Mining for nuggets in the IRS APA report.
As a practitioner who has worked extensively with the APA program, I am impressed by the Report's utility and candor. It assembles pertinent information in a far more coherent and usable way than a roomful of redacted APAs would have. Along with insight into the IRS's approach on some important issues, it provides a menu of ideas for structuring -- or restructuring -- APAs.
Taxpayers with potential transfer pricing issues -- whether in the context of an annual pricing study or a potential APA -- should carefully review the report. This article is intended to start that process by identifying many substantive "nuggets" of information to help taxpayers understand the IRS's APA approaches and accordingly plan transfer pricing structures or design APA applications.
Background of the Report
The APA program was instituted by the IRS in 1991 through the issuance of Rev. Proc. 91-22, 1991-1 C.B. 526. The program was part of a multi-faceted government attempt to improve compliance and reduce extended audit controversies with respect to the application of Code section 482 to intercompany (primarily cross-border) transfer pricing of goods, services, and technology. The purpose of the APA program was to permit reasonably prompt resolution of potential transfer pricing issues in advance of the covered transactions to provide taxpayers and the IRS alike with certainty in respect of compliance with section 482. During the early years of the program, the IRS was simultaneously working to issue substantial revisions of the substantive section 482 rules; proposed regulations were issued in 1992 and 1993 and finalized in 1994 ("the Regulations"), based in part on the concepts explored in the 1988 Treasury/IRS intercompany pricing "White Paper."(4)
The APA program provided the procedural framework for resolving transfer pricing issues on a prospective basis, with the Regulations providing the substantive framework. The Regulations, however, do not provide a cookbook approach to transfer pricing. Rather, they set forth broad principles and rules to be applied on a case-by-case basis to each taxpayer's particular facts and circumstances. The IRS rejected establishment of safe harbors in the Regulations because of concerns about the inappropriate use of published measures, as well as the objections of U.S. treaty partners.
This has generated a keen interest in how the IRS applies the Regulations' precepts in real cases, i.e., on audit or in APAs. Early in the APA program, the IRS expressed an intent to provide generic guidance regarding the application of section 482 standards in particular industries or situations as soon as it could do so without violating the confidentiality of the taxpayers involved. An example (and, unfortunately, the only one) was IRS Notice 94-40, which set forth APA approaches for global trading operations. The IRS also published a model APA agreement (Notice 98-65), but it provided only procedural, not substantive, guidance.
Despite hope that additional guidance would be issued as the IRS gained experience with the growing APA program and final Regulations, this was not forthcoming. Some attributed this lack of additional guidance to the filing (beginning in 1996) of FOIA requests by Tax Analysts and the Bureau of National Affairs ("BNA"), which sought disclosure of the individual APAs. The IRS resisted these requests on the ground that APAs were tax return information and therefore protected under section 6103 -- the position espoused in Rev. Proc. 91-22 and its successor, Rev. Proc. 96-53. The result was a lawsuit filed in federal district court by BNA. The IRS's failure to issue any further substantive guidance emanating from the APA program may have reflected litigation sensitivities.
As the first decade of the APA program drew to a close, there were almost 200 completed APAs but little official information on their content. Suddenly, in reaction to an unfavorable court decision regarding the release of IRS field service advice,(5) the IRS changed its position in the pending BNA case and announced that it would release all APAs on a redacted basis.(6) An outcry from affected taxpayers ensued and two amicus curiae briefs were filed in opposition to the IRS's proposed capitulation (one by Tax Executives Institute and one by the author on behalf of several anonymous corporate taxpayers and two trade associations). Legislative efforts were also initiated, which eventually resulted in late 1999 in the enactment of section 521 of Pub. L. No. 106-170.(7) The BNA lawsuit, rendered moot by this legislation, was dismissed in January 2000.
Section 521 amended sections 6103 and 6110 of the Code to provide that APAs and related background information are protected tax return information and cannot be disclosed publicly, even on a redacted basis. At the same time, the legislation required the Secretary of the Treasury to publish a detailed annual report on the status of APAs, including both statistical and procedural information regarding the program as well as substantive information concerning the APAs and the approved methodologies. The report is subject to statutory confidentiality strictures: it may not include information that (1) would be protected from disclosure under the section 6110(c) restrictions applicable to private rulings and other written determinations or (2) could be associated with or otherwise identify, directly or indirectly, a particular taxpayer.
Requiring this report was in effect the trade-off for legislatively guaranteeing the confidentiality of individual APAs, and all affected parties have hoped that it will make meaningful strides toward quenching taxpayers' (and publishers') thirst for guidance on the practical application of section 482. A Joint Committee report on section 521 further noted that the legislation "is not intended to discourage the Treasury Department from issuing other forms of guidance, such as regulations or revenue rulings, consistent with the confidentiality provisions of the Code."(8)
The IRS invested significant effort in reviewing its full APA library to cull and characterize the prescribed information for public consumption. It issued the 53-page Report (covering the entire history of the APA program through the end of 1999) on schedule. The Report responds to all items on the statutory list without breaching taxpayers' confidentiality.
