Tax Notes logo

Arm’s Length: Principle or Cult?

Posted on Oct. 26, 2020
Bill Parks
Bill Parks

Bill Parks is a retired finance professor and founder of 100 percent employee-owned NRS Inc. in Moscow, Idaho.

The author thanks Mary Hollenbeck, Kenzie Lauritzen, and particularly David Morse for their contributions to this article.

In this article, the author questions the arm’s-length principle, which he argues enables international tax avoidance, conceals problems, and fails to produce decent tax regimes.

Copyright 2020 Bill Parks. All rights reserved.

The OECD tax committee, and now the United Nations tax committee with its article 12B, keep edging closer to sales factor apportionment. The formerly monolithic support for the arm’s-length principle, separate accounting, and permanent establishment requirements continues to fray. Questioning the old system is not only appropriate but overdue. It’s time to end any tax regime that disadvantages domestic companies competing with multinational enterprises.

Despite being the accepted doctrine for over 90 years, from the beginning, the arm’s-length principle was subject to manipulation.1 It’s time to retire these failed principles that enable international tax avoidance, hide problems, and fail to produce decent tax regimes. As the late Edward D. Kleinbard pointed out, the OECD “have made their lives very hard for themselves by insisting on the arm’s length principle as an untouchable sort of axiom because I don’t think there are ways in which one can fairly price internal intangibles within a multinational group . . . . It’s nonsensical to chase a model that is inconsistent with both reality and with the best economic theory of the firm.”2 When advocates defend the arm’s-length principle in the face of overwhelming evidence of its defects, they become a cult or, at a minimum, guilty of cultish behavior.3

The OECD historically defended the so-called principles despite mounting pressure to make changes. Kleinbard called the principles untouchable, conjuring up a caste system. However, while “untouchable” can refer to the poor and downtrodden, in this case, domestic businesses are the aggrieved parties because they must unfairly compete with the privileged MNEs that pay little or no corporate tax. The privileges are untouchable as permitted by the arm’s-length principle rules.

It is unfair to blame only international organizations, such as the OECD, for their continued support. However, this obstinate refusal to accept any other perspective, despite the mountains of evidence of failure, does suggest a cultlike following. The three MNE examples that follow show some of the many ways that corporations limit their profits in high-tax countries.

Example 1: As soon as pharmaceutical company A is granted a patent, it sells the patent to its Bermuda subsidiary for a nominal amount. Then, if it becomes a successful drug, the Bermuda subsidiary leases rights to produce and market the drug to its U.S. subsidiary at such a high price that, despite company A’s billions in worldwide profits and the majority of its sales in the United States, the subsidiary has shown consistent and substantial losses in its U.S. operations.

Example 2: Company B’s U.S. subsidiary lends money to customers in the United States at an average interest rate of 4 percent. Company B’s Cayman Islands subsidiary finances the U.S. subsidiary and charges an average interest rate of 6 percent. Therefore, despite having a successful U.S. operation, the U.S. subsidiary doesn’t show a profit.

Example 3: Company C has most of its sales in the United States, but as a foreign company, it pays very little federal income tax. It uses its reinsurance subsidiary in Bermuda, where it shows much of its profit. Further, company C has an agreement with Bermuda that if Bermuda charges corporate taxes in the future, company C will not be liable for corporate taxes until 2035. Thus, it insures itself against having to pay taxes in Bermuda, despite legal changes. This tax planning may prove an effective deterrent if the OECD pillar 2 minimum tax would otherwise require it to pay corporate taxes on its Bermuda profits.

From Principle to Cult

In the above examples, the arm’s-length principle was supposed to ensure that goods and services were transferred between countries at the prices paid by unrelated parties. However, the actual application allowed a wide variety of prices because so many products are proprietary and they have no open market equivalents. By taking advantage of those price differences, adherents glorified a practice that was flawed from the beginning. It is unclear when the principle became a cult. By the time transfer pricing specialists were hired by MNEs to minimize corporate tax liability, it was clearly cultish. It had no claim to be a principle, unless paying less tax than domestic competitors could be called principle. The words of experts in the field of arm’s-length pricing reveal what is involved.

