal or theoretical background to efficiently determine the validity of a convertible instrument. Moreover, tax authorities do not have the resources to commit to this costly, case-by-case enforcement endeavor. This, however, is not an avenue of significant potential abuse under the Identification Rule. As demonstrated below, the benefits that accrue to MNEs from inflating holdings to skew the formulary allocation would be low under the Identification Rule.

Once a subsidiary is considered part of an MNE, other MNE members cannot deduct payments from financial transactions entered into with that subsidiary. This means that MNEs cannot engage in financial transactions to shift income to the subsidiary. (65) The treatment of payments made by subsidiaries to MNEs is slightly more complicated. Ideally, it should be disregarded for tax purposes, meaning that subsidiaries should not be allowed to deduct payments and payments should not be included in the net taxable financial income of MNEs. Going back to the previous Part, however, if one assumes that not all countries endorse the Proposal, a strict enforcement of the Identification Rule may allow MNEs to enjoy double dipping. This double dipping may occur when a subsidiary's tax jurisdiction allows a deduction but the payments made to an MNE are not included in its taxable income. Accordingly, to the extent that a subsidiary enters into financial transactions with other MNE members and payments to them are deductible, it may be wise to consider including those payments in that MNE's net financial income. Under the semi-strong version of the Proposal, dividends made by the related subsidiary are recognized for tax purposes.

After determining whether a specific subsidiary is part of an MNE group, the Identification Rule would determine the share of the subsidiary's financial income to be included in the formula. If an MNE holds more than 80% of a subsidiary's actual equity, the entire financial income of the subsidiary would be included in the MNE's formula under the Proposal. The MNE's actual equity ownership does not include convertible instruments not yet exercised. This rule would, at least in the case of the United States, correspond to the notion that more than 80% ownership of equity is tantamount to complete control of a subsidiary. Complete control implies that the interests of the two corporate entities are so intimately entangled that it would be wrong, and administratively costly, to see them as separate entities for tax purposes. This 80% figure is arbitrary. It is by no means carved in stone and therefore could be decreased or increased.

In all other cases, where a subsidiary is part of the MNE but the MNE holds less than 80% of the equity, only part of the subsidiary's financial income would be included in the net taxable financial income of the MNE. The percentage of income included would be equal to the percentage of actual equity ownership in the subsidiary. Additionally, when determining the total financial income that should he allocated under the Proposal to the subsidiary, the formula should only include a percentage (equal to the MNE's equity holding) of the payroll and tangible assets factors of the subsidiary.

This rule has two advantages. First, it prevents the arbitrariness of an all or nothing distinction. Thus, unlike the 50% bright line rule, there would be no colossal difference between tax treatment of subsidiaries in which MNEs hold 49.9% voting power or equity and the tax treatment of subsidiaries where MNEs hold 50.1%. This point is crucial since it reduces--but does not eliminate--the benefits of tax planning. The notion that a subsidiary's unitary/formulary classification does not necessarily entail a binary treatment of the subsidiary's entire income is a policy point that has by and large escaped existing literature.

Second, the Identification Rule's focus on actual ownership reduces the incentives of MNEs to artificially inflate convertible holdings in subsidiaries. Once tax authorities categorize a subsidiary as part of an MNE group, the MNE loses the ability to strip income or to shift its financial income via that subsidiary. The MNE will only be able to allocate the percentage (equal to its actual ownership percentage) of the subsidiary's financial income to its net taxable financial income.

An example should clarify the Identification Rule. Assume Letni became a strategic investor in a relatively small high-tech corporation in Israel (IsSub). IsSub has only one class of stock, which holds proportional equity rights and voting power. The shares in IsSub are held by Letni's subsidiaries in Ireland and Bermuda. Letni purchased only 15% of IsSub's equity and voting power but supplied it with additional funding in the form of loans and convertible debt. An Israeli mutual fund holds 50% of the equity and voting rights in IsSub. The mutual fund also holds some convertible debt, although this is not taken into account when determining whether IsSub is part of the Letni MNE. To determine whether IsSub's financial income should be included in Letni's financial income, one has to determine its equity and voting rights by assuming that Letni (but not the mutual fund) exercised all of its convertible rights. It would be useful to assume the following two scenarios. First, if exercised, Letni's convertible rights would give it 37% (or less) of the equity and voting rights in IsSub. Such a dilution would reduce the mutual fund's share to 37% (or more). In this scenario. IsSub is not identified as part of the Letni MNE group for the purposes of the Proposal. Even though Letni may have more than 35% of IsSub's equity, IsSub is not considered part of the MNE because there is a shareholder with a holding in IsSub that is greater than (or similar to) Letni's holding. When distributing dividends, IsSub's dividends would be included in Letni's net taxable financial income and sourced by the Proposal's formula. Letni's financial transactions with IsSub are therefore recognized for tax purposes. IsSub would be able to deduct interest payments to Letni and those payments would be considered as part of Letni's financial income.