Do the abstracts of the IRS collection (1) satisfy the statutory requirements, (2) meet the goal of providing useful guidance to taxpayers about the practical application of section 482, and (3) temper publishers' assertions regarding private practitioner libraries accessible only to a few? (1) Yes; (2) Yes, within reason; and (3) probably No. While the Report clearly responds to all identified items and provides significant, if not soup-to-nuts, guidance, one can expect continued requests for more specific guidance despite the resulting intrusion into taxpayers' legitimate privacy spheres. The balance of this article reviews the Report's contents to permit readers to make their own judgments.
The Report initially provides information regarding the APA program procedures. It describes the composition of an IRS APA team, including the APA Office Team Leader, an economist, the local revenue agent and his/her manager or case manager, a District Counsel attorney, perhaps a technical branch attorney with particular relevant expertise, and, for bilateral cases, an analyst from the Tax Treaty Division (i.e., the U.S. Competent Authority office). A full description of the APA process itself is set forth and the current version of the mode] APA is attached as an appendix to the Report.
Statistically, the program has grown steadily, with 40 new APAs and 13 renewals completed in the most recent year. The proportion of completed unilateral APAs(9) has gradually declined (in 1999, to less than 40 percent), perhaps reflecting IRS's encouragement of bilateral solutions, taxpayers' increasing interest in bilateral APAs, and accelerating success in reaching such agreements. Somewhat disturbing is the increasing number of withdrawals from the program (13 in 1999 vs. 46 during the entire 1991-99 period), although the Report offers no comment on or explanation of this statistic.
The data reveal predictable, albeit significant, differences in the time needed to complete unilateral APAs (an average of 20 months in 1999) compared with bilateral APAs (35 months). More startling are the subtrends that can be extracted from the data: the time to complete either a unilateral or bilateral APA has been increasing and renewals of bilateral APAs take as long as new bilateral APAs (vs. half the time in the case of unilateral APAs).
* The above data demonstrate that bilateral APAs, though desired, are not particularly efficient undertakings; the three-year average plus the necessary pre-submission planning and preparation time puts the taxpayer well into the covered period before an agreement is reached. While this may still be faster than an audit, the current pace of business changes calls for a more expeditious process if any "advance" feature is to remain in APAs. A taxpayer considering a bilateral APA should explore the realistic prospects for reaching a prompt agreement with the pertinent country with the IRS at the pre-filing conference.
Functions and Risks
The Report notes the importance of a reasonably detailed functional analysis of the parties to the covered transactions. Identification and location of risks are relevant to which party is entitled to pertinent income (or loss) as well as to that party's appropriate rate of return. The Report reminds taxpayers that in reviewing this information, the IRS will assess the ability of a particular party to fund losses that might be expected to occur as the result of the assumption of a risk and the extent of a particular party's managerial or operational control over the business activities that directly influence the amount of income or loss.
* Taxpayers should anticipate these inquiries and investigate these factors in preparing an APA request or a pricing study. Focus on relative risks could be particularly important with profit-split methods.
Section 521(b) of the 1999 legislation specifically required the Report to include information on approaches for sharing currency or other risks. Dealing with significant cross-border currency risks and their effect on relative profits is a constant source of concern to many taxpayers.(10)
* By suggesting why, in the majority of cases, no currency adjustment has been required, the Report indirectly suggests ways to deal with currency issues. The Report identifies the following possible reasons why adjustments have not been required: the comparable companies may experience similar currency exposure [a hint for selecting or evaluating comparables]; the tested party may not bear any currency risk [presumably this can be arranged by agreement]; the currency fluctuations may not have been material; or the taxpayer may be able to pass through substantially all of its currency risk to end users [suggesting, for example, the need to study the terms of contracts with major customers, market price elasticity, and the currency exposure of primary competitors]. * The Report identifies two accepted methods of dealing with currency risk. The first involves a formula that adjusts the tested party's gross margin by a specified portion of the percentage change in the exchange rate. The second uses a "no-adjustment" band of exchange rate movements within which no adjustment is required; if the exchange rate change exceeds the specified band, an adjustment is made to the tested party's operating margin depending on the extent of the exchange rate fluctation. Not surprisingly (although it is helpful to know this), both types of adjustments have generally been two-sided, requiring positive or negative adjustments depending on the direction of the currency movement. Both methods presumably call for financial and economic analysis to determine appropriate adjustments, but the Report does not provide guidance on this issue.
