Published in the Journal of International Taxation, Thomson Reuters, Vol. 35, No. 4, April 2024.

Cost Allocation Base Selection for Transfer Pricing Purposes

By Anthony Malik

INTRODUCTION

The successful implementation of a cost-based transfer pricing policy necessitates the group of companies to devise robust internal cost allocation methodologies. The reasonability of the cost allocation method employed is largely contingent upon the prudent choice of an appropriate cost allocation base. The allocation base, a representative of the phenomena that drive costs, is the central variable or basis that ultimately governs the entire allocation split of a given cost allocation method. It just so happens that selecting an allocation base from among the identifiable alternatives can be a particularly tricky and effortful task.

In the ensuing sections of this paper, we will discuss considerations for selecting an allocation base for transfer pricing purposes in general and for the transfer pricing of intercompany services in particular.[1] Due to the abject lack of any elaborative, substantive provisions in the statute,[2] we will go directly to the relevant regulatory guidance to which the courts have historically and customarily granted significant deference.[3] After reviewing the relevant tax law we will infer practical implications and considerations based on what one can reasonably glean from the authoritative guidance. To keep the scope of the topic manageable, in the conversation to follow, we will use the term cost in the conventional sense. In practice the term cost is not always literal and can be imbued with some rather abstract qualities once microeconomic factors are weighed in.

REVIEW OF THE TAX REGULATION ON POINT

Regarding selecting an appropriate cost allocation base, Treasury Regulation Section 1.482-9 does not provide procedural details for taxpayers to follow. It does, nonetheless, succeed in conveying the basic intuition behind selecting an allocation base. This imparted intuition serves to subsequently lay the foundation to support one’s practical considerations. Treasury Regulation Section 1.482-9 has a clear instructional preference to emphasize judging the relative benefits enjoyed by each recipient of the intercompany services.

If the benefits of a cost-inducing activity undertaken by a member of a controlled group extend to other members of the controlled group, the costs must be allocated among the benefitted members.[4] For this purpose, “An activity is considered to provide a benefit to [a member] if the activity directly results in a reasonably identifiable increment of economic or commercial value that enhances the [member’s] commercial position.”[5] The costs for such value-adding intercompany services “must be allocated among the [members] based on their respective shares of the reasonably anticipated benefits from those services.”[6] Thus the costs must be allocated to the respective members in a manner commensurate with the quantum and quality of the benefit enjoyed by the respective members rather than on the basis of some “generalized or non-specific benefit.”[7]

The regulation’s explicit guidance regarding selecting an allocation base succinctly specifies that “consideration should be given to all bases…, including, for example,… assets, sales, compensation, space utilized, and time spent.”[8] The regulation also provides some helpful examples in an attempt to concretize its guidance on selecting an allocation base.[9] It is worthwhile to consider the essential kernels of a selection of these examples due to their relevance to our discussion:

  • In the first case,[10] consider a parent company that centrally processes invoices for all its subsidiaries. In evaluating the respective shares of the reasonably anticipated benefits of each member, the total value of the merchandise sold may not be a reliable measure because it “does not bear a relationship to the anticipated benefits from the underlying [intercompany] services.”[11] Contrarily, “The total volume of orders and invoices processed may provide a more reliable basis for evaluating the shares of reasonably anticipated benefits from the data processing services.”[12]
  • In the following example,[13] a parent company centrally performs human resources functions for all its subsidiaries. “In evaluating the shares of reasonably anticipated benefits from these centralized services, the total revenues of each subsidiary may not provide the most reliable measure of reasonably anticipated benefit shares because total revenues do not bear a relation to the shares of reasonably anticipated benefits for the underlying services.”[14] Contrarily, “Employee headcount or total compensation paid to employees may provide a more reliable basis for evaluating the shares of reasonably anticipated benefits from the [intercompany] services.”[15]
  • Also instructive is an example[16] wherein a parent company, Company A, with two wholly-owned subsidiaries, Company B and Company C, hires an outside consultant for advice regarding a manufacturing process used by Company A and Company B. Company C, on the other hand, uses a manufacturing process sufficiently different such that it does not benefit from the outside consultant’s recommendations. In this example, when applying the chosen allocation base (i.e., sales), Company C’s weighted allocation base is excluded from the overall cost allocation equation. Accordingly Company C is excluded from the assignment of the costs incurred by Company A to hire the outside consultant. The point to absorb is that upon judging the suitability of an allocation base, whenever necessary, taxpayers must also consider implements to exclude the weighted allocation bases of any non-benefitting members.[17]