The second scenario would assume that if exercised, Letni's convertible rights would give it 85% ownership of IsSub's voting power and equity rights. IsSub would thus be considered as part of Letni for the purposes of the Proposal. Any interest payments made by members of the Letni group to IsSub would not be deductible. If Israel adopts a formulary sourcing regime similar to the Proposal, it would have to prohibit IsSub from deducting interest payments it made to corporate entities within Letni. In this case, Letni's financial income should not include the interest payments made by IsSub to Letni. If Israel allows IsSub to deduct payments made to Letni, however, it may be wise to require that the interest payments made by IsSub be included in Letni's net taxable financial income. Letni's actual equity ownership in IsSub is 15%. Thus, only 15% of IsSub would be included in Letni's financial income, and only 15% of its tangibles' value and payroll factors would be included in the Proposal's allocation formula.

It should be noted that the Identification Rule could be used by the tax planning industry for tax avoidance purposes in the same manner as any equity ownership benchmark. (66) Therefore, every equity ownership standard should be supplemented by a special anti-avoidance rule. (67) This anti-avoidance rule should only nullify corporate ownership structures aggressively designed to primarily avoid the unitary definition. It should be based on the hard-to-implement economic definition of business entity. It is important to note that the subtle nature of tax avoidance obscures the solution to general anti-avoidance rules, encumbering them with long and difficult evidentiary inquiries. Since no verbal formula can encapsulate the idea of tax avoidance, this type of rule offers no structural solution but only a flimsy last line of defense. (68) While this anti-avoidance rule may indeed contribute to the administration of formulary and unitary regimes, it should only be used in cases of extreme abuse by taxpayers. This characterization of the anti-avoidance rule as the mechanism of last resort is necessary for the clarity of the Identification Rule to be maintained. As mentioned, clarity is crucial for the Identification Rule to be consistently applied by different sovereigns.

While the Identification Rule and the bright line 50% rule discussed by the article offer no magic solution, the Identification Rule is not necessarily an enormous obstacle to the operation of the Proposal's formulary allocation. This is because, unlike related party income shifting, the manipulation of holdings may have a real price tag for MNEs because tax planning involving ownership manipulation may jeopardize MNEs' genuine business objectives.

VI. IMPLEMENTING FORMULARY ALLOCATION--ASSESSMENTS OF THE PROPOSAL

MNEs vary considerably in their propensities. They follow numerous business structures, operate all around the globe and have different levels of financial income. The Proposal's formulary allocation method should at least be implemented on the financial income of the 500 largest MNEs. These MNEs control the lion's share of international commerce, hold enormous assets, and engage in extensive tax planning. Large MNEs set standards of tax compliance and fuel the tax planning industry. Thus, large gains from implementing the Proposal would justify the difficulties of applying the Identification Rule to complicated corporate structures and amending and renegotiating DTT obligations. Moreover, if income that these large MNEs earn from financial assets is not taxed through allocation formulas, income shifting and deferral abuse of these dominant actors will continue to impair the integrity of the corporate income tax. The erosion of the corporate income tax base carried out by the largest corporate taxpayers is bound to negatively impact notions of equity in source taxation and the taxation of business entities more generally.

On a conceptual level, this article's discussion of the Proposal in a nonideal reality promotes the idea of formulary allocation as a nonutopian alternative. This article establishes that formulary sourcing does not need to include all commercial activities of MNEs but only those that are difficult to source in accordance with the arm's length standard, like financial activities. The key to this insight is that only certain types of activities, which have no adequate market comparables, should be sourced by the formulary method. In the absence of such comparables tax authorities cannot determine the "correct" price and/or structure of a transaction. Even affiliated transactions with market comparables could shift income through contractual risk shifting. (69) This risk shifting could be more easily identified and countered through the use of CPM methods. This is not the case with affiliated transactions involving financial assets since there is no one correct way to channel them. Thus, to offer a valuable alternative to policymakers, academic literature should not engage in promoting a unitary formula that encompasses all MNE income. Academics should rather aim to form formulary methods that source only the types of affiliated transactions in which market comparables are unavailable or unreliable.

This Part argues that a formulary arrangement is possible without a multilateral tax authority to administrate it. Admittedly, it is unrealistic to recommend that, in the near future, tax authorities start negotiating DTT provisions such as the one suggested by the Proposal. Over time, however, DTT practices aligned with the article's suggestion may develop. The importance of the suggestion is that it does not require an reformation of the international income tax regime. It thus offers policymakers a nonutopian institutional platform for advancing a more equitable source tax allocation.

Finally, the Identification Rule makes an important contribution by stressing that formulary sourcing is not necessarily an all-or-nothing mechanism. The gradual implementation of the Identification Rule somewhat smoothes the arbitrariness of traditional bright line classification rules. Although more complex than bright line rules, the Identification Rule is clear, fair, and possible to apply consistently by different sovereigns.