More than 20 industry categories have 3 or more completed APAs. Three-quarters of the APAs are in 9 industry categories: financial institutions and products; computer hardware/software; chemicals and related products; transportation equipment; electrical equipment and components; food and beverages; consumer electronics; miscellaneous services; and metals and metal products.
* The noticeable paucity of telecommunications and "new economy" businesses suggests transfer pricing challenges in those businesses as well as the fast pace of "Internet time." The APA Office recently held several prefiling conferences on e-commerce issues, but the direction of APAs in this field is as yet unclear.(11)
The Report's statistics identify the types of "tested parties" (the related party or parties to whose results the transfer pricing methodology is applied) in the completed APAs. There are meaningful numbers of each: distributors, manufacturers, service-providers, participants in cost-sharing arrangements, licensors and licensees of intangible property, dealers in financial products, and dealers in commodities.
* Each category has roughly the same number of foreign and U.S. entities, providing important, and unexpected, evidence that the APA Office is receptive to methodologies that rely on foreign data for compliance testing. * The Report states that in only one case has the tested party been a "publisher and web site operator." The fact that this description appears in the list of types of tested parties suggests that the IRS is avoiding the issue concerning whether this business should be classified as a service or a license of intangible property.
TPMs: Transfers of Tangible and Intangible Property
The Regulations provide for a number of "specified" transfer pricing methods ("TPMs"), including transactional methods such as resale price, cost-plus, comparable uncontrolled price ("CUP"), and comparable uncontrolled transaction ("CUT"), as well as profit-based methods such as the comparable profits method ("CPM") and several profit-split methods. Clearly, the TPM most commonly used in APAs covering transfers of tangible or intangible property is CPM, accounting for almost 60 percent (or 114) of the APAs for these types of transactions.(12) There were meaningful, albeit smaller, numbers of CUPs (11) and CUTs (12), and a similar number of resale price (10) and cost-plus (10) TPMs. The Report observes that cost-plus has proven to be the easiest transactional method to apply because the taxpayer's costs are identifiable and it is easier to identify functionally comparable transactions than closely similar products (which is needed for a CUP). A dozen cases involved royalty-type TPMs and profit-split approaches accounted for 26 cases.
* Every TPM specified in the section 482 Regulations has been employed in APAs. Moreover, as discussed in the following pages, these were far from cookie-cutter applications.
Among the interesting TPM variations noted in the report, 2 of the 11 CUPs were based on published market data because commodity-type products were involved. The TPM in two other cases consisted of an "operating income point that depends on sales change and on internal management measure of profitability," which suggests a creative approach possibly involving econometrics.
Profit Splits: In the profit-split category (26 APAs), there was only one "comparable profit split" TPM (not surprising in view of the difficult comparability requirements of Treas. Reg. [sections] 1.482-6(c)(2)), but 14 residual profit-split TPMs.
* Taxpayers should not feel tied down by the two types of profit splits specified in the Regulations, i.e., the comparable and residual profit-split methods. There have been eight "other" profit-split cases. The Report identifies some of these "unspecified" types, basing allocation of profits on, for example: a weighted allocation formula with operating assets and certain operating expenses as factors; relative value of the parties' contributions; or compensation and activities similar to the Notice 94-40 profit split used for certain financial businesses. One APA (apparently involving financial products) even permitted the taxpayer's internal profit allocation method as the TPM, since its reliability was enhanced by the fact that the taxpayer used it in determining arm's-length payments such as compensation and bonuses. Also intriguing -- and indicative of the program's flexibility -- is the TPM described as "profit set to sum of a certain return on assets and a certain operating margin; this method combined with an other profit split." * The Report observes that profit-split methods are used "most often" when both sides own valuable nonroutine intangibles. The Report explains that, if all the intangibles were owned by only one side, the other side would usually be a simpler party whose functional contribution could be more easily valued. The Report does not, however, proscribe use of a profit-split method in that case.
PLIs: Within CPM, operating margin is the prevalent profit level indicator ("PLI"), accounting for half of the cases. The remaining CPM cases are divided approximately equally between those using as a PLI gross margin, return on assets ("ROA") or return on capital employed ("ROCE"), Berry ratio, and markup on costs. While these data suggest the program's flexibility, the statistics are not broken down between types of businesses (e.g., distributor vs. manufacturer) or types of transaction (e.g., tangible property vs. intangible property). Thus, it cannot be determined whether the program tends toward a standard PLI in certain categories of cases (e.g., operating margin for distributors). This is an area where additional detail or correlation might have been usefully provided without compromising taxpayer confidentiality.
