The court’s decision applies the regulations to the facts of the Coca-Cola controlled transactions and arrives at its conclusion in a principled manner. A combination of critical factors led to the court to side with the I.R.S., and hold that the Supply Points used intangible assets in their businesses, but were not entitled to a return to valuable intangible property at arm’s length as a result of legal ownership or exercise of Supply Point control over the intangible property. See the January edition of Insights.
The 2020 decision of the Tax Court in The Coca-Cola Company and Subsidiaries[1] (“Coca-Cola”) $3 billion transfer pricing case may cause Coca-Cola to petition the appropriate U.S. Court of Appeals for the same reason it petitioned the Tax Court. A large amount of tax is at stake over a number of open tax years. But in comparison to other transfer pricing cases, this may be one of the only reasons this controversy went to trial and may yet again proceed to trial. The Tax Court’s decision highlights several themes in transfer pricing best practice and global policy development.
Background
During the years in issue, Coca Cola products were produced by independent bottlers around the world using trademarks, product names, logos, patents, secret formulas, and proprietary manufacturing processes licensed by Coca-Cola to seven controlled companies or Supply Points. The Supply Points produced beverage concentrate for sale to bottlers, and granted bottlers the right to use the Coca-Cola trademarks, product names, and logos to package beverages. Supply Points were allocated advertising, marketing and head office costs by Coca-Cola, but did not develop or execute territorial marketing or advertising. The task of local market advertising and consumer marketing and maintaining relations with bottlers fell to approximately 60 Service Companies (“ServCos”). As advertising and marketing is an important, concerted worldwide undertaking for Coca-Cola, external ad firms and marketing firms were engaged, in some cases by the ServCos.
The issue in dispute was the pricing of the licensed rights between Coca-Cola and the Supply Points. A secondary issue was whether dividends paid to Coca-Cola by the Supply Points should be regarded as payment for the licensed rights, thereby reducing the I.R.S. income reallocation to Coca-Cola.
The value of certainty
Coca-Cola settled an I.R.S. transfer pricing examination of its 1987-1995 tax years using a residual profit split agreed with the I.R.S. The closing agreement for the examination of the 1987-1995 tax years stipulated an allocated 50% of Supply Point residual profit in excess of 10% of gross sales to Coca-Cola in satisfaction of the company’s Code §482 requirement. This arrangement came to be known as the 10-50-50 method, and was used to allocate income between Coca-Cola and the Supply Points in tax years subsequent to 1995. The closing agreement allowed the Supply Points to meet their licensing payment requirements by repatriating funds in a variety of forms, including the payment of dividends from Supply Points to Coca-Cola.
The 10-50-50 method was agreed after a lengthy I.R.S. examination and represented what appeared to Coca-Cola to be a reliable authority on which to base its transfer pricing position for future tax years. The closing agreement did not make the 10-50-50 method binding on the I.R.S. or Coca-Cola for tax years after 1995, lacking the certainty otherwise obtainable under a unilateral A.P.A. Coca-Cola nonetheless placed a high future certainty value on the I.R.S. position without contemplating the future income effect of the compliance mechanism that performs annual tests of the key assumptions underlying a typical unilateral A.P.A. A.P.A.’s and Competent Authority agreements did vary the 10-50-50 method to resolve select controversies, but the perceived wisdom of the 1995 closing agreement endured. It is unclear why variation of the 10-50-50 method by Competent Authority was not a clearer signal to Coca-Cola or a question dealt with in more depth in the Tax Court decision.
Coca-Cola argued that the I.R.S. abandoned the 10-50-50 method in the 2007-2009 tax years, but was unsuccessful in its argument as the Tax Court found there was no agreed method to abandon.
Written agreements
The inter-company agreements that could be located served as the court’s basis for understanding the intent of Coca-Cola and the Supply Points, the rights and obligations of the parties, and the terms used in the controlled transaction. First impressions were not entirely favorable, as the court characterized the agreements as terse and incomplete. court.