The notion that formulary arrangements are infeasible prevented the advancement of international income lax regime sourcing rules for the last two decades. Persuaded that they are deadlocked, policymakers and tax authorities have encumbered lax rules with the arm's length standard. Problems arose as a result in recent years as the volume and sophistication of affiliated transactions increased. It is absurd to endure enormous litigation, compliance, and administration costs related to transfer pricing given transfer pricing's failure to prevent tax avoidance. The discussion in this Part paves a way out of the deadlock by showing that advancements are attainable even within a politically constrained reality.

VII. THE ALLOCATION PHASE: SOURCING AS TAXING INCOME ON A NET BASIS

In two previous articles, the author identified a new tendency in the policymaking of the international income tax regime regarding the taxation of financial income. The articles surveyed the international income tax regime's treatment of interest payments over the last ninety years and the problem of earnings-stripping. (70) This new policy trend was labeled the Allocation Phase, in hopes that it represented the beginning of an emerging new sourcing paradigm in the international income tax regime. The features of this phase reflect policymakers' recognition that the main challenge and duty of the international income tax regime is not to eliminate abusive transactions but to develop economically sound, administrable, and fair sourcing allocation methods for financial income. The Allocation Phase is in the initial stages. Nevertheless, recent developments (71) recognize that the problem of sourcing financial income may not be solved on the quicksand of the ad hoc arm's length standard approach and that there is a need to provide a comprehensive perspective of how to allocate financial income.

Few advocate the integrity of the transfer pricing or source rules. (72) There is a fundamental tension between the notions of source and income taxation. An income tax is a tax on the net gains of a given taxpayer. Income is thus a holistic notion that looks to the aggregated economic position of taxpayers. In contrast, source taxes are categorically anti-holistic because they levy taxes only on activities in a given jurisdiction. Income source taxes are therefore bound to be schizophrenic for taxpayers with activities in a number of jurisdictions.

Tax authorities' attempts to allocate the earnings of MNEs to different geographical sources undermine the objective of the income tax, which is to impose tax burdens according to a taxpayer's net taxable income. The most obvious example of this is when an MNE is comprised of a hollow parent holding company and two foreign branches located in two different countries. When one branch has a loss of $200 and the other a gain of $100, the MNE's net economic income is -$100. Source rules, however, prescribe that an MNE is not allowed to offset losses in one jurisdiction against gains in another. (73) Thus, the jurisdiction that had a profitable branch must pay source taxes even though the MNE's net income was negative, which may seem unfair. As in any other case where legal form is inconsistent with economic substance, however, this inconsistency is used by MNEs to shelter earnings and deductions in locations with the most tax efficient results. Scholars and policymakers have recognized this tension from the very beginning of the international income tax regime in the 1920s. (74) The source rules, appeared antithetical to an equitable--and possibly progressive--taxation in accordance with taxpayers' ability to pay and incompatible with an intra-nation standard of horizontal equity. (75)

Under conventional source rules, which for MNEs rely on the arm's length standard, there is indeed an irreconcilable rift between the notions of source and income taxation. Source taxes, even when laid as corporate income taxes, are never levied on the net taxable income of MNEs because their commercial activities are not confined to one jurisdiction. Because source taxes have a jurisdictional priority over residency taxes, (76) the nature of the source taxes haunts the integrity of the international income tax regime, exposing the regime to abuse and reducing its political legitimacy.

This article attempts to narrow the gap between taxation at source and taxation of income. As the Proposal demonstrates, formulary allocation allows tax authorities to determine the net taxable income figures on the MNE level with regard to financial income. After such net taxable figures are extracted for a given MNE, they are allocated to the different jurisdictions in which it operates in accordance with the relative level of economic activity in those entities (as determined by formulary proxies). If recognized as viable sourcing arrangements, formulary and unitary sourcing methods may offer a quantum leap in the sense that they would allow source taxation on net taxable income. Furthermore, as the semi-strong version of the Proposal for MNEs demonstrates, the notion of source taxation supremacy under formulary allocation methods indeed weakens residency taxation although it does not eliminate residency taxation altogether.

Before concluding, one additional issue merits attention. The allocation formula operates under an expenses-equivalence assumption concerning money spent in different jurisdictions. Thus, as mentioned earlier, for allocation via payroll and property, the allocation formula is likely to produce regressive inter-nation wealth distribution outcomes, because of the immense disparities in wage rates and value of real property between different source jurisdictions. (77)

At the outset, it is important to note that the allocation formula cannot, by itself, promote redistributive objectives. The flexibility of the allocation formula's factors may be formulated, however, to take into account inter-nation distributive objectives. This feature of formulary sourcing gives it an enormous distributive potential. For example, this could be done by multiplying a constant figure inversely correlated to the average national GNP per capita when computing the revenue share of developing countries, so that poorer countries get a greater share. The current arm's length standard conceptual framework is flawed normatively because its bilateral perspective atomizes national claims for revenue. This atomization narrows the questions of sourcing to the identification of jurisdiction where income was produced. The allocation formula and unitary system methods offer a multinational perspective that would permit issues of global wealth distribution to be addressed in a theoretically comprehensive manner.

