Starting the Conversation: A Formulary System For Dividing Income Among Taxing Jurisdictions

The author, embarking on what will be a series of articles evaluating formulary apportionment as a viable alternative to the arm's-length standard, notes the intense disagreement between those on either side of the debate--and the lack of any detailed comparison between the two approaches. In this, his first installment, he sets forth the goals of any system for dividing multinational companies' income among taxing jurisdictions.  


By Michael C. Durst, Contributing Editor

Michael C. Durst has been a practitioner, author, and speaker in the field of international tax and transfer pricing for more than 20 years and served as director of the Internal Revenue Service's Advance Pricing Agreement Program from 1994 to 1997. His most recent work has focused on base erosion and other issues facing developing countries' tax administrations.

This is the first of a series of articles that, taken together, will:

• evaluate whether current arm's-length transfer pricing rules for dividing taxable income among countries should be replaced by a system of formulary apportionment similar to that used by the U.S. states and the Canadian provinces; and

• identify in technical detail how, if one or more countries should decide to adopt a formulary approach, they might do so most effectively.


The project on which these articles embark is unavoidably controversial. Arm's-length transfer pricing rules have been in effect around the world for decades, and they have been vigorously defended on the ground that moving to a formulary approach would involve prohibitive complexity and international confusion.

Against this, critics of the arm's-length approach argue that the premise on which such rules are based--namely, that transactions among commonly controlled members of multinational groups are comparable to transactions among unrelated entities--is incorrect, so that despite intensive efforts over the years to enforce arm's-length transfer pricing rules with reasonable efficiency and effectiveness, no country has yet succeeded in doing so.

Critics also argue that arm's-length rules have been interpreted over the years to allow multinational groups to shift large amounts of income to low- and zero-tax countries, resulting not only in damage to public treasuries around the world but also to a loss of public respect for the fairness of tax systems. In recent years, global media coverage of the shifting of income to low- and zero-tax countries has given the debate between arm's-length and formulary rules substantial public attention, raising the likelihood that one or more governments will give serious consideration to adopting a formulary approach.

Heated Debate, Little Analysis

Perhaps because of the large amounts of money potentially at stake for both taxpayers and governments, the debate between arm's-length and formulary approaches to dividing income has become quite heated. The intensity of disagreement has been so great that for any national government, or an international organization such as the Organization for Economic Cooperation and Development or the United Nations, even to appear to be giving serious consideration to a formulary approach would be seen by many as an unacceptably provocative move in itself.

A result is that despite the energy expended on both sides of the debate between adherents of arm's-length and formulary approaches, the two systems have received little detailed, point-by-point comparison. Further, surprisingly little serious effort appears to have been devoted to addressing, systematically, the technical issues that would arise if a country or countries were to seek to implement formulary rules.

These articles will not take upon themselves the task of trying to resolve the debate between arm's-length and formulary rules for dividing income, but they will attempt to contribute in at least a small way to alleviating the deficit of systematic analysis that seems historically to have afflicted the debate.

This initial article offers a conceptual background for a study of formulary apportionment by describing the objectives of a successful system for dividing income. The articles immediately following will:

• describe the different systems for dividing taxable income that are now in effect around the world, both at international and subnational levels (such as the multistate apportionment system in the United States);

• assess the relative strengths and weaknesses of the systems now in operation;

• discuss the problem of double taxation as it arises under both arm's-length and formulary systems; and

• consider which kind of system is most likely to minimize problems arising from the inconsistent tax treatment of taxpayers by the different countries in which they conduct business.


Articles beyond these will explore in detail the primary technical issues that should be addressed in designing satisfactory rules for international formulary apportionment, including how best to define the “unitary group” to which formulary apportionment of income should be applied and--drawing on historical experience in the United States and Canada--which factors might be included in a formula. In addition, future articles will contain sample statutory and regulatory language for implementing a system for formulary apportionment as well as rules for accomplishing the transition from an arm's-length to a formulary system.