One firm specializing in the profit-split method is Transfer Pricing Solutions Asia. It bills itself as a “boutique transfer pricing firm located in Singapore who works directly with your team, applying our experience and expertise in transfer pricing to provide, prepare, document, and assist in defending your international related party transactions.” Why does a firm need to have its related-party transactions defended? Perhaps because within the first few pages, transfer pricing books invariably define the purpose of international transfer pricing as “minimizing the taxes paid by the firm,” or its equivalent.

In the age-old fable, observers of the emperor’s new clothes knew they were supporting a lie, and yet they applauded the beauty of nonexistent garments. The same goes for the arm’s-length principle fanatics, praising the beauty of transfer pricing, and turning a blind eye to the distortions separate accounting enables. It took just one parade and a little child to unmask the emperor’s lack of clothes, but it has taken 90 years of tax avoidance to admit that the arm’s-length principle has no legitimate use in the modern world. Despite mounting evidence, its loyalists remain firm.

So what is to be done? First, admit that the arm’s-length principle has no legitimate claim to be an honest tax broker in the modern world. Second, recognize that the forces behind minimizing tax are too powerful to allow the arm’s-length principle to be completely abandoned. The appropriate action for now is to remove some, but not all, profits from the total and allocate them to market countries.

Adam N. Michel suggests that the OECD consider the experience of the various states that have had involvement with apportionment for over 100 years to learn how countries might react to apportionment.4 He mentions the earlier three-factor standard of payroll, property, and sales. For many years, that was how almost all states taxed corporations, and it was stable. The consensus later eroded because states realized they could make themselves more desirable to companies if they taxed only sales. Michel states that any cartel will be unstable, as will any uniform formula-based corporate tax system. However, only the Chicago School of Economics followers believe that cartels are necessarily unstable and prone to cheating. The American experience is that despite billions of dollars in fines, cartels exist.5

Michel suggests that the OECD is moving away from separate accounting and PE, which he calls longtime international tax principles. As explained above, they can no longer be principles, but enablers of tax avoidance. Again, he builds on state experience to advise the OECD. Although the rest of his observations and suggestions are spot on, he cites the state experience changing from three-factor to more heavily weighting sales as the cause of the lowering of tax rates. However, correlation does not necessarily mean causation. In this case, the rise of the TEA (taxed enough already) party, with its low-tax mantra, coincided with the shift from three-factor to more heavily weighting sales or on sales alone.

Cara Griffith, CEO of Tax Analysts, suggests that while a single sales factor may be inevitable, it shouldn’t prevail in state cases. “A strong argument can be made that the sales factor is a poor indicator of a company’s activity and should be minimized and that property and payroll would be better indicators. [However,] states are seeking to encourage job creation and investment within their borders and shift some of the tax burden to out-of-state companies.”6 On the international stage, a move to taxing based on sales would correspond with Griffith’s statement, except the shifting of the tax burden would be to the MNEs that are now avoiding tax at the level that domestic companies pay.

Of course, this requires a shift away from using only PE to determine a tax nexus. PE relies on a physical presence or the presence of an agent with “habitual” engagement in the tax jurisdiction. The information age has disrupted these antiquated and much-abused 20th century concepts. Both Griffith and Michel are concerned that apportioning based on sales is somehow suspect and that companies should be taxed based on their activities instead. However, most in the business world consider the overriding factor to be sales to a customer. Without the customer, all the other factors are just costs that have no value except to produce losses. A company cannot outsource a customer.