The article's notion that MNE corporate taxation should play a role in inter-nation redistribution is controversial because any corporate income should, at the end of the day, be attributed economically to individuals. Thus, since incidences of corporate tax are not entirely clear, especially for MNEs, it is hard to see how one can establish a distributive claim through corporate taxation. The article nevertheless considers the dominance of corporations as investment vehicles to justify a business entity tax level on investments of foreign taxpayers. MNEs are intermediate configurations between the country of investment and foreign investors, and, as such, global wealth redistribution should be a factor when determining how to tax them. The justification for corporate tax as a source tax on the MNE level does not mean that countries should adopt specific investor level tax treatment for returns on either domestic or foreign investments. Put differently formulary taxation of MNEs should promote "internation [allocative] equity," which in the case of international trade has primacy over "internation equity." (78) Once inter-nation allocative equity concerns are settled, each country has the prerogative to promote different notions of intra-nation equity among its individual residents through various forms of corporate-shareholder integration.

The Proposal's sourcing formula is just one of many possibilities that may unravel the Gordian knot of the arm's length standard. By way of conclusion, this Part summarizes the principled common denominator features of formulary regimes. First, every longstanding source regime should aim to identify the share in the income tax base of a specific taxpayer in a specific jurisdiction and should avoid levying withholding taxes on gross income. This requires source rules to use broad formulary sourcing arrangements to eliminate the need for withholding taxes. (79) Second, one should construct a regime that allocates the right to tax among jurisdictions in light of the material factors of production employed in each jurisdiction. In the context of financial activities, one must remember that capital is a fungible resource and that holding of capital assets cannot be traced among affiliated entities. The Proposal represents a sourcing arrangement that overcomes much of the capaciousness plaguing the present regime. The pillars of the Proposal are (1) an emphasis on sourcing the net financial income of an MNE--only at the source level and only through a net taxable income corporate tax--and (2) a sourcing formula that relies on difficult-to-manipulate proxies that correspond, roughly, with how financial income is generated by MNEs.

VIII. CONCLUSIONS

For many years, the international income tax regime policy has been deadlocked. Since its infancy, it has been haunted by the decision to source MNEs affiliated transactions according to the arm's length standard. Despite its theoretical problems, the arm's length approach has arguably provided a reasonably administrable solution for many years. In the last quarter of a century, however, the arm's length paradigm has reached a point where it is so broken that policymakers cannot avoid fixing it. It simply cannot coherently and systematically break down many types of affiliated transactions associated with MNEs. Most notably, it fails to adequately source financial income and intangibles-related affiliated transactions of MNEs. Transfer pricing loopholes are of great concern given the growing impact of MNEs on the global economy, the integration of markets and sectors, and the increase in the volume and sophistication of cross border affiliated transactions. The structural loopholes jeopardize tax regimes' ability to achieve even minimal standards of horizontal equity between taxpayers; additionally regimes are unable to promote notions of income tax progressivity as they are expected to in most democratic countries.

The article provides a novel approach to the old challenge of fixing the international income tax regime. Rather than offering small incremental changes to the current regime, a number of politically reasonable, practical, and administrable suggestions have been proposed. The Proposal supports the claim that it is both necessary and plausible to advance international income tax regime policy out of its deadlock by promoting an Allocation Phase in which significant portions of MNEs' income are formulary allocated.

(* A Visiting Assistant Professor at Northwestern University Law School. I wish to thank Michael Graetz for his thoughtful comments and supervision. I wish to thank Reuven Avi-Yonah, Gadi Benshalom, Jennifer Brickey. Charlotte Crane, Sarah Fritsch, Daniel Markovits, Philip Postlewaite, Ali Nardeli, Adam Rosenzweig, and Tyler Southall for their helpful comments. Most of all I thank Avital, Alon, and Tamar Benshalom for their love and patience.

(1)See Julie Roin. Can the Income Tax Be Saved? The Promise and Pitfalls of Adopting Worldwide Formulary Apportionment, 61 Tax L. Rev. 169 (2008).

(2) See REUVEN S. AVI-YONAH, ET AL., ALLOCATING BUSINESS PROFITS FOR TAX PURPOSES: A PROPOSAL TO ADOPT FORMULARY PROFIT SPLIT (Univ. Michigan Law & Econ., Ohlin Working Paper No. 09-003) (providing the most recent example of the optimistic group); Roin, supra note 1 (providing the most recent example for scholarly work that believes a formulary approach is not feasible).