All of these articles, which will be published serially over the next year, inevitably will touch on questions on which informed readers are likely to hold various views, including views different from the author's. In addition, many of the conceptual and technical issues to be addressed in designing a formulary apportionment system are complex and difficult, and resolving these issues satisfactorily requires the exercise of judgment. The hope is that spacing these articles over the course of a year will provide opportunity for constructive debate on policy as well as technical questions. Those who wish to participate are urged to send their comments to this publication.1

Goals of a System

A first step in comparing different possible systems for the international division of income, and in designing an effective system, is to identify the goals such a system should achieve. For anyone familiar with the complexities and uncertainties of international business, and particularly the virtually infinite ways in which multinational businesses generate income around the world today, this task will immediately appear discouraging.

The heart of any system for the geographical division of taxable income presumably should consist of an attempt to identify where a particular taxpayer, engaged in a particular business activity, has generated each identifiable component of added value. Even in relatively simple situations of cross-border business activity, however, the question of where value is added can be difficult and complex--even to the point of appearing impossible to resolve with reasonable confidence.

Auto Manufacturing--a Straightforward Example?

Consider, for example, the relatively straightforward situation of the manufacture of automobiles by a member of a multinational group in one country, and the sale of those automobiles at wholesale, to unrelated dealers, through distribution personnel of the same multinational group based in another country. The income earned from the manufacture and sale of the cars would appear to derive from a number of activities and factors, only some of which can convincingly be attributed to a single country, and the relative importance of which cannot be isolated and measured.

The income derives from the efforts of numerous people performing various functions--research and development, design, administrative, purchasing of materials, fabrication, marketing, logistics, and a host of others. Some of these functions might be performed in the country where the cars are manufactured, some might be performed in the country where the cars are sold, and some might be performed in other countries where the multinational company conducts business. And, of course, sometimes the performance of a particular function, such as R&D, might be spread among a number of different countries.

A complicating factor is that the value added by all of these functions is greatly enhanced by the fact that they are all harnessed together, by a single corporate management, to generate the most attractive product possible for sale on the world markets. That is, the profit produced from the manufacture and sale of the cars arises from the efforts of people performing allthe functions comprising the design, manufacture, and marketing of cars, and doing so together.

Indeed, much of the value created by a multinational business is likely to arise not only from the different functions performed by the personnel located around the world, but also from the fact that those personnel have been integrated into a commonly managed multinational business. There is no theoretical basis on which to conclude confidently that this increment of value--what economists call the economies of integration--is attributable to the contributions of personnel in any particular country.

Further complicating the analysis, factors outside a multinational company's control are likely to influence the level of the group's profit, sometimes strongly. Examples are the relative desirability of the manufacturer's particular brand of cars (while conscious marketing efforts of company personnel may have some effect, much depends on consumer taste); changes in the price of fuel (which can affect relative demand for cars of differing levels of fuel efficiency); changes in consumer interest rates; and particularly, for cross-border activities, fluctuations in currency exchange rates, which can greatly affect the profitability of manufacturing a product in one country and selling it in another.

Even using a relatively straightforward example like this one, in which many of the activities can at least be tied to physical facilities in identifiable locations, there is simply no way of sorting out and valuing the influences of all the interacting factors that contribute to the level of profitability of a manufacturing and distribution activity.

Harder Cases

Moreover, many of the largest businesses in international commerce offer far fewer geographic reference points for use in dividing income than does the manufacture and sale of automobiles. In the global insurance business, for example, there is no equivalent to the manufacturing plant; for companies that design and sell software, there generally is no analogue to the distribution center.

Further, many businesses generate income through heavy reliance on intangible assets such as patents and trademarks, which may have gained their value as a result of events occurring in many jurisdictions, some of which (such as the gradual development of consumer loyalty) may be impossible to observe. It is thus not surprising that an authority no less lofty than the U.S. Supreme Court, in a case involving income division at the state level, has said that allocating income among various taxing jurisdictions “bears some resemblance … to slicing a shadow.”2

Indeed, it quickly becomes apparent to anyone involved in cross-jurisdictional taxation that an indisputably correct division of income, by reference to some kind of identified theoretical norm, cannot possibly be achieved. Any system for dividing income among jurisdictions must be assessed according to criteria of practical suitability rather than of theoretical purity. At best, this system can be “good enough.” That is, it can provide answers, with reasonable consistency, that are likely to be accepted as plausible and unbiased by both taxpayers and tax authorities.