* The Report commentary suggests that ROA/ ROCE may be more reliable for a manufacturer than for a distributor because the level of operating assets may have a higher correlation to profitability. The indicated rationale is that a manufacturer's operating assets -- such as property, plant and equipment -- may have a greater effect on profitability than a distributor's operating assets, since the primary value added by a distributor is often based on services provided, which may in turn be less dependent on the level of operating assets. Without here debating the economic validity of this point, this comment at least suggests a steep slope for using ROA/ROCE for a distributor. * Gross margin is considered a disfavored PLI because categorization between operating expenses and cost of goods sold (COGS) may be subject to manipulation. * The Berry ratio has been used as the PLI where services are the main source of value added and the expenses in providing those services are classified as operating expenses rather than costs of goods sold. An example given is "a low-risk distributor providing marketing and distribution services." (While the Report does not elaborate on the distinguishing characteristics of a "high-risk" distributor, presumably more extensive inventory, currency, or credit risks, for example, are involved.) * The Report suggests that the APA Team will often look for convergence between several different PLIs to confirm the reliability of the selected PLI. Moreover, the APA Team may scrutinize divergences between PLI results to find anomalies in the taxpayer's data (for instance, a largely depreciated but still valuable asset base). * The Report suggests that markup on costs is an appropriate PLI when the taxpayer's sales are to controlled parties (since the cost figure is uncontrolled), or where it is common industry practice to set prices by reference to costs (e.g., contract manufacturers). Total costs are normally used for this purpose with some exceptions such as product-specific taxes reimbursed by the purchaser.
Intangible Property: While the Report lumps tangible and intangible property together for some purposes, TPM issues peculiar to intangibles are fairly easy to spot. Most royalties used in APAs were fixed rate (7).
* There were two APAs where the royalty rate varied with operating margin and one where the rate varied with the ratio of R&D to sales. The first type of variable rate TPM suggests an attempt to develop a flexible "commensurate with income" royalty; the Report commentary calls this a "sliding scale" or "step royalty" and explains that the methodology is designed to effectively prove whether (and when) the intangible has great value by determining whether it generates a high level of profits. The second type could reflect significant variability in the taxpayer's R&D efforts from period to period or significant sensitivity of the comparable data to expenditure levels.
Of the 67 services cases, the TPM in the majority involved charging cost plus a mark-up. Most of the remaining cases used cost without a mark-up. The selection of TPM was generally based on the "integral part" test under the Regulations, which allows cost without a markup to be used where the services are not an "integral part" of the business activity of the service renderer or the service recipient.(13)
* The Report observes that a cost-plus method has been used for non-integral services "where it was otherwise determined that parties dealing at arm's length would not have charged out [just] the cost of services." This appears as a taxpayer option in the Regulations.
* Some more unconventional methods for pricing services included:
* Mark-up on costs, but limiting R&D expenses to a certain percentage of sales. * Asset-proportionate share of system-wide return on assets, but limited to a certain cost-mark-up range. * For contract R&D services, a mark-up on cost plus a percentage of sales of patented products resulting from the R&D and other factors.(14) * For real estate management, a fee set as a percentage of rents plus a percentage of the total value of new leases, but not less than a certain mark-up on costs. * Dollar cap on management fee. * Profit split using five-factor formula. * Profit split subject to a floor on operating margin.
TPMs: Financial Products
Half of the 43 financial products TPMs were based on a multi-factor profit split of the type described in Notice 94-40. The remaining cases generally involved inter-branch allocations, using statistical comparisons of related and unrelated party transactions. Other TPMs noted were residual profit split (2), market-based commission (2), and taxpayer's internal allocation system (1). Identifying a TPM as a profit split, however, provides little guidance compared to the detail in Notice 94-40.
TPMs: Cost-Sharing Arrangements
Half of the 14 TPMs for cost-sharing arrangements (see Treas. Reg. [sections] 1.482-7) allocated costs on the basis of sales, and several others used a combination of sales and other factors (production costs or profit). Only two allocated costs on the basis of profit and one used raw material costs. This data show that the program maintains flexibility with respect to cost-sharing approaches, although the sample is small.
* Six APAs addressed the methodology for computing cost-sharing buy-in payments, a controversial and often difficult aspect of cost-sharing arrangements. Two of the APAs based the buy-in on capitalized R&D, and another two used the sum of capitalized R&D and an amount based on a residual profit-split analysis. One other buy-in calculation was based on market capitalization, and another on residual profit split with a "comparable acquisitions check." These last two cases suggest creative developments that may prove helpful in an increasingly technology-based market environment.
The Report contains a wide-ranging discussion of the role to be played by critical assumptions and the different types of critical assumptions. Critical assumptions are objective business and economic criteria, the continued existence of which are material to the TPM. They may deal with externally driven events, such as a significant variance from budgeted sales volume, or be caused by a taxpayer's actions, such as a change in business strategy or the cessation or transfer of a covered business segment.