(3) SOL PICCIOTTO, INTERNATIONAL BUSINESS TAXATION 32-33 (1992).

(4)For an explanation of why this erosion is inequitable and inefficient, see Ilan Benshalom, Sourcing the " Unsourceable": The Cost Sharing Regulations and the Sourcing of Affiliated Intangible-Related Transactions. 26 Va. Tax Rev. 631 (2007) [hereinafter Benshalom. Sourcing the "Unsourceable"].

(5) See generally Daniel Shaviro, The Long-Term U.S. Fiscal Gap: Is the Main Problem Generational Inequity?. 76 GEO. WASH. L. REV. (forthcoming 2008).

(6) David A. Weisbach, Line Drawing, Doctrine, and Efficiency in the Tax Law, 84 CORNELL L. REV. 1627 (1999) (arguing for the value of maximizing efficiency in a world with second-best tax legislation).

(7) See Roin, supra note 1. at 221-22.

(8) Benshalom. Sourcing the "Unsourceable", supra note 4; Ilan Benshalom. The Quest to Tax Financial Income in a Global Economy: Emerging to an Allocation Phase, 28 Va. Tax Rev. 165 (2008) [hereinafter Benshalom, Allocation Phase].

(9)The allocation method from intangibles could be operated unilaterally. In fact, some have argued that the "arm's length methods" through which tax authorities currently allocate intangibles-related transactions are really no more than ad-hoc formulary allocation mechanisms. Benshalom, Sourcing the "Unsottrceable", supra note 4. Operating a unitary system for FMNEs requires tax-coordination of only a limited number of developed countries. Benshalom. Allocation Phase, supra note 8, at 203-04.

(10) Benshalom. Allocation Phase, supra note 8.

(11) Id.

(12) Ilan Benshalom, The Quest to Tax Interest Income: Stages in the Development of International Taxation, 27 Va. Tax Rev. 631 (2008) [hereinafter Benshalom, Stages].

(13) One alternative used frequently is the profit-split transfer-pricing method. Tax authorities use this method to bifurcate the income of functionally integrated MNEs. It requires the different MNE entities to split the total income derived from an activity according to each party's contribution. It aggregates the income derived from the activity and then divides this aggregated figure among the relevant parties. This is very different from the traditional arm's length inquiry, which hypothesizes how unrelated parties would price the transaction. Compliance with this nontransparent case-by-case formulary allocation mechanism is costly for both taxpayers and tax authorities because it is amorphous as a matter of policy design. It depends on documentation and fact-finding requirements. It is also biased in favor of taxpayers, given tax authorities' inferior information and lack of audit and litigation resources. Furthermore, the particularity of the profit-split analysis, as well as taxpayers' ability to partially control whether and how to apply it, render it difficult to generate any general principles out of it. The OECD also deviates from the arm's length standard in its discussion of the branch allocation rules.

(14) For example, the proposed dealing regulations emphasize the use of the profit-split method. Prop. Treas. Reg. [section] 1.482-8(e). 63 Fed. Reg. 11,177 (Mar. 6. 1998).

(15) Benshalom. Allocation Phase, supra note 8.

(16) Id.

(17) Anti-abuse concerns sometimes require broad and simplified rules to prevent an array of potential abuses. On other occasions, they require adjustability so as to counter innovative tax planning schemes.

(18) See supra Part II. C.

(19) JOANN M. WEINER, U.S. TREASURY, OFFICE OF TAX ANALYSIS, USING THE EXPERIENCE IN THE U.S. STATES TO EVALUATE ISSUES IN IMPLEMENTING FORMULA APPORTIONMENT ATCTHE INTERNATIONAL LEVEL 2-6 (1999).

(20) The single most important expense that the Proposal's allocation formula does not include in computing MNEs' net taxable financial income is salaries. Unlike FMNEs, the Proposal offers a formulary method that allocates the financial part of MNEs' income, which may be a comparatively small component of their aggregated income. Consequently, any attempt to attribute non-financial losses or expenses to the calculation of MNEs' net taxable financial income would be administratively hard and conceptually flawed. The administrative hurdle results from tax authorities' difficulty in bifurcating expenses--such as salaries--from MNEs' general-expense pools and attributing them only towards the income of financial activities. Most of these expenses are not attached inherently to MNEs' financial resources but to their activities in a specific location. This bifurcation of expenses is also antithetical with the notion introduced earlier that MNEs' financial income is derived primarily from capital and not from financial services. The notion justifying bifurcating MNE financial income and allocating it through formulary methods is that financial assets' mobility and fungibility carry an enhanced abuse potential under arm's length methods. Thus, since the formulary appropriation method targets primarily the problem of asset mobility--and not the location of financial activities--there seems to be less justification to deduct salaries of financial officers against MNEs' net financial income. Therefore, they should each be deducted from the subsidiaries income and not MNEs' general pools of (net) taxable financial income.

