Adjusting uncontrolled profit-based benchmarks for differences in operating expense structure

Transcription

1 Adjusting uncontrolled -based benchmarks for differences in operating expense structure Vladimir Starkov NERA Economic Consulting, Chicago Reprinted from the August 2008 issue of BNA International s Tax Planning International Transfer Pricing

2 Adjusting uncontrolled -based benchmarks for differences in operating expense structure Vladimir Starkov NERA Economic Consulting, Chicago When an intercompany transaction is tested using financial ratios based on operating (for example, return on sales or return on operating assets), consideration should be given to the possible differences in the composition of operating between controlled and uncontrolled transactions. This article argues that ability ratios based on operating obtained from data on uncontrolled transactions may have to be adjusted to reflect lower intensity of certain operating costs, such as, often associated with intercompany transactions. The need for such adjustment arises because operating observed in uncontrolled transactions incorporates the return on certain activities such as that may have only minor presence in intercompany transactions. The adjustment for a different structure of operating between uncontrolled and controlled transactions is indicated both when benchmarking of the controlled transaction is performed using the tested party s transactions with unrelated parties and when benchmarks are derived from a sample of unrelated s. This adjustment is particularly important when an industry is characterised by a high proportion of in the operating of the uncontrolled transactions. I. Basis for adjustments The transfer pricing regulations commonly considered a point of reference for practitioners and tax authorities, such as the U.S. regulations under Section 482 and Organisation for Economic Co-operation and Development transfer pricing guidelines, 1 stipulate that the application of -based transfer pricing methods (such as the s method or the transactional net margin method) should consider differences in the structure of operating between controlled and uncontrolled transactions. If such differences affect the reliability of the analysis, they should be adjusted for. 2 The fact that the different structure of operating between the uncontrolled and controlled transactions may significantly affect the comparability and reliability of transfer pricing analysis has been recognised in the literature. For example, David Broomhall discusses differences in selling, 2 general, and administrative (SG&A) structures between controlled and uncontrolled distributors and suggests an adjustment that would take these differences into account. 3 This article further develops the argument for adjusting ability measures of intercompany transactions by demonstrating that ignoring differences in the intensity of operating between controlled and uncontrolled transactions ( intensity is defined here as amount of operating per dollar of sales) can lead to a bias in computing the required ability of an uncontrolled transaction. The proposed adjustment method is illustrated with a numerical example. II. Profit-level indicators Financial ratios or level indicators (PLIs) based on operating (for example, return on sales, return on operating assets, and return on total costs) are widely applied in practice. The benchmark PLIs are usually derived from a sample of uncontrolled transactions, which include either the tested party s own transactions with unrelated parties or financial data of unrelated s. When PLIs based on operating are used to benchmark the ability of controlled transactions, it is important to consider whether the intensity of operating observed in the uncontrolled transactions (that is, operating per dollar of sales) is similar to that of the controlled transaction in question. In many cases, the intensity of certain operating such as sales,, pricing, invoicing, collection, and customer support may be significantly higher for unrelated-party transactions than for intercompany transactions. The bias that can be introduced by equating the rate of return derived from uncontrolled transactions to the return required for a controlled transaction is illustrated by Example 1 below. The example describes a case where do not give rise to a unique and valuable intangibles return that is difficult or impossible to benchmark reliably. In the case where such non-routine intangibles exist, a split analysis may be more appropriate. Example 1 A manufacturer, ABC, sells substantially identical products to both an unrelated distributor in its home country, X, and to a controlled distributor in a foreign country, Y (see Figure 1). For simplicity, assume that other factors affecting comparability of these hypothetical controlled and uncontrolled transactions (such as contractual terms,

3 Figure 1 Uncontrolled distributor Controlled manufacturer Controlled distributor Uncontrolled parties Uncontrolled parties in uncontrolled transactions Other operating in inter-company transactions General and administrative General and administrative Direct and indirect manufacturing costs Direct and indirect manufacturing costs Home country X Uncontrolled transaction value chain Inter-company value chain Foreign country Y assignment of risks, and economic conditions) are the same. 4 Further, assume that substantially all of the of ABC are dedicated to supporting transactions with the unrelated distributor, while the controlled foreign distributor is responsible for its own sales and activities. 5 ABC s other operating support both uncontrolled intercompany sales with equal intensity per dollar of sales. 6 The operating earned by the manufacturer in the uncontrolled transaction provides a return on all of the activities in the value chain including activities absent from the intercompany transactions. Accordingly, assigning to the intercompany transaction the same operating per dollar of sales (or costs) as the one observed in the unrelated transaction would lead to an incorrect result for the benchmark return on activities supporting the intercompany transaction. The higher is the proportion of the expenditures in the overall operating, the larger is the bias. The sign of this bias depends on whether the return on activities present in uncontrolled transactions but absent from intercompany transactions is higher or lower than the return on activities of the entire value chain. For example, if the return on is higher than the return on total costs of the tested party, the benchmark of the intercompany transaction that involves minimal activities would be overstated. III. Benchmarking with independent s When comparability between the uncontrolled and controlled transactions of the same entity cannot be established, or where such uncontrolled transactions do not exist, the ability of the uncontrolled transaction needs to be benchmarked using the data from independent s. The returns derived from the s sample, however, can still misstate the true arm s length ability of the controlled transaction whenever the structure of the s operating costs and those of the intercompany transaction remain dissimilar. This potential for bias in establishing an arm s length return for intercompany transactions indicates a need for an adjustment to capture the difference in the structure of operating between the uncontrolled and controlled transactions. The following adjustment would eliminate the return on activities absent from intercompany transactions: Subtract the cost of activities that are performed only in uncontrolled transactions (for example, ) from the total cost of the uncontrolled transaction for a self- or from the total cost of each of the independent s. Subtract an arm s length return on functions not performed in intercompany sales (for example, return on functions) from the operating of the uncontrolled transaction for a self- or from the reported operating of the each of the independent s. Recalculate the value of the PLI selected for the analysis using the adjusted figures for total cost and operating. Example 2 illustrates this adjustment. Example 2 Assume the same facts as those described in Example 1. Assume further that manufacturing entity ABC is selected as the tested party. 7 A sample of uncontrolled manufacturers is selected, and after appropriate adjustments to increase reliability and comparability, returns on total cost (ROTC) are calculated according to Table 1. 3