The court’s review of the agreements pointed out several gaps that would otherwise be absent in an agreement between independent parties, but focused on the rights and obligations of the Supply Points concerning the use and development of licensed intangible property. The court’s analysis noted the inclusion of Coca-Cola in bottler agreements when ordinary sublicensing arrangements would render this unnecessary, but most importantly traced the history of cancellation and assignment of Supply Point licensing agreements through company reorganizations. This reorganization history demonstrated that Supply Point agreements were terminated without remuneration to the exiting Supply Point.
The nil terminal value of the rights granted to the Supply Points undermined Coca-Cola’s argument that the Supply Points had accumulated the right to earn high returns from valuable intangible property by funding marketing and advertising activity. Ordinarily, an arm’s length party with a defensible claim to intangible property rights and resulting expected returns would be paid a positive amount to give up such rights. The consolidation of Supply Point operations around the world over time provided a narrative of Supply Points with limited intangible property rights at arm’s length.
The other series of company agreements that were examined by the court were those between Coca-Cola and its subsidiaries concerning the allocation of U.S. headquarters expenses or “pro-rata.” Though the court suggested that the terms of the pro-rata allocations were often unclear, together with the evidence from the licensing agreements, the allocations of advertising and marketing expenses under the pro-rata agreements were used by the I.R.S. to argue that allocations of expense to the Supply Points did not result in the right to earn high returns to licensed intangible property. This line of argument was accepted by the court, following the intangible asset ownership conditions of Treas. Reg. §1.482-4(f).
The Coca-Cola agreements largely stood on their own and were used extensively to evaluate the issue in dispute. This stands in contrast to other transfer pricing cases where fact witnesses were required to complete the court’s understanding of the interpretation by the parties of key inter-company agreements and the division of duties, risks, and rights in intangible assets that followed from that interpretation.
Impracticality of a residual profit split
Coca-Cola used the residual profit split method and other transfer pricing methods to support its position at trial, but relied principally on the 10-50-50 method or residual profit split method to determine income in its 2007-2009 tax years.
The factual underpinning of a residual profit split is critical to method selection, best method analysis, and selection of a reliable split metric when applying the method. The court’s decision describes a consolidation of Supply Point operations over time, a variability in pro-rata allocations, and some fluidity in operational managerial responsibilities between Coca-Cola and its Supply Points over time. All of these factors point at minimum to a reconsideration of the validity of the assumptions underlying the 10-50-50 method in each tax year. From the court’s decision, it seemed that the force of attraction to the method owing to the assumed precedent value of the 1987-1995 closing agreement dominated the annual requirement to select and apply the best method.
The residual profit split method is a two-sided method often employed when reliable data and uncomplicated fact patterns concerning intangible property rights are not present. The I.R.S. position adopted by the court suggests that an objective annual review of the facts and available comparable data might have resulted in a Coca-Cola income result much closer to its own. Once the assumption of Supply Point ownership of valuable intangible assets could be relaxed based on factual evidence, the way was open to apply a one-sided method such as the C.P.M.
C.U.T.s both ways
The court’s decision marks the departure from the methodological dominance of the comparable uncontrolled transaction or C.U.T. method that has served taxpayers well in past cases tried before the Tax Court.[2] The court’s thorough best method analysis and evaluation of comparability factors allowed a robust application of the C.P.M. or comparable profits method to define the I.R.S. income adjustment sustained by the court.
Coca-Cola did introduce an expert report that used an application of the C.U.T. to support the company position. The expert’s C.U.T. analysis was found by the court to rest on shaky factual foundations and to resemble a “Rube Goldberg machine.”[3] An unkind C.U.T.
Corroboration by a secondary method
For any economist that finds graphs to be persuasive visual support for ordinarily not-so-visual or exciting modes of communication, the Coca-Cola decision includes a handy rejoinder to attorney colleagues’ variously-deployed graph and economist jokes. This section is presented without a graph. Consider how much better it could be with a graph as you read the next paragraph.
