The origins of transfer pricing’s comparable profits method

By Ednaldo Silva, Director, RoyaltyStat, Bethesda, Md.

In memory of Kim Palmerino (1947–2019), the most intrepid promoter of my CPM creation.

The US’s 1968 transfer pricing regulations (under section 482 published in the Federal Register (33 FR 5848), April 16, 1968) specified three transfer pricing methods: a disaggregated individual price comparison (CUP) between the controlled price and the internal or external comparable prices; and two aggregated gross profits (resale price and cost-plus) methods.

The old notion that gross profit methods are “transactional” (i.e., price-wise disaggregated) is nonsense.

A key premise of the 1968 regulations was that (using transfer pricing) inter-group profit shifting occurred by controlled exporters under-reporting prices in revenue and controlled importers over-reporting purchases in COGS.

The 1968 regulations presumed that the controlled party operating expenses (XSGA) were at arm’s length, meaning that XSGA was devoid of related party accounts.

Today, this assumption is considered false. Major transfer pricing violations are booked in XSGA, including non-arm’s length management fees, outbound royalties, and misallocated intangible creating expenses, such as research and development, software development, and marketing and sales.

CUP method

A CUP (comparable uncontrolled price) violation can be expressed by the inequality:

1) (p – τ) q < 0

2) Else, the equality (p – τ) q = 0 corroborates CUP transactions.

The controlled price p and the uncontrolled price τ are scalars or row vectors.

In the absence of τ-comparables, the gross profits methods would apply.

Gross profits violations were expressed in two cases, depending on whether the controlled taxpayer was an importer or exporter of goods and services.

First, the importer can over-report-controlled purchase in COGS, and reduce gross profit compared to uncontrolled revenue. Second, the exporter can under-report-controlled revenue and reduce gross profit compared to uncontrolled COGS.

Comparable profits method

In the late 1980s, I was hired by the IRS to assist the Manhattan District audit of over 60 inbound distributors (controlled importers), which showed the incongruous existence of supposed arm’s length gross profits combined with persistent operating losses and the resulting accumulation of unwarranted net operating loss carryovers.

With my economic assistance, an audit of these operating loss inbound distributors showed that after adjusting (reducing or annulling) controlled outbound management fees, outbound royalties, or advertising expenses, the persistent annual operating losses would disappear, challenging the accumulated net carried losses.

I developed the CPM (comparable profits method), quorum pars magna fui, in response to the audit discovery that (even if the purchase was at arm’s length) the controlled importer’s annual operating expenses were not arm’s length (contradicting a premise of the 1968 regulations).

Two empirical violations of the 1968 assumption were discovered.

First, outbound payments for management fees or for royalties in the deductions of the controlled importer led to persistent domestic operating losses.

Second, advertising and promotion in the deductions of the controlled importer led to persistent domestic operating losses. The reimbursement (or not) by the foreign legal owner of the marketing intangibles for the importer’s marketing expenses was decisive in the assessment.

Persistent losses from advertising deductions prompted the famous cheese examples of the 1994 regulations section 1.482-4(f)(3)(iv) (Examples 2, 3 and 4), involving a foreign producer of cheese not paying for controlled domestic marketing expenses.

Excessive outbound payments were solved under the CPM by aggregating COGS and XSGA (i.e., by summing these two direct cost and indirect expense accounts), such that transfer pricing violations are captured if they occur above or below the gross profit accounts of the domestic controlled entity.

Performing service functions to benefit the foreign (offshore) intangible owning affiliate was more complex to solve because cumulative research and development and advertising and promotion deductions can create self-developed intangibles. In turn, intangibles can generate premium profits.

Three solutions to what the OECD now calls DEMPE expenses were proposed, and I shall refer to their auteurs.

My solution was to add back (i.e., to deny) the misallocated controlled deductions to the operating profits of the importer, providing an economic (no free lunch) theory for the disallowed deductions practiced by the tax auditors.

Kim Palmerino’s solution was to assert co-ownership, meaning that the intangibles created by cumulative deductions for advertising and promotion belonged to the domestic controlled entity.

I held this position as lead outside economic expert for IRS in GlaxoSmithKline Holdings (Americas) Inc. v. Commissioner, No. 5750-04 (TC 2006), where (like in the referred cheese examples) the key issue was that the domestic subsidiary of the foreign parent was regarded to be the economic owner of the marketing intangibles.

The IRS and OECD are converging toward a third solution, proposed by Christine Halphen, where the controlled deductions represent services provided by the domestic entity to the foreign legal owner of the intangibles. The controlled services must be compensated with cost reimbursement plus an arm’s length markup.

Halphen’s solution is like my solution, except that the arm’s length markup is added to the added-back advertising and promotion expenses of the domestic controlled entity.