(21) Daniel N. Shaviro, Does More Sophisticated Mean Better? A Critique of Alternative Approaches to Sourcing the Interest Expense of U.S. Multinationals, 54 Tax L, Rfv. 353, 381-82 (2001).

(22) Benshalom. Stages, supra note 13, at 676-82.

(23) Bcnshalom. Allocation Phase, supra note 8. at 199-200 (arguing that payroll serves as a good proxy for the location of FMNEs' intangible assets).

(24) As the author has written elsewhere, some of this income, which arises from intangibles, should also be sourced through formulary methods and the rest by reliance on the existing transfer-pricing methods, particularly the CPM method. Benshalom. Sourcing the "Unsourceable". supra note 4. at 681-82.

(25) See id. at 681-84.

(26) As in the case of FMNEs. the APA alternative can help policymakers to reduce the compliance and auditing costs of private MNEs. See Benshalom, Allocation Phase, supra note 8, at 211 11.133.

(27) The formulary allocation the author offered in a previous article suggests that the relative factor that should be allocated to the tangible properly factor versus the payroll factor should correlate with the relative value of tangible and intangible assets within the MNE. The aggregate value of MNE intangibles would be determined by subtracting their total market value from the value of its tangible assets. This requires. of course, that the formulary allocation of MNEs' financial income be supplemented with another de facto threshold requirement: that MNEs would be a publicly traded company in a recognized stock exchange. The reason for this requirement is that for privately owned MNEs tax authorities may find it costly to make such valuations. Since, however, the vast majority of large MNEs are publicly traded, if the proposal is adopted, it would be uniformly imposed on them. That being said, especially for medium-sized MNEs. a broad implementation of the proposal may result in excessive compliance costs. Although these inequities could be reduced by APAs. policymakers may seek to avoid them altogether by restricting the formulary allocation of financial income only to large publicly traded MNEs. Benshalom, Sourcing the "Unsourceable ".supra note 4.

(28) In the context of the formula proposed for FMNEs, one of the ways tax authorities can attribute income to the sales factor is to allocate part of the income attributed to the Residual Value through a combined payroll and sales formula. See Benshalom, Allocation Phase, supra note 8, at 196-200. The precise weight that should be given to the payroll and sales factors could be determined through an arbitrary standard (e.g., a 1:1 ratio) or according to the total estimated value of consumer-based intangibles.

(29) KPMG, Corporate and Indirect Tax Rate Survey 2008 11,19 (2008).

(30) Benshalom, Allocation Phase, supra note 8, at 208 (providing an example of such planning).

(31) Id. at 206.

(32) Id. at 193-96, 206. This difference merits explanation. Unlike MNEs, FMNEs have a number of distinctive features. First, most major FMNEs operate in an industry confined by extensive regulation. They have an array of broad, albeit limited, categories of professional positions, which make it administratively possible for tax authorities to distinguish these positions. Moreover, these professionals' activities comprise the core of FMNEs value-generating processes. A payroll formulary arrangement that relies on the compensation of these professionals would produce sourcing outcomes that correspond with MNEs' volume of economic activity in different jurisdictions. Second, the entire net income of FMNEs is subject to unitary allocation. Thus, FMNEs that manipulate their payroll factor may reduce significantly their tax liabilities. Because of this, the article emphasized the importance of anti-abuse concerns. Given the administrative difficulties of having a broad definition for employees, the article's solution narrowed it to include only the hard-to-manipulate and core, professional positions.

(33) An exception could be made regarding hedging transactions, through which MNEs engage in future transactions related to some of their operational exposures. Under current law. these types of hedging transactions are limited in scope and should he recorded as hedging transactions at the day in which the position is entered into.

(34) The most startling example is withholding taxes on financial payments. While dividend payments are typically subject to withholding taxes, interest payments and income derived from financial derivatives are typically exempt by double taxation treaties from withholding source taxes. This discontinuity is ridiculous given taxpayers' ability to replicate equity investments with the use of hybrid financial derivatives.

(35) The system will not be purely territorial because foreign sourced income from non-financial transactions (e.g., profits from leasing transactions and part of the profits from exports) would still be taxed on a worldwide income basis.

(36) See Benshalom. Solacing the "Unsourceahle", supra note 4, at 632-33; Benshalom. Allocation Phase, supra note 8, at 196.

(37) MNEs can also receive dividends from unrelated parties. As explained in detail in Part VI, supra, these dividends would be considered financial income.

(38) See supra Part V (discussing the definition of related parties).

(39) Benshalom. Allocation Phase, supra note 8. at 178-79.

(40) Jack M. Mintz. Globalizing the Corporate income Tax: The Use of Allocation, 56 FINANZARCHIV 388. 414-18 (1999). Note, however, that Canada does not employ a corporate group concept. Id. at 415.