Although critical assumptions are often viewed as conditions that may "undo" the APA (triggering discussions with the IRS and possible revision or cancellation of the APA), they may also inject flexibility for the taxpayer if changing business conditions make continued application of the specified TPM inappropriate.
* Some critical assumptions simply call for an automatic adjustment in the TPM. Taxpayers should consider whether this might be a useful mechanism for dealing with potential business trends (e.g., changes in sales growth or market share) or a technique to enable a longer APA term.
A standard critical assumption (contained in the model APA) requires the business activities, functions, risks, assets, and accounting methods and categories of the taxpayer to remain materially the same. Apart from this, the Report categorizes six types of critical assumptions. The examples illustrate the variety of circumstances presented in APAs and the versatility of the APA process to take these into account. Taxpayers should review this part of the Report carefully in designing an APA request.
Operational Critical Assumptions: These critical assumptions, which are the most frequent type, primarily involve limits on change. A list of items that have been required to remain substantially the same includes: customers, products, risks, functions, sales levels, business methods, pricing policies, the absence of catastrophic events, the presence of a cost-sharing agreement, the presence or absence of intangible assets, and continuation of specified personnel. A frequent critical assumption involved the definition or computation of costs or expenses, along with limits on the amount and manner of variation. Based on the example given (that U.S. deductions for restructuring fees not exceed a stated maximum amount), one suspects that this sort of critical assumption is generated by IRS concerns over characterization or manipulation of expenses or the difficulties of making appropriate adjustments to comparable companies. The Report suggests that baseline levels of restricted costs (e.g., a percentage of sales) were sometimes provided, with a permitted variance. An operational critical assumption may involve specified record maintenance.
* Several critical assumptions cited in the Report related to new products, e.g., that new products will, or will not, be covered. This should remind taxpayers that carefully delineating covered transactions is important and that, without adequate anticipation, business changes can dilute the effectiveness of the APA.
Legal Critical Assumptions: The Report cites a number of critical assumptions that relate to competent authority agreements, e.g., their effectiveness or rollback effect.
* One interesting example required continued satisfaction of a related competent authority agreement, which required system profits (i.e., combined U.S. and foreign profits on the subject transactions) to remain above a specified minimum level. Presumably, this was directed at avoiding commitment to a method or result when system profits are low or negative -- a problem area for the IRS.
Other critical assumptions in this category appear directed at maintaining key legal factors within certain parameters, e.g., customs law, customs duties levels, regulations, and import or export barriers. These suggest that taxpayers should identify -- and be sensitive to -- the key legal obligations affecting risks and returns.
* An example given requires a parent company to maintain existing guarantees of its subsidiary's liabilities.
Tax Critical Assumptions: Eleven tax-related critical assumptions are noted, which appear to be mostly one-of-a-kind. While details are not provided, one can imagine the purpose for most of the assumptions. For example: statute of limitations [to ensure the ability to make corrective adjustments]; tax effect of specified expenses [to ensure reporting consistency]; ability to change a specific tax election [to avoid tax "gaming"]; permanent establishment [to confirm which tax regime applies]; and entrance into a closing agreement for earlier "rollback" years [to ensure consistent treatment, since resolution of rollback years is often part of the APA "package" negotiation].
Financial Critical Assumptions: Financial critical assumptions may require that various financial ratios (profit splits, Berry ratios, operating margins or gross margins) be maintained within prescribed ranges or limits. Note that this is a separate issue from, and applies in addition to, the agreed TPM. Perhaps one reason for these types of assumptions is to reconcile different methodological approaches of the foreign counterparty in bilateral APAs. Critical assumptions may be a useful technique in situations of this sort.
* The example given -- setting ranges for the subsidiary's gross margin and for the combined operating margins of parent and subsidiary -- sounds like this type of assumption. That is to say, requiring results to fall within a resale price method range even where that is not the TPM could be a way to effectively satisfy the treaty partner's preferred TPM, perhaps on a less precise basis; limits on combined margins would be a way to avoid having the APA apply in extreme factual situations involving unusual profits or losses. The example notes an exception -- the case where the variance is owing to valid business reasons or attributable to economic conditions beyond the parent's control -- that could significantly improve flexibility or at least provide the framework for IRS or competent authority negotiations.
Other types of financial assumptions set boundaries on particular factual uncertainties -- limitations on system loss, intangible profit projections or buy-in payments [conditions relevant to cost-sharing agreements], and valid business reasons for debt [perhaps for comparability purposes].