Kim Palmerino, Christine Halphen, and John Dean (later Tax Court judge) were the IRS Office of Chief Counsel “field attorneys” who vetted the CPM in its infancy and brought my novel CPM to the attention of the IRS Office of Chief Counsel.

The CPM introduced several new ideas (some are praxis in economics and statistics):

First, absent CUP transactions, operating profits became the most reliable indicator of “true” taxable income because they capture transfer pricing violations in COGS or XSGA, thereby making gross profit methods obsolete.

Gross profit methods survived, but they should be deleted from the transfer pricing regulations because they lead to unreliable income tax assessments.

Second, like other empirical measures, the arm’s length operating profit indicator is subject to random statistical errors, which are captured in the concept of arm’s length range. The 1968 regulations’ practice of point estimates of gross profit indicators, without considering standard errors, were debunked by the novel CPM.

Third, and this was Christine Halphen’s insistence, the arm’s length interval (which later became range) of profit indicators must be based on comparable functions performed, assets employed, and geographic markets.

Except for intangible differences, such as differentiated versus undifferentiated goods and services, product comparability is irrelevant to gross and operating profits methods because they are aggregate estimates combining many goods and services.

Major risks assumed are incorporated in payroll and assets employed, and they are captured under the statistical ranges of the selected operating profit indicator. The effects of balance sheet assets employed can be tested under trivariate regression models (I continue to eschew the widespread use of univariate quartiles).

I introduced quartiles and Tukey’s outlier detection whiskers in transfer pricing to identify transfer pricing audit candidates, not to determine the arm’s length range of the selected profit indicator because the interquartile range is wide (impractical or unreliable).

To determine the arm’s length range of the profit indicator, I introduced statistical confidence intervals. Regulatory and enforcement degradations made quartiles of univariate profit indicators bad practice.

The original arm’s length “range” was calculated by statistical confidence intervals (which are narrow and become narrower as sample size increases). The original comparable profits method was published in the proposed 1992 regulations made mandatory and called the comparable profits interval (CPI). See Proposed transfer pricing regulations of January 30, 1992, 57 Fed. Reg. 3571 (1992), as corrected in 57 Fed. Reg. 27716 (1992). See also the New York Bar Report 738, October 22, 1992, extensive commentary to which I was a listed (my name is misspelled) recipient, accessed here.

CPM is a multiplier theory

To grasp the legalese of the 1968 regulations, I translated the three transfer pricing methods into algebra and found a multiplier formula tying them together.

Like in multiplier theory, I created one definition (accounting) equation and one stochastic equation system and obtained the reduced-form equation to estimate the price or profit indicator. 

Voilá, from a single reduced-form equation, the CUP, resale price, cost-plus, and CPM methods can be derived:

1) p = (c + z), which means that unit price = unit cost plus profit.

2) z = µ p, which is a stochastic profit margin equation.

3) p = λ c, which is the markup price (CUP) equation.

where the reduced-form slope coefficient is λ = 1 / (1 − µ) is the profit markup.

If the CUP cannot be found, equation (3) is multiplied by q, the quantity sold, and revenue becomes proportional to total cost:

4) p q = λ c q, or

5) R = λ C

The markup λ multiplier remains the same after aggregation over many products. A major advantage of my single reduced-form equation is that it shows that the 1968 regulations specified one pricing equation expressed in disaggregated (equation (3)) or aggregated (equation (5)) form. The idea of three transfer pricing methods is innocent.

The difference between gross profit and operating profit methods is the concept of cost employed in the analysis and resulting concept of profit. The reduced-form equation is the same, only the cost variable is measured, COGS excluding or including XSGA (operating expenses).

If cost = COGS, the analyst is in the unreliable gross profit paradigm of the 1968 regulations. If cost = (COGS + XSGA), the analyst is in the more reliable operating profit paradigm of the 1994 regulations.

Origins of the CPM

Pretenders claim that Notice 88-123, 1988-2 CB 458, 1988-49 (White Paper) contained the CPM.

Textual analysis shows this is false. The White Paper’s BALRM, which was rejected in the 1994 regulations, was devoid of the CPM’s novel features, such as the imbued (integral) concept of statistical intervals.

References

Apropos, my self-attribution is to the Roman poet Virgil, The Aeneid, book 2, lines 5–6 (1999): quorum pars magna fui.

My translation: “quorum which I played a great part,” referring to the “great part I played” in developing the CPM. The IRS Chief Counsel field attorneys and the IRS Manhattan District Administration supported my development of the CPM as a rival paradigm to the impaired gross profit methods of the 1968 regulations.

I translated Ovid’s expression because the translation in the Loeb Classical Library is tedious (windy). It reads: “wherein I played no small role.” Translating magna as “no small” is inaccurate. See Virgil, Aeneid, Books 1-6, Loeb Classical Library, 1999 [1935, 1916].