(41) Diane M. Ring, Prospects for a Multilateral Tax Treaty. 26 Brook. J. Int'l L. 1699 (2001) (addressing some of the inherent problems of obtaining a good, multilateral tax decision making framework).

(42) See generally Michael J. Graetz & Alvin C. Warren Jr.. Income Tax Discrimination and the Political and Economic Integration of Europe. 115 Yale L.J. 1186(2006).

(43) See Charles H. Gustofson, Robert J. Peroni, and Richard Crawford Pugh, Taxation of International Transactions 54-67 (3rd ed. 2001).

(44)See Christina M. Lyons, The Constitutionality of the Worldwide Combined Reporting Method of Taxation of Multinational Corporations: Barclays Bank v. Franchise Tax Board, 37 B.C. L. Rev. 183 (1995) (describing the United Kingdom's retaliatory threats to tax following California's decision to employ a formulary method to determine Barclays' income tax liability).

(45) See Yesim Yilmaz, Tax Havens, Tax Competition, and Economic Performance. 43 Tax Notes Int'l 587 (Aug. 14, 2006) (arguing that the United States is a tax haven).

(46) Robert T. Kudrle & Lorraine Eden, The Campaign Against Tax Havens: Will It Last? Will It Work?. 9 Stan. J.L. Bus. & Fin. 37. 41 (2003) (providing a brilliant taxonomy of the different types of tax havens and their different functions).

(47) Id.

(48)For a further discussion on this issue, see Benshalom, Allocation Phase, supra note 8, at 173 & n.26, 177 & n.43. 194 n.89.

(49) Both countries could average the allocated figures, they could agree to provide taxpayers with some type of relief for the over-taxation in their jurisdictions.

(50) The figures in parenthesis represent figures used to determine the relative share of each jurisdiction in Lenti's net income under the proposed DTT arrangement and not the actual income allocated to it.

(51) Thus, to accommodate the United Kingdom formula's sourcing. the United States allocated 56.25% of Letni's income to the United Kingdom (because that was the relative percentage the United Kingdom was entitled to under its formulary arrangement), 56.25% to itself and 12.5% to Ireland, as determined by its own formula. Letni's income would be inflated to 125% (56.25 * 2 + 12.5 = 125). Hence, the relative share of the United States under the treaty modified formula would be 56.25/125, or 45%.

(52) For example, the United States could prescribe that only 50% of the financial income losses may be deducted from MNEs' ordinary income. Alternatively, it could prescribe that ordinary income and financial income are separate baskets that cannot be consolidated.

(53) This attribute may help developing countries seeking to attract foreign investments.

(54) See Benshalom. Stages, supra note 13, at' 700, 707 (arguing for the necessity of creating an allocation system that does not require tax regime harmonization and noting the historical failure to create such a system).

(55) David R. Tillinghast, Tax Treaty Issues, 50 U. Miami L. Rev. 455 (1996) Tillinghasl mentions the several key points. First, he notes that treaties "are incredibly slow-moving creatures. They are time-consuming to negotiate and impossible to update on a regular basis." Id. at 455. Second, he notes that treaties are obsolete;

The concepts embodied in the existing tax treaties ... were largely conceived in the days of a "brick-and-mortar" industrial economy ... [For example,] [r]ules designed to apply to the physical delivery of tangible goods or the provision of physical labor, ... do not always work well when applied to the delivery of software or on-line services. At the same time.... the use of derivative financial instruments to bundle or unbundle economic interests, synthesize securities or confer the economic equivalent of the ownership of property without actually transferring that ownership raises treaty issues that require resolution.

Id. at 456. Third, treaties' distinction between debt and equity, for the purposes of withholding source taxes, is somewhat inconsistent with the exemption of income arising from derivative financial instruments, which could be used to create financial positions that are equivalent to ownership over debt and equity resources. See id. at 457-58.

(56) John F. Avery Jones, Are Tax Treaties Necessary?, 53 Tax L. Rev. 1, 17 (1999) (mentioning that DTTs' tendency to exempt capital gain from source taxes encourages taxpayers to seek character arbitrage by structuring their transactions to elicit capital gains rather than ordinary income).

(57) See Benshalom, Stages, supra note 13, at 676, 700 (2008) (discussing the practical and historical difficulties of negotiating and implementing precise rules capturing the economic substance of transactions).

(58) See Benshalom, Allocation Phase, supra note 13. at 216, (59) This offsetting relationship may spark a sophisticated debate regarding whether parties to DTTs should entangle the provisions dealing with the allocation formulas of FMNEs and MNEs to increase their effectiveness.

(60) See id. at 175-76, 179. 203 (characterizing the dynamic of sovereigns competing for the tax base as a prisoner's dilemma, and the Proposal as an alternative to the currently existing responses to this dilemma).