Accounting Critical Assumptions: Rather surprisingly, accounting assumptions seldom occurred (seven instances). Some examples of such assumptions included: mark-to-market accounting, consistency of accounting computations for all related parties, accounting for currency gains and losses, and similar accounting treatment by U.S. and foreign-related parties with respect to specified items (e.g., manufacturing costs). Given the common use of financial statement amounts to determine compliance with the agreed TPM (particularly in respect of CPM), the infrequency of this type of critical assumption suggests that the program principally relies on generally accepted accounting principles (GAAP) standards to police potential abuses.
Economic Critical Assumptions: Only eight critical assumptions involved economic and financial conditions, such as assumptions regarding interest rates or market share. Perhaps this reflects taxpayer reluctance to hinge the effectiveness of an APA on hard-to-predict economic circumstances, or the two-edged-sword nature of such items.
The approaches to the selection of comparables outlined in the Report are unremarkable. The identification of some of the common databases used, including ones for foreign companies, may be useful to taxpayers considering transfer pricing analysis seriously for the first time, although the U.S. databases at least should be common knowledge to most economic consultants involved in preparing pricing studies or designing APA proposals. The Report commentary states that information from trade publications is occasionally reviewed for identifying comparables.
* The Report emphasizes that application of multiple quantitative screens to select comparables will not, by itself, suffice. Business descriptions and 10K SEC filings must also be reviewed.
The Report usefully suggests some criteria that are applied in selecting comparables. Frequently used criteria involved the level of the market served, sales levels, tests for financial distress, the maturity of the company (e.g., minimum or maximum number of years of operation), or the geographic market served. The "geographic market" criterion may involve not only geographic areas but also the percentage of government sales.
* The APA Office generally requires the comparables to have complete financial data available for a specified period of time (e.g., three years) to avoid including companies in different stages of economic development or using atypical years of a company that is subject to cyclical fluctuations.
* Tests for financial distress include: operating losses in a given number of years, an unfavorable auditor's opinion, or bankruptcy.
Taxpayers would be well served to review their proposed comparables (whether in pricing studies or proposed APAs) for these factors before the IRS does.
* The Report provides useful examples of how to screen comparables for the presence or absence of intangibles. For example, one can screen for R&D expenditures as a percentage of sales or costs or compare book and market values of a company. Implicitly, a considerable disparity between book and market values would indicate the presence of intangibles, although it is questionable whether that is a useful standard in today's market; perhaps industry baselines should also be established.
The SIC coding of companies is not always accurate, particularly in respect of companies' principal activities. The Report suggests some ways to sort this out, e.g., looking at property, plant and equipment ("PP&E") as a percentage of sales or assets to identify distributors (which tend to have a lower ratio) vs. manufacturers (which tend to have a higher ratio). The Report also notes that testing the ratio of operating expenses to sales or to total costs can help determine whether a company has a significant marketing and distribution function.
Adjustments to enhance comparability where income-statement-based PLIs are used almost always involve "asset intensity" adjustments for differences in levels of accounts receivable, inventories, and accounts payable, whether or not the differences are material.
* An appendix to the Report provides detailed asset-intensity formulas used in many APAs, and newcomers to APA/transfer pricing analyses will find this particularly useful. Important aspects are that mid-year averages are used for balance sheet items, that formulas should avoid using data that are being tested (e.g., sales), and that a short-term interest rate is used. * The APA program requires data to be compared on a FIFO inventory/COGS accounting basis, so that adjustments are needed to correct for LIFO reporting. * The Report notes the potential importance of adjustments for differences in relative levels of PP&E, although they are less frequently seen. If these differences relate to variations in the age of assets, buy/lease distinctions, or capital-labor choices -- rather than fundamentally different functions and risks -- mechanical adjustments akin to the asset-intensity adjustments above can be used (with, however, a longer-term interest rate).
Other infrequent adjustments relate to differences in other balance sheet items, operating expenses, R&D, or currency risk. The Report also acknowledges that there have been adjustments "in rare or singular cases" for items such as start-up costs, cost of capital variations, nonroutine intangibles, sales shocks, manufacturing functions, and product liability; the indicated choice of language suggests that proposals to adjust for these factors may not be particularly welcome.
Finally, the Report notes that accounting adjustments (such as reclassifications) may be made when warranted to increase the reliability of particular comparisons, although comparable company data are often not available in sufficient detail to permit this.
A major innovation in the 1994 Regulations was authorization of the use of ranges for testing compliance with section 482.
* A surprising statistic is that most "ranges" used in APAs provide for a "specific result" (144) rather than an interquartile range(15) (46). "Specific result" does not mean a specifically identified range (such as X percent to Y percent), but a specific point result. The Report explains that this single or specific result is used to avoid the possibility of manipulation to produce a result near the bottom of a specified range and that the specific point is often (but not always) the median of the set of comparables' results.