Profits derived from sales also reflect a more accurate taxation model to account for the world of 2020, not activities. Proposals to alter tax jurisdictions and method toward the basis of sales can be found in various publications, from the 2007 Hamilton Project paper7 to a September 2020 Tax Notes article.8

MNE resistance with support from Treasury Secretary Steven Mnuchin9 reflects the faltering yet persistent resistance from within the system. Fear of systemic change has brought out transfer pricing experts to warn the OECD and the U.N. to stay away from real reforms.10 Sales factor apportionment may be the best solution for most countries, but realistically, an initial compromise proposal will be required. That compromise should have a clear set of parameters to reduce tax avoidance through arm’s-length pricing or a related method.

There are three notable proposals for improving the taxing of MNEs: one from the U.N. tax drafting subcommittee on article 12B; pillar 1 from the OECD; and a proposal from the Independent Commission for the Reform of International Corporate Taxation (ICRICT). The solution might be to draw from these three proposals to create a compromise that would also allow for the appropriate evolution to a stable taxing system.

The U.N. Article 12B Proposal

The U.N. Department of Economic and Social Affairs tax committee proposed the new article 12B to tax “income from automated digital services.” The proposal addresses digital tax issues and makes some intriguing proposals, including a new taxing right based on sales to a country. It specifically rejects a nexus requirement and substitutes a revenue one, and it implicitly covers MNEs with much lower sales. But the most helpful part of the proposal is the 30 percent of a nonresident digital MNE’s profits to be allocated to market countries.

Article 12B suggests that destination countries can tax 30 percent of the proportion of a nonresident digital company’s profits that corresponds to the percentage of sales in the country. Also, the proposal doesn’t require any threshold for taxing and explicitly rejects any PE requirements. The proposed legislation focuses on digital services and ignores tax avoidance allowed through other means and industries.

The OECD Pillar 1 Proposal

The OECD has been wedded to the so-called arm’s-length principle, but at last, perhaps reluctantly, conceded that a part of residual profits would be available for taxing by market countries. The new taxing proposal consists of two proposals (reduced from three).11 It also doesn’t cover extractive industries. Pillar 1 establishes a new taxing right for market jurisdictions over a share of residual profit calculated on a unified basis.

This new taxing right on multinational groups also does not require PE. The critical aspect of the reallocation of taxing rights to market jurisdictions is how to divide a portion of an MNE group’s profit. Under the OECD’s proposal, routine profits and residual profit are to be set by specified percentages. To be subject to this tax (in-scope), the OECD requires the businesses to be automated digital services or consumer-facing businesses. This scope is too restrictive to be fully effective, but it isn’t the only limitation.

The threshold initially included a global revenue minimum of €750 million (roughly $885 million) that might be reduced gradually.12 These high thresholds leave tax jurisdictions and their domestic businesses at the mercy of those who can creep under the minimums. However, the initial amounts haven’t been finalized. Some smaller unitary groups and the various countries’ tax administrations will be overwhelmed if the implementation occurs on a specific start date. Pacing the implementation through threshold limits to be reduced every year solves a significant technical difficulty.

Unfortunately, although the OECD tiptoes into apportionment, it still relies on some arm’s-length pricing and transfer pricing. Part of the proposal taxes a portion of residual profit in market countries. Even though the proposal may define residual profits in specific terms, there is still room for tax planning in the division between residual and ordinary profits. The OECD’s unwillingness to entirely disavow this flawed approach in its reforms indicates it will be “retaining Transfer Pricing rules where they work relatively well.”13

The ICRICT Proposal

ICRICT’s independence allows it to make proposals without worrying about MNEs or their supporters’ reactions. ICRICT helps less-developed countries tax corporations, particularly MNEs, that now often pay them little tax. While the OECD entertains a limited global approach, ICRICT proposes that all corporate profits be allocated half on sales and half on personnel based on headcount.14

While ICRICT suggests a one-size-fits-all solution, including headcount, it remains more suited to less-developed economies. The ICRICT formula is particularly appropriate for those less-developed countries that rely heavily on extractive industries. Developed market jurisdictions benefit most by allocating profits based on sales. However, some developed markets contain extractive industries or commodities. From a tax standpoint, commodities share many of the characteristics of extractive industries.

