IRS Reports on Advance Pricing Agreements for 2020

ANNOUNCEMENT AND REPORT
CONCERNING
ADVANCE PRICING AGREEMENTS

March 23, 2021

This Announcement is issued pursuant to § 521(b) of Pub. L. 106-170, the Ticket to Work and Work Incentives Improvement Act of 1999, which requires the Secretary of the Treasury to report annually to the public concerning advance pricing agreements (APAs) and the Advance Pricing and Mutual Agreement Program (APMA Program), formerly known as the Advance Pricing Agreement Program (APA Program). The first report covered calendar years 1991 through 1999. Subsequent reports covered each calendar year 2000 through 2019 separately. This twenty-second report describes the experience, structure, and activities of the APMA Program during calendar year 2020. It does not provide guidance regarding the application of the arm's length standard.

Part I of this report includes information on the structure, composition, and operation of the APMA Program; Part II presents statistical data; and Part III includes general descriptions of various elements of the APAs executed in 2020, including types of transactions covered, transfer pricing methods used, and completion time.

John C. C. Hughes
Director, Advance Pricing and Mutual Agreement Program

Part I. The APMA Program — Structure, Composition, and Operation
[Pub. L. 106-170 § 521(b)(2)(A)]

In February 2012, the former APA Program was moved from the Office of Chief Counsel to the Office of Transfer Pricing Operations¹ within the Large Business and International Division of the IRS and combined with the U.S. Competent Authority staff responsible for transfer pricing cases, thereby forming the APMA Program.

In September 2018, APMA restructured its management and realigned its teams. As of December 31, 2020, the APMA Program comprised 64 team leaders, 21 economists, 9 managers, and 3 assistant directors. Each assistant director oversees 3 managers who lead teams comprised of both team leaders and economists. The APMA Program's main office is in Washington, DC, and it also has offices in northern California (San Francisco and San Jose), southern California (Los Angeles and Laguna Niguel), Chicago, and New York.

On August 31, 2015, new revenue procedures governing requests under the mutual agreement procedure (MAP) and APA applications were published in 2015-35 I.R.B. on pages 236 and 263, respectively. Revenue Procedure (Rev. Proc.) 2015-41 provides guidance and instructions on filing APA requests as well as guidance and information on the administration of APAs. Rev. Proc. 2015-41 updates and supersedes Rev. Proc. 2006-9, 2006-1 C.B. 278, as modified by Rev. Proc. 2008-31, 2008-1 C.B. 1133, which is also superseded. Rev. Proc. 2015-40 provides procedures and guidance on requesting assistance from the U.S. Competent Authority where the taxpayer believes that the actions of the United States or a treaty country result or will result in the taxpayer being subject to taxation not in accordance with the applicable U.S. tax treaty. Rev. Proc. 2015-40 updates and supersedes Rev. Proc. 2006-54, 2006-2 C.B. 1035.

Model APAs appear as appendices to this report. Appendix 1 is the model for APAs covered by Rev. Proc. 2006-9. Appendix 2 is the current model APA for APAs covered by Rev. Proc. 2015-41. A list of primary APMA contacts is available at https://www.irs.gov/businesses/corporations/apma-contacts.

Part II. APMA Program Statistical Data
[Pub. L. 106-170 § 521(b)(2)(C)(i-viii)]

Table 1: APA Applications Filed
§ 521(b)(2)(C)(i)

	Unilateral	Bilateral	Multilateral	Total
Filed 1991-19992				401
Filed 2000-2019	622	1,621	26	2,269
Filed in 2020	15	103	3	121
Total Filed 1991-2020				2,791

The charts above illustrate the number of complete applications filed per year and the bilateral requests received in 2020 by foreign country. As of December 31, 2020, APMA had also received 25 user fee filings that were not yet accompanied by substantially complete APA applications, in addition to the 121 complete APA applications.

Table 2: Executed³ and Pending APAs
§ 521(b)(2)(C)(ii-vi)

	Unilateral	Bilateral	Multilateral	Total
Total Executed 1991-2019	643	1,280	17	1,940
Total Executed in 2020	19	105	3	127
Total Executed 1991-2020	662	1,385	20	2,067
Total Pending as of 12/31/2020	43	384	21	448
Renewals Executed in 20204	11	64	0	75
Renewals Pending5 as of 12/31/2020	25	154	8	187

 

In 2020, the percentage of renewals executed remained fairly consistent (59 percent in 2020 versus 57 percent in 2019). The charts above illustrate the total number of APAs executed per year and the countries involved in the executed bilateral APAs.