Criteria for a Satisfactory System

A system for dividing international income must be accepted widely as fair, in the interest of political stability. It also must operate within the capability of real-life, resource-constrained companies and revenue administrations--that is, taxpayers' cost of compliance and the government's costs of implementation and enforcement must be reasonable.

Results under the system should be reasonably predictable, and it should not seek to influence business decisions (for example by offering incentives to locate factories or other places of employment in particular jurisdictions).


A system for dividing income is likely to remain stable only if taxpayers, as well as voters and other political decision-makers, perceive it to be reasonably fair. The concept of fairness in connection with dividing income is elusive; it is not possible to judge the theoretical correctness of any particular set of rules for dividing income, and therefore there is no scientific way of evaluating a particular system's fairness. Nevertheless, interested parties can and do make judgments about fairness. When either business interests or large groups of voters and policy makers perceive that a system produces anomalous results, political pressures to replace the system will arise.

From a taxpayer's standpoint, whether a system divides income fairly depends on whether it reasonably approximates the company's perceptions of relative income or losses in the different jurisdictions where it operates. An income tax, as a general matter, serves as a dampener of a business's economic volatility. That is, in times of success, an income tax reduces the taxpayer's resulting economic profit; when a company incurs losses, the deductibility of the losses (perhaps through carryovers) reduces the financial consequences of the loss.

This dampening effect is generally seen as a desirable feature of an income tax. If the multinational company views this dampener as operating effectively--mitigating losses while imposing high tax liabilities only where operations have been successful--it is more likely to perceive the tax rules as fair.

In the context of rules for dividing income, however, there is an important built-in ambiguity in determining whether a tax system appropriately tracks the level of a taxpayer's net income during a particular period.

Consider, for example, a multinational business that is thriving in much of the world but experiences a setback in a particular country owing to facts (for example, the occurrence of a national disaster) unique to that country. Should the company's tax burden in that country be reduced, perhaps to zero, even though the taxpayer as an integrated economic unit has earned high overall returns from its worldwide business? Or should each country in which the taxpayer conducts business be permitted to reach a reasonable share of the taxpayer's income from global operations without regard to the taxpayer's relative profitability in any particular country during a given year?

Taxpayers, for the most part, probably view their global, rather than local, income as more appropriately governing their tax liabilities because global income affects their overall ability to pay tax. Some governments, however, may feel that a taxpayer generating added value in a particular country should pay tax reflecting that value even if its global economic performance has been unfavorable.

As will be discussed in future articles, the differing views about basing tax on local as opposed to global economic performance have posed a continuing problem in the design of rules for dividing income. There appears to be no satisfying intellectual basis on which to resolve this dilemma--but it must be resolved in some manner.

In addition to businesses and governments, a country's voters--those who, in democratic systems, ultimately are entitled to control a country's tax policies--must perceive rules for dividing income as fair if the rules are to survive over the long term. From the public's standpoint, a fair business tax system arguably is one that subjects all companies to tax in reasonable proportion to the amount of activity they conduct in the country. In recent years, the perceived tendency of transfer pricing rules to allow movement of income from a country where business is conducted to low- and zero-tax jurisdictions where little if any observable business is conducted have fueled public demands for reconsideration of international tax rules, including rules for dividing income among countries. It is fair to say that heightened public interest in the movement of income to low- and zero-tax countries, fueled in large measure by extensive media coverage around the world,3 today represents an important political influence in deliberations about international tax policies.

Administrability and Cost-Effectiveness

A tax system should achieve its objective--the transfer of the legally correct amount of money from taxpayers to the government--at a cost that is low relative to the volume of taxes collected. To that end, it should:

• minimize the need to compile and analyze factual data;

• minimize the need for subjective judgment;

• maximize predictability; and

• minimize economic distortion.


Minimizing Information-Related Costs

Multinational businesses already are required, for accounting and various regulatory purposes, to compile large volumes of data about their businesses. Rules for dividing income should minimize the amount of information taxpayers must gather and retain beyond what already is required for other purposes. Thus, a system for dividing income should, to the extent possible, require only information already contained in the taxpayer's financial books and records. Limiting the amount of information companies must retain reduces their costs as well as those of tax administrators, who generally are expected to audit and evaluate the additional data.