The review of the relevant paragraphs of Treasury Regulation Section 1.482-9 should make it apparent that the suitability of an allocation base is highly context-dependent. Rigid legal rules dictating the uses of allocation bases are unlikely to consistently produce outcomes that closely parallel the economic substance of the pertinent underlying business functions and arrangements. It is seemingly for this reason that the regulation liberally states that “Any reasonable [allocation base] may be used to allocate and apportion costs under [Section 482]”.[18] The guidance evidently leaves much to the discretion of taxpayers in devising their own cost allocation methodologies because taxpayers themselves, as opposed to an independent administrative agency, are in a position to be most familiar with the relevant business context.

Along this vein, there is further assurance that the IRS will potentially, though not conclusively, consider reliable taxpayers’ internally-devised cost allocation methods for their various legitimate non-tax business purposes.[19] Thus if a taxpayer reasonably concludes that it allocated its intercompany service costs “on a basis [(i.e., an allocation base)] that most reliably reflects [each members’] respective shares of the reasonably anticipated benefits attributable to such services… the Commissioner may not adjust such allocation basis.”[20]

PRACTICAL IMPLICATIONS AND CONSIDERATIONS

As can be gleaned from the regulatory guidance hitherto discussed, in practice, the main difficulty in selecting an allocation base is that one can often initially identify multiple ostensibly suitable allocation bases that bear a causal relationship to the allocable indirect costs accumulated in a given cost pool. The uncertainty regarding the suitability of an allocation base is an uncertainty regarding the nature of the cost pool—that is, the difficulty in judging the suitability of an allocation base stems from a difficulty in understanding how the allocation base pertains to the cost pool. Since there is no directly traceable causal relationship between the cost pool and the allocation base, there is usually no single obvious choice. Resultantly, selecting an allocation base necessarily involves subjective professional judgment.

Good judgment in the selection of an allocation base is critical because it can be highly consequential to the pertinent cost objects during the cost allocation process. Different allocation bases will invariably bear different causal relationships to their cost pools. The difference between these causal relationships can be quite significant. One must not mistake these differences as something entirely random or devoid of meaning. Rather, it reflects the differences between the relationships between the considered cost pool and the business processes and activities underlying the respective allocation bases. An examination of these underlying business processes and activities aids ascertaining the degree of relevance of each allocation base to the given cost pool.

The degree of an allocation base’s relevance is dependent upon the quality, degree, and strength of the causal relationship it bears to the considered cost pool. There need to be present some aspects that conjoin these two variables. Naturally, the relevance of an allocation base is vitiated if such vital aspects are weak. On the other hand, the stronger the causal relationship between the allocation base and the cost pool (i.e., the greater the allocation base’s degree of relevance to the cost pool), the greater its efficacy in relating the allocable indirect costs to the cost object. The importance of selecting the most relevant allocation base identifiable or available is that even a marginally less relevant allocation base can result in too much or too little overhead being assigned to a given cost object thereby distorting the arm’s length charges.[21]

As can be seen, wisely selecting an allocation base requires the practitioner to holistically view and understand the business without being hemmed in by disciplinary boundaries. Selecting an allocation base is inescapably somewhat arbitrary. Consequently, the entire process of cost allocation is inherently imprecise. Notwithstanding, upon implementation some methodologies will clearly prove themselves to be inferior to others. Thus practitioners must aspire to devise methodologies that will result in the least amount of cost distortion. Identifying with this objective will enable the practitioner to bypass the deadlock wherein an arm’s length transfer price needs to be established but the practitioner is at the same time incapable of reifying this need.

To help concretize what we have hitherto discussed, let’s look at a couple of examples wherein each of the same two different allocation bases being considered for implementation can be more or less relevant depending on the nature of the indirect costs being allocated. We will see that an imprudent choice of an allocation base can significantly distort the arm’s length charge.

Suppose that the AAA engineering firm, headquartered in Country X, has two wholly-owned subsidiaries, BBB and CCC, in Countries Y and Z, respectively. Of all the three entities, AAA is the longest-established and employs the most experienced engineers. Firms BBB and CCC, due to their different competencies, will soon collaborate on a newly-secured engagement. Both subsidiaries, at the outset of the engagement, receive broad-based technical training from AAA in order to work most efficiently. The cost to AAA for providing this technical advisory is $10,000. Both BBB and CCC benefit from the training in some capacity and the cost of the training is not directly traceable to any single subsidiary. Under these circumstances, upon what basis should the costs of training be allocated to BBB and CCC?