As the top chart illustrates, the number of pending requests decreased slightly relative to December 31, 2019. As of December 31, 2020, over half of the pending bilateral APA requests involved either Japan, India, or Canada.

Table 3: APAs Revoked or Cancelled and Applications Withdrawn § 521(b)(2)(C)(vii)

	Unilateral	Bilateral	Multilateral	Total
Revoked or Cancelled in 2020	0	0	0	0
Total Revoked or Cancelled 1991-20206				11
Applications Withdrawn in 2020	2	5	0	7
Total Applications Withdrawn 1991-20207				272

Table 4: APAs Executed in 2020 by Industry
§ 521(b)(2)(C)(viii)

Industry
Manufacturing	43
Wholesale/Retail Trade	40
Services	22
Finance, Insurance, and Real Estate	11
Management	9
All Other Industries	2

Table 4a: Manufacturing APAs Executed in 2020

Manufacturing
Chemical Manufacturing	8
Transportation Equipment Manufacturing	8
Miscellaneous Manufacturing	7
Food Manufacturing	6
Computer and Electronic Product Manufacturing	5
Machinery Manufacturing	3
Plastics and Rubber Products Manufacturing	3
All Other Manufacturing	3

Part III. General Descriptions of APAs Executed in 2020
[Pub. L. 106-170 § 521(b)(2)(D) and (E)]

Nature of the Relationships
§ 521(b)(2)(D)(i)

As in prior years, more than half of the APAs executed in 2020 involved transactions between non-U.S. parents and U.S. subsidiaries.

Covered Transactions, Functions and Risks, and Tested Parties
§ 521(b)(2)(D)(ii-iii)

Although most of the transactions⁸ covered in APAs executed in 2020 involve the sale of tangible goods or the provision of services, approximately 25 percent of transactions covered in APAs executed in 2020 involve the use of intangible property, which can be among the most challenging transactions in APMA's inventory.

In the majority of APAs, the covered transactions involve numerous business functions and risks. For instance, with respect to functions, APAs involving manufactured products typically involve a controlled group that conducts research and development (R&D), engages in product design and engineering, manufactures the product, markets and distributes the product, and performs support functions such as legal, finance, and human resources. Regarding risks, the controlled group may assume a variety of risks, including market risks, R&D risks, financial risks, credit and collection risks, product liability risks, and general business risks. In the APA evaluation process, a significant amount of time and effort is devoted to understanding how the functions and risks are allocated amongst the controlled group of companies that are party to the covered transactions. For methods requiring selection of a tested party, the tested party that is chosen generally will be the least complex of the controlled taxpayers.

Consistent with prior years, a majority of tested parties⁹ in 2020 were U.S. distributors, U.S. manufacturers, or U.S. service providers.

Transfer Pricing Methods Used
§ 521(b)(2)(D)(iv)

In 2020, the most commonly used transfer pricing method (TPM) for both the sale of tangible property and the use of intangible property continued to be the comparable profits method/transactional net margin method (CPM/TNMM). The CPM/TNMM was used for 84 percent of transfers of tangible and intangible property.

For covered transfers of tangible and intangible property that used the CPM/TNMM, the operating margin (OM) is still the most common profit level indicator (PLI) used to benchmark results. It was used 69 percent of the time. Other PLIs, such as the Berry Ratio and net cost plus, made up the other 31 percent. As used here, “OM” is defined as the ratio of operating profit to sales,¹⁰ and “Berry Ratio” is defined as the ratio of gross profit to operating expenses.¹¹ Most services transactions (85 percent) also used the CPM/TNMM with the OM and operating profit to operating expense being the most common PLIs (used 57 percent of the time).¹²

Sources of Comparables, Comparables Selection Criteria, and Nature of Adjustments to Comparables or Tested Party Data
§ 521(b)(2)(D)(v-vii)

For the APAs executed in 2020 that involved CPM/TNMM with a North American tested party, the most widely used data source for comparables was Standard and Poor's Compustat/Capital IQ database. Different sources were used in other cases (e.g., where the tested party was not a U.S. or Canadian entity or where transaction-based methods were applied). The other most commonly used databases are listed in the table below.