Avoiding Subjectivity

A system for dividing income also should rely as much as possible on computations that can be performed directly on the basis of quantitative financial data. In particular, the system should minimize the need for the taxpayer to apply subjective judgment to the interpretation of financial data in order to translate the data into a division of its income. Generally, the more a system relies on a subjective evaluation of a taxpayer's business operations, the higher will be a taxpayer's costs in estimating its annual tax liability each year.

Further, a system that relies extensively on a company's subjective evaluation of factual data puts the tax administration in the difficult position of attempting to evaluate and second-guess the taxpayer's determinations, perhaps to the satisfaction of a court. The more tax rules force businesses and tax administrations to rely on subjective judgment, the greater the likelihood of intractable and expensive tax controversies.

Maximizing Predictability

Financial predictability is important to economic decision makers in both the private and public sectors. In the private sector, an investor is more likely to make an investment, such as an investment in a cross-border business activity, if it can predict with reasonable certainty the after-tax income it will earn from the venture. Thus, predictability of result, under a system for dividing taxable income among countries, is likely to encourage cross-border investment and therefore to promote global economic growth.

Accordingly, countries should adopt systems for dividing income that allow investors to predict how much income each country with potential taxing jurisdiction will claim the right to tax. If the rules in one or more countries are vague--for example, if the investor can expect to obtain reasonable certainty only years into the future, perhaps after a period of controversy with one or more governments--taxpayers may be discouraged from entering into cross-border investments, inhibiting economic growth.

Certainty about a country's rules for dividing income internationally is important to governments as well as investors. Governments often must budget for multiple-year periods, and the more confidence they have in the predictability of tax revenues, the more willing they will be to commit to making investments in infrastructure and other long-term public needs. In short, from the standpoint of both the private and the public sector, social welfare is maximized when the rules enable taxpayers and governments to predict results reasonably accurately in advance.

Minimizing Economic Distortion

Taxes on business income inherently and unavoidably discourage investment and hence distort economic decision making. This is true regardless of the system that is used for dividing income among jurisdictions. The only way fully to eliminate the economic distortion caused by a tax on business income is to eliminate the tax. Nevertheless, it may be possible for the choice of a particular system to affect the nature and extent of the economic distortion arising from a tax.

For example, the U.S. states have tended over the years to base income apportionment more heavily on the distribution of sales revenues among the different states than on the distribution of payroll expenses or the location of tangible property. The states have emphasized the sales factor based largely on the view that sales-based apportionment tends to minimize disincentives to locate employment and physical plant within a particular state. As another example, it has been argued that the current U.S. international tax rules encourage U.S.-owned businesses to establish plants and operations overseas rather than at home, because it is easier under the laws for U.S.-owned businesses to transfer income to low- and zero-tax countries from non-U.S. jurisdictions than from within the United States.4

Determining the extent to which particular kinds of rules will affect the distortions arising from a tax system is a difficult and sometimes subtle exercise, and opinions can differ concerning the significance of particular distortive effects. Nevertheless, the extent to which particular rules for dividing income might affect the kinds and extent of economic distortions resulting from the imposition of a tax is an important consideration in the evaluation of alternative systems of rules.

1 Readers are invited to email the author at or the editor at

2 Container Corp. v. Franchise Tax Board, 463 U.S. 159, 192 (1983).

3 See, for example, Jesse Drucker, “IRS Auditing How Google Shifted Profits,”, Oct. 13, 2011; Charles Duhigg and David Kocieniewski, “How Apple Sidesteps Billions in Taxes,” The New York Times, April 28, 2012; Rajeev Syal and Patrick Wintour, “MPs Attack Amazon, Google and Starbucks,” The Guardian, Dec. 2, 2012; and Hugo Duncan and Tamara Cohen, “Starbucks 'Treats Tax Like a Church Collection Plate’: Treasury Chief Secretary Attacks Coffee Chain,” Daily Mail, Dec. 9, 2012.

4 Edward Kleinbard, “The Lessons of Stateless Income,” 65 Tax Law Review 99 (2011).