After careful consideration AAA has identified two potentially suitable allocation bases (i.e., the number of billable hours worked by the employees of each subsidiary and the amount of revenue earned by each subsidiary) and is now trying to judge which one of the two is a comparatively superior metric. AAA first considers the number of billable hours worked by the employees of both BBB and CCC as the allocation base because it has a readily discernible causal relationship to the allocable costs. If in the completion of the engagement BBB’s employees work 100 billable hours while CCC’s employees work 150 billable hours, this allocation base would result in the following cost allocation:

Cost Allocation to BBB Cost Allocation to CCC
= Allocation rate × Weighted allocation base = Allocation rate × Weighted allocation base
= (Cost pool ÷ Allocation base) × Weighted allocation base = (Cost pool ÷ Allocation base) × Weighted allocation base
= [$10,000 ÷ (100 + 150)] × 100 = [$10,000 ÷ (100 + 150)] × 150
= ($10,000 ÷ 250) × 100 = ($10,000 ÷ 250) × 150
= $4,000 = $6,000

At first glance the results generated by this allocation base appear perfectly reasonable; following the training, each of BBB and CCC worked 100 and 150 hours, respectively, to complete their tasks. Since the benefit of the training extends over 50 more billable hours from CCC, it seems logical that the allocated costs should be proportionately higher to CCC. However, the appropriateness of this allocation base comes into question once we consider the disparate seniority levels and billing rates of the employees of BBB and CCC.

Suppose BBB’s employees are more experienced and credentialed than CCC’s. Consequently, during the course of the engagement, BBB’s employees perform higher profile services that ultimately add more value to the recipient. BBB’s firm-wide hourly billing rate is $250 per hour. On the other hand, though CCC’s employees also perform vital specialized services, they perform services of a lower profile than BBB’s employees. CCC’s firm-wide hourly billing rate is $125 per hour. Keeping this in mind, using the revenue earned from the engagement (based on the respective number of billable hours worked and hourly billing rates) by each of BBB and CCC as the allocation base results in the following cost allocation:

Cost Allocation to BBB Cost Allocation to CCC
= Allocation rate × Weighted allocation base = Allocation rate × Weighted allocation base
= (Cost pool ÷ Allocation base) × Weighted allocation base = (Cost pool ÷ Allocation base) × Weighted allocation base
= {$10,000 ÷ [(100 × $250) + (150 × $125)]} × (100 × $250) = {$10,000 ÷ [(100 × $250) + (150 × $125)]} × (150 × $125)
= [$10,000 ÷ ($25,000 + $18,750)] × $25,000 = [$10,000 ÷ ($25,000 + $18,750)] × $18,750
= ($10,000 ÷ $43,750) × $25,000 = {$10,000 ÷ $43,750} × $18,750
= 0.22857 × $25,000 = 0.22857 × $18,750
= $5,714 = $4,286

In the example above, the revenue earned from the engagement by each of BBB and CCC seems a superior allocation base. BBB evidently benefitted more from AAA’s training because it enabled it to provide services of a higher profile and resultantly earned a greater share of the revenue from the engagement compared to CCC. This is suggestive of a stronger causal relationship to the training costs. Comparing the costs allocated to each of BBB and CCC using the two different allocation bases quantitatively highlights the consequentiality of the choice of the allocation base to the cost objects and how an imprudently chosen allocation base contributes significantly to cost distortion; there is almost a complete reversal of the financial charges to each of BBB and CCC under the two different scenarios (i.e., $4,000 to BBB and $6,000 to CCC versus $5,714 to BBB and $4,286 to CCC).

Now let’s qualitatively consider a different indirect cost with respect to which the suitability of the same two allocation bases is arguably reversed. Suppose that AAA centrally provides remote desktop IT support services to help the employees of each of AAA, BBB, and CCC resolve issues accessing the various enterprise-wide proprietary software and databases. The incidence of providing these IT services is strongly correlated with the amount of billable hours worked by the employees of AAA, BBB, and CCC. There is no evidence or reason to believe that more complex or higher caliber services performed by any given employee additionally strain the enterprise-wide IT infrastructure resulting in more system glitches requiring troubleshooting. In fact, based on service tickets, there is some evidence that BBB’s more experienced employees, compared to CCC’s relatively less experienced ones, are more adept at self-resolving many of the routine software glitches and hence less likely to tie up organizational resources.