This approach is more understandable in the context of a royalty (because it reflects typical business practices), in the context of profit-split methods, or with transactional methods such as CUPs, even though it would be possible to specify ranges in those cases (and modest ranges have occasionally been allowed). The Report does not indicate to what extent the predominance of "specific results" cases refers to these situations, but it does suggest that specific results were used "many times" in situations where a set of comparables could yield a range of results (presumably CPM cases). The number of CPM cases (114) compared with the number of interquartile range cases (46) appears to bear this out. The IRS's apparent preference in the prospective context of APAs should be carefully considered by taxpayers considering entry to the process because it may be more restrictive than methods applied on audit.
Not surprising is that only a handful of cases (5) involved the comparables' full range, since the Regulations only permit this approach with close-to-perfect comparables (i.e., when adjustments have been made to eliminate the effect of all differences).(16)
In a number of cases, a ceiling (4 cases) or a floor (20 cases) was specified, e.g., where business conditions might change. Some APAs involving financial products used "statistical confidence intervals" (e.g., 95 percent) for comparability purposes, where large sets of transactions were being compared.
Adjustments to Results
In making adjustments to bring results within a specified range, the adjustment was "sometimes" to the closest edge of the range, and "sometimes" to another point such as the median. Unfortunately, the Report provides no relative quantification between these two significantly different adjustment approaches. Adjustment mechanisms are not always automatic, and they may require negotiations between the Competent Authorities or even result in the cancellation or revocation of the APA.
* One interesting feature in "some" cases permitted automatic adjustments unless the result was far outside the specified range. The Report explains that this was designed to allow flexibility and efficiency but guard against abuse of the adjustment mechanism.
The statistics on the length of APAs form a Bell curve, with 89 percent of APA terms being three to six years, with a peak at five years. Outliers were two one-year APAs and two ten-year APAs.
The Report notes that these statistics do not include "rollback" years (i.e., pre-APA years to which the same TPM is applied to resolve similar transfer pricing issues), due to some data inadequacies, although section 521 requires this information. To fill the gap, the Report includes data obtained in an unrelated 1999 IRS survey, which indicated that approximately one-fourth of APAs included rollbacks, generally covering three to four years. At the extreme was a 16-year rollback.
* Formal rollbacks may not be critical or necessary to address taxpayer problems: In 11 cases, "the APA process facilitated a settlement of back years, though the methodology was not rolled back."
The Report and attached model APA list typical annual report documentation requirements: GAAP-based audited financial statements; statements of material differences in business operations, risks, etc., and accounting methods and classifications; failure of critical assumptions; compensating adjustments; and financial analysis demonstrating compliance with the TPM and reconciliation thereof to the financial statements. Taxpayers outside the APA program would be well advised to make and maintain such reconciliations.
Documentation may be tailored to the nature of the records kept by specific industries; as an example, the Report lists types of records peculiar to financial products businesses. Cost-sharing arrangements may generate additional documentation requirements, such as product and intangible summaries, cost reconciliations, and internal documents used in calculating the annual cost-sharing payment. Some case-specific requirements are also noted. For instance, information regarding the worldwide ratio of R&D expenses to sales may shed light on the R&D functions being performed by a domestic subsidiary as compared to those of a foreign parent.
The Report notes that some annual reports have required information on third-party transactions (e.g., to support a CUT analysis) and U.S. Customs filings (if there is an issue regarding the inconsistent valuation of imported tangible property). Business plans, or reconciliations between financial projections and actual results, may be required to ascertain whether the financial projections underlying the TPM approximated the actual financial results. It is not clear whether these types of requirements are related to a TPM feature or critical assumption or are instead included for informational purposes in anticipation of the renewal of the APA.
* In some of the covered APAs, taxpayers were required to explain extraordinary transactions with a foreign parent that exceeded a certain dollar limit.
Efforts to Ensure Compliance with APAs
The Report helpfully explains the process followed by the APA Office after taxpayers file the annual reports required after an APA has been completed. First, the annual reports are reviewed by an APA Office professional staffer who is assigned primary responsibility for reviewing most annual reports received and maintaining a compliance database. If the level of complexity makes it more efficient for a person already familiar with the case to review the report, the initial team leader may be assigned responsibility for the review; or, if a renewal is in process, the team leader handling the renewal may also review the annual reports. The APA Office review consists primarily of ensuring that required information is included and testing compliance, along with classification issues, but generally does not audit the accuracy of the numbers contained in the report.
If compliance questions arise, the APA Office, in coordination with the IRS district, will contact the taxpayer for further discussions. Absent compliance questions on the face of the annual report, it is forwarded to the IRS district, which is responsible for deciding whether or to what extent to audit the underlying data.