Combining the Proposals

Each proposal has clear positives and negatives. However, the right ideas are in discussion now. Of course, transfer pricing professionals lament the move away from the arm’s-length principle.15 The U.N.’s proposal was declared unworkable because it believes the United States would never settle for 30 percent.16 That might be an appropriate counter, but not if the alternative is 100 percent allocated to market countries.

However, the U.N. 12B proposal is one of the most substantial and obvious ideas to address corporate tax avoidance. By taxing 30 percent of the proportion of a nonresident company’s profits that corresponds to the country’s percentage of sales, the system becomes much more straightforward. Further, U.S. companies, especially domestic ones, desire a simpler territorial tax. The U.N. 12B proposal’s primary flaw is the narrow scope limited to digital companies, not the percentage allocation.

The OECD pillar 1 proposal is desirable to U.S. reformers because it widens the scope of the new taxing right to consumer-facing companies, instead of limiting it primarily to digital U.S. companies. Other industries, such as pharmaceuticals, represent a significant portion of the U.S. tax avoidance losses.17 The OECD tax committee has also considered the technical difficulties of implementation and has suggested doing it gradually. Much to the transfer pricing proponents’ dismay, the case-by-case nature of the OECD reform would be replaced with a specified percentage, significantly reducing after-the-fact tax liability disputes.

If the OECD would accept the U.N. 12B proposal of 30 percent of profits18 to market countries, it would make significant headway in resolving the tax avoidance techniques. The OECD would also have to be willing to reduce its threshold on global revenue over time. Granting additional time allows smaller unitary business groups to transition to this system.

The OECD initial threshold limited to company groups with revenues exceeding €750 million is too high, but it makes sense as an initial level with phased reductions over five years. After five years, any unitary group with sales of more than €50 million or the equivalent should be subject to tax in its market jurisdiction. The lower limit might be further reduced depending on experience.

The phased-in implementation proposal would resolve the legitimate concerns of smaller company groups coping with tax changes around the world. The U.S. sales tax is far more complicated than the international corporate tax system, yet much smaller companies have to navigate it because of the 2018 Wayfair decision.19 Software systems help companies of all sizes comply. Requiring almost all companies to pay corporate tax would soon result in the offer of software assistance from multiple companies at reasonable prices as they learn from the examples set by larger companies.

ICRICT will be unlikely to accept a compromise based on sales only because it focuses on many emerging and developing countries without developed retail markets. Therefore, undeveloped countries that export raw materials, commodities, or manufactured goods can and should be treated differently, but by the nature of the industries they primarily tax. Separating extracting and commodity products into an alternate formula could yield beneficial results for all countries. Market countries would benefit from apportioning 30 percent of all profits based on sales, except extractives and commodities. Developing nations would benefit from the commodities and extractives industries formula of 15 percent on sales and 15 percent on headcount for all corporations.

The longtime warnings against double taxation are misguided. Double taxation is acceptable when taxing individual income. In the United States, individuals may pay income tax to federal, state, and local governments, receiving benefits from all three. MNEs likewise could pay income tax in destination and residence countries, and if they object, they could decline to participate by not selling in the destination country. As long as the tax situation is known in advance, double taxation should be accepted. Mitigation could be offered, but is not required by residence countries.

This proposal is more stable than anything yet suggested, but it will still be transitory because countries, like states, will recognize that taxing more profits based on a sales formula increases not just the fairness between domestic and international companies, but the amount of revenue collected. They are, therefore, likely to increase the percentage taxed on destination sales, and later perhaps some, like many states, will go all-in and tax entirely based on sales.