Under these circumstances, due to the weaker causal relationship, revenue, compared to billable hours, would be the less relevant allocation base and would therefore contravene the tax law’s arm’s length standard. Contrarily, billable hours, though not perfect, due to its greater efficacy of relating the costs of the IT services to BBB and CCC, would be the superior allocation base because its usage would result in smaller cost distortion.

CLOSING THOUGHTS

Treasury Regulation Section 1.482-9 discusses the potential of alternatives among, and the comparative suitability of, different allocation bases but does not explicate the reasons as to what renders one alternative superior over another. The regulatory language, although not ambiguous, lacks conceptual fullness due to its brevity. By minimizing the depth of field, the guidance creates a sense of flatness with respect to evaluating the suitability of an allocation base. Indeed, the interpretive difficulty of the regulatory guidance is precisely attributable to this fact that it presents a field constituting only a surface. At no time does this become more apparent than when a practitioner turns to the relevant guidance while designing a cost-based transfer pricing policy to meet the tax law’s arm’s length standard.

In practice, without a sufficient epistemic understanding of what renders one allocation base suitable over another, it is unlikely that one will prudently select the best allocation base from among the alternatives. The professional subjectivity inherent in selecting an allocation base is such that the ultimate efficacy of even the best choice of allocation base is not to be found in a singular endpoint wherein all the fragments of a given cost pool neatly fit. As we discussed, the practical implication of the regulatory guidance is that it is the degree of the causal relationship between the allocable costs and the allocation base that is the most efficacious determinant of the shares of the reasonably anticipated benefits of the respective controlled group members. Discerning the causal relationships among the potentially suitable allocation bases and the allocable costs undoubtedly requires a comprehensive understanding of the underlying business activities, functions, and arrangements.

* This article is reprinted with the author’s permission from the April 2024 issue of the Journal of International Taxation, Thomson Reuters.

[1] Notably for the services cost method under Treas. Reg. § 1.482-9(b), the cost of services plus method under Treas. Reg. § 1.482-9(e), and the comparable profits method under Treas. Reg. §§ 1.482-5 and 1.482-9(f).

[2] See IRC § 482.

[3] “[No] challenge to the validity of [the] regulations issued under §482 on the grounds that they constitute an unreasonable interpretation of the statute or one plainly inconsistent with the statute has ever been successful.” Lepard, 551-2nd T.M., Section 482 Allocations: General Principles in the Code and Regulations. See, for example, Foster v. Commissioner, 80 T.C. 34 (1983); PPG Industries, Inc. v. Commissioner, 55 T.C. 928 (1970); Kahler Corp. v. Commissioner, 58 T.C. 496 (1972); and Latham Park Manor, Inc. v. Commissioner, 69 T.C. 199 (1977).

[4] Treas. Reg. § 1.482-9(k)(1).

[5] Treas. Reg. § 1.482-9(l)(3)(i).

[6] Treas. Reg. § 1.482-9(b)(7)(ii)(B).

[7] Supra note 4.

[8] Treas. Reg. § 1.482-9(k)(2)(i).

[9] Treas. Reg. § 1.482-9 more often uses the alternative cost accounting term allocation key instead of allocation base.

[10] Treas. Reg. § 1.482-9(b)(8), Ex. 16.

[11] Ibid.

[12] Ibid.

[13] Treas. Reg. § 1.482-9(b)(8), Ex. 17.

[14] Ibid.

[15] Ibid. See also Treas. Reg. § 1.482-9(b)(8), Exs. 18-19 which are also centered around the idea of relating the allocation base to the taxpayers’ reasonably anticipated benefits.

[16] Treas. Reg. § 1.482-9(k)(3), Ex. 2.

[17] See also Treas. Reg. § 1.482-9(l)(3)(ii) the implication of which is that recipients of a remote or negligible benefit from a shared intercompany function or service are to not be allocated a share of the costs of such intercompany function or service (i.e., they are not to be assigned a weighted allocation base).

[18] Supra note 8.

[19] Treas. Reg. § 1.482-9(k)(2)(ii).

[20] Treas. Reg. § 1.482-9(b)(7)(ii)(B).

[21] Ever since the promulgation of their first incarnation, the regulations under the predecessor of § 482 in the Revenue Act of 1934 have specified an arm’s length charge standard for intercompany transactions. Reg. 86, art. 45-1(b) (1935).

View the galley version of this paper in PDF format: Cost Allocation Base Selection for Transfer Pricing Purposes (Galley Version)