4 Assume that for the tested manufacturer, the proportion of (SM) in total operating (SG&A) supporting its uncontrolled transactions is 30 percent. Further assume that the uncontrolled manufacturers exhibit a similar proportion of SM in their SG&A. 8 To reflect the structure of operating characteristic of the intercompany transactions, the SG&A of each of the manufacturers is reduced by the representative proportion of the SM costs that is, by 30 percent (see Table 2). An appropriate return on functions is derived using the return on SG&A for independent companies performing similar functions. distributors may provide a reliable source for benchmarking returns on SG&A functions due to the general similarity between the functions performed by distributors and selling and activities performed by full-fledged manufacturers. 9 Assume that the benchmark return on SG&A derived from the sample of independent distribution s is found to be 40 percent. This return is then subtracted from the operating of each of the manufacturers to obtain their operating s net of functions (see Table 3). Adjusted ROTC ratios for the manufacturers are computed using the adjusted total costs and adjusted operating (see Table 4). In this example, the adjustment results in reducing the interquartile range of ROTC from an initial range of eight percent to 16 percent to the adjusted range of six percent to 15 percent. The interquartile range is reduced because, based on the assumptions this example, the return on Table 1 Sales COGS SG&A Total costs Company A % Company B % activities is higher than the overall return on the total cost for each of the manufacturers. IV. Conclusion ROTC Company C % Company D % Company E % Formula A B C D=B+C E=A-D F=E/D -based ability ratios derived from uncontrolled transactions may require adjustment to reflect different intensity of certain operating costs between uncontrolled and controlled transactions. One example of such costs is. Return on activities that do not support intercompany transactions may be significantly different from the return on the overall value chain of the intercompany transactions. Therefore equating returns of uncontrolled and controlled transactions without adjusting for differences in the operating cost structure between these types of transactions may lead to a bias in calculating the benchmark ability of a controlled transaction. This article proposes an adjustment that mitigates this bias and demonstrates how the adjustment would be implemented. Table 2 SG&A (30% * SG&A) SG&A net of COGS Total Cost Net of Marketing Company A Company B Company C Company D Company E Formula C G = 30% * C H=C G B I=B+H Table 3 Table 4 Return on functions (40% * SM) net of return net of return Total cost net of Adjusted ROTC Company A Company B Company C Company D Company E Formula G J = 40% * G E K=E J Company A % Company B % Company C % Company D % Company E % Formula K I L=K/I 4

5 Vladimir Starkov, Ph.D., is a senior consultant at NERA Economic Consulting in Chicago and may be contacted at: vladimir.starkov@nera.com The author thanks Harlow Higinbotham and Stuart Harshbarger of NERA for their review and valuable comments. Any errors and omissions are the sole responsibility of the author. 1 Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations, OECD (1995). 2 Regs. Sec (c)(2)(iii); OECD guidelines Para SG&A and Distributor Profits in Transfer Pricing, Tax Management Transfer Pricing Report 288, August 9, When comparability between the controlled and uncontrolled transaction of the same entity cannot be asserted, one would seek to establish comparability between the controlled transaction in question and transactions among unrelated parties after making appropriate adjustments. 5 When the foreign affiliate utilises some of the activities performed by another member of the controlled group, incurred on behalf of the foreign affiliate should be charged to that affiliate with an appropriate mark-up. 6 When intensity of operating other than is significantly different between the controlled and uncontrolled transactions, benchmarks for ability of the controlled transaction could be adjusted further to account for these differences using the same adjustment method as the one discussed in this article. 7 Testing a manufacturer s in a controlled transaction may be the most reliable application of -based transfer pricing methods when (1) a controlled manufacturer performs routine functions (that is, the controlled manufacturer does not own valuable intangibles); (2) a controlled entity to whom the manufacturer sells performs non-routine functions or owns valuable intangibles; or (3) reliable returns on distribution activity of the controlled entity are difficult to obtain (for example, due to lack of functional comparability, lack of reliable data, or lack of geographical comparability). Profitability of a controlled manufacturing entity may also be benchmarked in the context of a residual split method or for transfer pricing planning purposes. 8 In a real-world situation, the proposition that the tested party and s have similar intensity of may be tested by checking whether the tested party s ratio of SM to SG&A lies within an interquartile range of SM to SG&A ratios for uncontrolled companies. The for the s can be found using trade sources and dedicated publications such as Advertising Ratios & Budgets (Schonfeld & Assoc., Annual Reports) and Ad$pender (TNS Media Intelligence, Annual Reports). 9 companies performing significant advertising and functions may not be the best benchmarking candidates since independent advertising and companies are likely to own intangibles of significant value. Note also, as indicated in Broomhall (2007), that it may be important to draw such observations of return on SG&A from distributors having ratios of SG&A to sales to that of the tested party. 5