Both the OECD and the U.N. proposals expect many changes to come through bilateral treaties and domestic law changes. Therefore, tax legislation would generally have an effective date 18 months or two years in the future to allow time for modifying existing treaties. In anticipation of this transition, the option should be made available for nations to recognize a full sales formulation and a 50 percent split (sales and headcount) in commodities and extractives. While unrealistic to expect the OECD to bless a shift to taxing solely based on sales, the corporate tax systems need to be simplified. Accepting this solution as a transitional proposal would allow the simplification to start.


Although sales factor apportionment has proven to be the best and most stable way to tax corporations at the state level, powerful forces preclude full global enactment at this time. The OECD is subject to strong forces from member countries that have dug in to support the status quo and fight the automatic widening of countries’ taxing authority. Kleinbard pointed out the problem. However, no one listened, and unfortunately, he is no longer here to pursue it. Because of its many problems, it is past time to put the arm’s-length cult to rest. Calling it a principle has delayed its demise, but now that the OECD and its government members have studied the problems, perhaps they are now ready to create a stronger system by making some important changes.

By limiting the proposed changes to a simple allocation to all countries, it may be possible to make a change grounded on the U.S. states’ experience. Though the United States’ resistance to digital services taxes purports to be about fairness, it owes much to the desire to protect the most influential U.S. MNEs. Making the changes universal removes any question about fairness. Taking a go-slow approach may be the best option at this time. As countries become aware of its advantages, formula usage will likely erode upward. Except for commodities and extractive industries, sales factor apportionment will likely become the standard for most countries.


1 Multinational enterprises importing or exporting commodities such as wheat or coal used subsidiaries in tax havens to provide insurance, financing, logistics, and other inputs to avoid profits in any high-tax country.

2 In an interview with Erik Cederwall of the Tax Foundation, Making Sense of Profit Shifting” (May 15, 2012).

3 A misplaced or excessive admiration for a particular person or thing.

4 Michel, “What the OECD Can Learn From U.S. States About Corporate Income Apportionment,” Tax Notes Int'l, Dec. 2, 2019, p. 821.

5 Jonathan Tepper and Denise Hearn, The Myth of Capitalism 23 (2018).

6 Cara Griffith, “Single Sales Factor Apportionment May Be Inevitable, But Is It Fair?” Forbes, Sept. 18, 2014.

7 Kimberly A. Clausing and Reuven S. Avi-Yonah, “Reforming Corporate Taxation in a Global Economy, a Proposal to Adopt Formulary Apportionment,” Hamilton Project (2007).

8 Jeff Ferry and Bill Parks, “Public Company Corporate Tax Under the TCJA and Sales Factor Apportionment,” Tax Notes Federal, Sept. 21, 2020, p. 2229.

9 Mnuchin letter to José Ángel Gurría, Treasury Department (Dec. 3, 2019).

10 Oliver Treidler, “Are the U.N. and OECD Walking Together Into the Formulary Apportionment Abyss?Tax Notes Int’l, Aug. 31, 2020, p. 1191.

11 Ryan Finley and Stephanie Soong Johnston, “New Detail on OECD’s Pillar 1 Proposal Emerges in Draft Report,” Tax Notes Int’l, Aug. 10, 2020, p. 718.

12 OECD Centre for Tax Policy and Administration, “Pillar One Blueprint,” Steering Group of the Inclusive Framework on BEPS (Oct. 2020).

13 OECD, “Secretariat Proposal for a ‘Unified Approach’ Under Pillar One,” Public Consultation Document (Oct. 9, 2019).

14 ICRICT, “International Corporate Tax Reform: Towards a Fair and Comprehensive Solution,” at 5 (2019).

15 Treidler, supra note 10.

16 Id.

17 Michelle Chen, “Pharmaceutical Giants Have Avoided Paying About $2.3 Billion in Taxes in the US Alone,” The Nation, Oct. 2, 2018.

18 Taxable income for allocation should be based on International Financial Reporting Standards or TGAAP depending on the headquarters of the MNE.

19 South Dakota v. Wayfair Inc., 585 U.S. 138 (2018).


Copy RID