To motivate the I.R.S. reason for investigating the 10-50-50 method, the expert hired by the I.R.S. considered the question of the profitability of a sample consisting of large, public manufacturers of food and beverages, Coca-Cola, and the seven Supply Points. He then produced a striking histogram, plotting the 2007-2009 return on assets in 5% bins on the Y axis at left, and the number of companies in each of the defined bins on the X axis. Coca-Cola and all but one supply point (Egypt) were found in bins considerably above the sample mean (the Y-axis reading at the point the histogram extends furthest to the right). How considerably? Between 90% and 205% in ROA terms. This served as a convincing reason for further investigation of the reason for relatively high profitability, as opposed to investigation of the similarity of royalty rates.
Graphical inspiration received, a line of inquiry about factors effecting profitability and measurement of profitability began. Ordinarily this might be an analysis performed as part of the application of a secondary method (a best method criterion under Treas. Reg. §1.482-1(c)), but with the 10-50-50 method having been agreed only as a way of settling a dispute, the C.P.M. became the primary transfer pricing method the court relied on to decide the case.
Evidentiary value of internal and other transactions
Coca-Cola differs from many multinational companies in that its final product is manufactured by bottlers around the world that are third parties. Moreover, these bottlers are in many cases public companies owing to the large capital requirements of bottling plants and distribution infrastructure. Many bottlers have raised capital on public securities markets and therefore report their financial results publicly.
While it is unclear whether Coca-Cola documented its reasons for not applying a transfer pricing method using bottlers financial data, the I.R.S. seized on the availability of bottler data and apply the C.P.M. A considerable portion of the dispute between Coca-Cola and the I.R.S. concerned comparability of the independent bottlers and the Supply Points. After a detailed presentation of arguments from each side, the court acknowledged several shortcomings of the C.P.M. using bottler data and allowed in most cases that identified differences between the controlled transaction terms and the attributes of the independent bottlers were not sufficiently significant to disqualify the C.P.M. In respect of certain differences where estimation of partial effects was not possible or reliable, the court reasoned that any resulting bias owing to a lack of comparability would serve to benefit Coca-Cola and therefore not overstate the proposed I.R.S. income adjustment.
When an intangible asset is not a valuable intangible asset
Many modern transfer pricing controversies concern the attribution of returns to valuable intangible assets as distinct from returns to more ubiquitous intangible assets used in business operations. A considerable share of the current policy development effort of the O.E.C.D. in respect of the digital economy is occupied with a similar question, and is seeking a relatively more mix-in-the-pan method of resolution than the Tax Court appeal and trial process conducted to decide the Coca-Cola issues. This article may inadvertently contribute to the debate by encouraging, of all things, a graphical approach.
Though exhaustive, the court’s decision applies the regulations to the facts of the Coca-Cola controlled transactions and arrives at its conclusion in a principled manner. A combination of critical factors led to the court to side with the I.R.S., and hold that the Supply Points used intangible assets in their businesses, but were not entitled to a return to valuable intangible property at arm’s length as a result of legal ownership or exercise of Supply Point control over the intangible property.
While many O.E.C.D. member market jurisdictions seek to argue for the equivalent of Coca-Cola’s position and the attribution of profit resulting from intangible assets in name only using profit split methods, an objective view of value capture as was argued by the I.R.S. provides an instructive view of future talks between treaty partners.
[1] Coca-Cola Co. & Subsidiaries v. Commissioner, No. 31183-15, 155 T.C. No. 10
[2] The persuasive nature of a CUT is discussed in M. Peggs, “Amazon Makes the C.U.T. – an Important Taxpayer Win, a Reminder to Consider Transactional Evidence,” Volume 4, Number 5 Insights p.45 (2017).
[3] Wikipedia explains a Rube Goldberg machine in the following language:
A Rube Goldberg machine, named after American cartoonist Rube Goldberg, is a machine intentionally designed to perform a simple task in an indirect and overly complicated way. Usually, these machines consist of a series of simple unrelated devices; the action of each triggers the initiation of the next, eventually resulting in achieving a stated goal.