Adjustments have been proposed with respect to 15 APAs, out of 239 annual reports reviewed as of December 31, 1999.
There is a wealth of information in this mother lode. The inclusion of pertinent reasoning and the manifest variety of approaches should help taxpayers plan for APAs as well as negotiate APAs already in progress.
Although taxpayers or the financial press may yearn for additional specific information -- such as exact CPM percentage ranges for distributors of particular products -- or for more correlation between various data, this was not required by the legislation, would most likely not have been obtainable from properly redacted APAs, and has the potential to be seriously misleading if presented out of context. There are no one-size-fits-all solutions under the current U.S. transfer pricing regime and the Report is not intended to obviate the need for the careful comprehensive APA process itself.
Because the Report covers an undifferentiated nine-year period, it is hard to place the more intriguing nuggets in chronological context. Thus, it is difficult to tell whether the growth of the program and concomitant efforts at consistency are leading to more or less flexibility and creativity. The 2000 APA Report, covering only one year, should be a valuable barometer of trends, if the same level of detail is maintained.
Although the IRS should still be encouraged to provide more detailed information on specific industries and issues where possible (such as Notice 94-40 for global trading activities), the Report has turned the APA program into a much more public library. The Report should be commended as an industrious and useful response for additional guidance on an inherently factual and taxpayer-specific topic.
Patricia Gimbel Lewis
(1) With thanks for contributory insights from my colleagues Christopher S. Rizek and Matthew W. Frank.
(2) Annual Report Concerning Advance Pricing Agreements, IRS Announcement 2000-35 (March 30, 2000) (hereinafter cited as the "Report").
(3) Lee A. Sheppard, The IRS Reports on Its APA Program, 87 TAX NOTES, 184 (April 10, 2000). This piece is less a critique of the Report than a presentation of the arguments for disclosure of individual APAs (see the background discussion below) and displays a lack of hands-on experience with the APA process, e.g., asserting that the IRS has "given up enforcement of the arm's length standard" and replaced it with "private law."
(4) IRS Notice 1988-123, 1988-2 C.B. 458.
(5) Tax Analysts v. IRS, 117 F.3d 607 (D.C. Cir. 1997).
(6) See IRS Information Release 1999-05 (January 1999).
(7) The Ticket to Work and Work Incentives Improvement Act of 1999.
(8) Staff of the Joint Committee on Taxation, SUMMARY OF CONFERENCE AGREEMENT ON H.R. 1180 RELATING TO EXPIRING TAX PROVISIONS AND OTHER REVENUE PROVISIONS (JCX-85-99) (Nov. 17, 1999).
(9) "Unilateral" APAs are agreements between the taxpayer and the IRS, whereas "bilateral" APAs also involve an agreement between the U.S. and foreign tax authorities.
(10) See Private Letter Ruling 9237008 (May 20, 1992), a technical advice memorandum discussing possible ways to evaluate adjustments for currency risk under section 482.
(11) Interview with APA Program Director Richard Barrett, 8 TAX MANAGEMENT TRANSFER PRICING REP. (No. 13) 588-89 (Oct. 27, 1999).
(12) Note: Although the Report covers 231 APAs, the fact that some APAs cover more than one type of transaction or industry and involve more than one type of method results in various statistics aggregating more than 231. For example, the statistics on TPMs have 196 entries for tangible and intangible property, 67 for services, 43 for financial services, 14 for cost-sharing arrangements, and 6 for cost-sharing buy-in payments.
(13) Treas. Reg. [sections] 1.482-2(b)(3).
(14) In 1997 public remarks, Richard Barrett, former Director of the APA program, hinted at a "kicker" approach to provide an extra benefit for participation in the development of particularly valuable or unique intangibles. 6 TAX MANAGEMENT TRANSFER PRICING REP. (No. 12), 345, 348 (Oct. 1, 1997).
(15) Five of the 46 cases used a "Tukey filter," a statistical technique that eliminates outliers before computing the interquartile range.
(16) Treas. Reg. [sections] 1.482-1(e)(2)(iii)(A).
PATRICIA GIMBEL LEWIS is a member of Caplin & Drysdale, a Washington, D.C. tax law firm. She has both law (JD) and business (MBA) degrees from Harvard University and an undergraduate degree from Wellesley College. Ms. Lewis served on the IRS Commissioner's Advisory Group in 1994-1996, and chaired the Taxation Section of the District of Columbia Bar in 1992-1995. She is the co-chair of this year's IRS/George Washington University International Tax Institute and a frequent author, columnist, and speaker on transfer pricing issues.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||advance pricing agreements|
|Author:||Lewis, Patricia Gimbel|
|Date:||May 1, 2000|
|Previous Article:||IN MEMORIAM.|
|Next Article:||The mechanics of California's R&D tax credit.|