(61) Lee Burns, Commentary, Are Tax Treaties Necessary?, 53 Tax L. REV. 39, 39-42 (1999): H. David Rosenbloom, Sovereignty and the Regulation of International Business in the Tax Area, 20 Can.-U.S. L.J. 267. 268 (1994).

(62) See OECD, Articles of the Model Convention with Respect to Taxes on Income and on Capital art. 7 (Org. for Econ. Cooperation & Dcv. 2003).

(63)An alternative benchmark could be 10% ownership, as required under United States Controlled Foreign Corporations Legislation. I.R.C. [section] 951(b). This is problematic, however, since one can think of many ownership patterns in which a 10% shareholder does not really control a subsidiary even if it is the largest shareholder. A 35% ownership benchmark reduces the likelihood that the main shareholder does not control the subsidiary.

(64) This should be done with reference to the ownership patterns in the markets in which the formulary system is employed. For instance, in markets with dispersed ownership, such as the market for publicly traded corporations in the United States, it may be sensible to reduce this threshold. In Europe, where ownership is more concentrated in the hands of institutional investors, it might be increased. The important point is that roughly the same threshold is used by all developed countries so that the actual percentage would be a matter of political compromise.

(65) This is the same concept as that contained in the related party rules under I.R.C. [section] 267.

(66) See Walter Hellcrstein & Charles E. McLure, Jr., The European Commission's Report on Company Income Taxation: What the EU Can Team from the Experience of the US Slates, 11 Int'l Tax & Pub. FlN. 199, 205 (2004) (noting that a "consolidated apportionable tax base predicated on solely on legal control could lead to manipulation by the careful adjustment of ownership interests in other corporations to minimize tax burdens").

(67) See id.

(68)Graeme S. Cooper, Business Purpose, Economic Substance, and Corporate Tax Shelters: International Experience with General Anti-Avoidance Rules, 54 SMU L. Rev. 83. 92,128 (2001).

(69) Richard J. Vann, Reflections on Business Profits and the Arm's-Length Principle, in THE TAXATION OF BUSINESS PROFITS UNDER TAX TREATIES 133, 142 (Brian J. Arnold, Jacques Sassevillc & Erick M. Zolt eds., 2003).

(70) Benshalom, Allocation Phase, supra note 8; Benshalom. Stages, supra note 13.

(71) These recent developments are in Ihin-capitalization and branch-allocation rules. See Benshalom, Allocation Phase, supra note 8, at 180-91.

(72) The United States source rules, for example, could be accused of accumulating a myriad of distortions and inequities. See Shaviro. supra note 21.

(73) The only notable exception to this is the EU, where the European Court of Justice (ECJ) established in its Marks & Spencer ruling that per-se prohibition of loss consolidation violates the freedom of establishment. Case C-446/03, Marks & Spencer pic v. Halsey, 2006 E.C.R. I-10,837.The ECJ did not, however, delineate a clear notion of when loss consolidation should be permitted and when disallowed. Id.

(74) See generally Ke Chin Wang, International Double Taxation of Income: Relief Through International Agreement 1921-1945, 59 Harv. L. Rev, 73, 81-96 (1945) (discussing the debate of source versus residence taxation when the issue double taxation was addressed in the 1920,1930s and 1940s).

(75) It previously has been noted that:

The source principle is a companion principle to capital import neutrality in that both require the abandonment of residence taxation..,. Under the source principle, a tax cannot be fair. That is, a country following the source principle cannot have a progressive tax or a tax that otherwise relates to the economic wellbeing of its taxpayers. What does occur is unbridled tax competition.

Michael J. Mclntyre, Commentary, The Design of Tax Rules for the North American 

Free Trade Alliance, 49 Tax L. Rev. 769, 772 (1994).

(76) Stephen E. Shay et at, "What's Source Got to Do With It?" Source Rules and U.S. International Taxation, 56 Tax L. Rev. 81, S3 (2002). The residence jurisdiction's entitlement is always secondary in the sense that if levied, it offers some type of relief to taxes paid in the source jurisdiction. One can argue that this concurrence is not a byproduct of recognition of the source jurisdiction's entitlement to first levy tax but a byproduct of the practice in which the source jurisdiction can take the first tax bite.

(77) See JEROME R. HELLERSTEIN & WALTER HELLERSTEIN, STATE AND LOCAL TAXATION: CASES AND MATERIALS 470,542 (7th ed. 2001).

(78) The phrases internation and intranational equity have been coined by Nancy Kaufman. See Nancy H. Kaufman. Fairness and the Taxation of International Income, 29 LAW & POL'Y INT'L BUS. 145, 153 & n.48 (1998).

(79) See Peggy B. Musgrave, Comments in Session on "Foreign Reaction to U.S. Tax Reform". 41 NAT'L TAX J. 365. 366 (1988) (arguing that to determine what comprises fair taxation on the source level requires a consolidated assessment of source corporate taxes and withholding taxes).