Table 5: Sources of Comparable Data

Avention (formerly known as OneSource)	Mergent
Bloomberg	Orbis
Bureau van Dijk (BvD)	Prowess
Global Vantage	RoyaltySource
ktMINE	RoyaltyStat
LoanConnector	Worldscope

In making comparability adjustments, typical balance sheet adjustments, as identified in Treas. Reg. §§ 1.482-1(d)(2) and 1.482-5(c)(2)(iv), were made in most cases, including adjustments for differing amounts of payables, receivables, and inventory. Where appropriate, adjustments for different accounting practices were made to convert from LIFO to FIFO inventory accounting, and a small number of cases also involved the accounting reclassification of expenses, e.g., from COGS to operating expenses.

Ranges, Goals, and Adjustment Mechanisms
§ 521(b)(2)(D)(viii-ix)

Most transactions covered in APAs target an interquartile range as described in Treas. Reg. §1.482-1(e)(2)(iii)(C). Where the transaction involves a royalty payment for the use of intangible property, both specific royalty rates and ranges have been used. Where the covered transaction is the sale or license of intangible property, and the payment for such transfer would be a royalty based solely on external comparable uncontrolled transactions, a secondary or confirming method, e.g., a test of the post-royalty operating margin or cost-plus mark-up, has sometimes also been used. The testing periods of the APAs executed in 2020 were either a single year, the term of the APA only, or the term of the APA plus rollback years.

APAs executed in 2020 included several mechanisms for making adjustments to the tested party results when the results fall outside the range or do not match the point required by the APA. Examples of the mechanisms used are an adjustment bringing the tested party's results to the closer edge of the range applied to the results of a single year, an adjustment to the closer edge of the range applied to the results over the APA term, an adjustment to the specified point or royalty rate, or an adjustment to the median of the range for a single year.

Critical Assumptions
§ 521(b)(2)(D)(v)

The model APAs used by the IRS (included as Appendix 1 of this report) include standard critical assumptions that there will be no material changes to the taxpayer's business or to its tax or financial accounting practices during the APA term. A few bilateral cases have also included critical assumptions tied to the taxpayer's profitability in a certain year or over the term of the APA. Pursuant to § 7.06(3) of Rev. Proc. 2015-41, APMA will cancel an APA in the event of a failure of a critical assumption unless the parties agree to revise the APA.

Term Lengths of APAs Executed in 2020
§ 521(b)(2)(D)(x)

Table 6: Term Lengths of APAs Executed in 2020

Term Length (years)	Number of APAs
1	1
2	2
3	3
4	4
5	62
6	15
7	24
8	11
9	4
14	1
Average	6

As described in § 3.03(1) of Rev. Proc. 2015-41, taxpayers should request an APA term that would cover at least five prospective years and may also request that the APA be “rolled back” to cover one or more earlier taxable years, although the appropriate APA term is decided on a case-by-case basis. Of the APAs executed in 2020, 11 percent included rollback years. A substantial number of those APAs with terms of greater than five years were submitted as a request for a five-year term, and the additional years were agreed to between the taxpayer and the IRS (or, in the case of a bilateral APA, between the IRS and the foreign government upon the taxpayer's request) to ensure a reasonable amount of prospectivity in the APA term.

Amount of Time Taken to Complete New and Renewal APAs
§ 521(b)(2)(E)

Table 7: Months to Complete New and Renewal APAs Executed in 2020

	Unilateral		Bilateral		Unilateral & Bilateral
	Average	Median	Average	Median	Average	Median
New	36.2	35.3	50.8	43.7	48.9	43.7
Renewal	25.4	21.0	34.1	30.3	32.8	29.6
New & Renewal	29.0	24.0	40.1	36.3	38.5	32.7

The median time required to complete an APA continued to decrease in 2020 to 32.7 months (from 38.8 months in 2019 and 40.2 months in 2018).

Efforts to Ensure Compliance with APAs
§ 521(b)(2)(F)

As described in § 7.02(1) of Rev. Proc. 2015-41, taxpayers are required to file annual reports to demonstrate compliance with the terms and conditions of their APAs. The filing and review of these annual reports are critical parts of the APA process. Through annual report review, the APMA Program monitors taxpayer compliance with APAs on a contemporaneous basis. Annual report review also provides current information on the success or problems associated with the various TPMs adopted in the APA process.

Nature of Documentation Required in Annual Report
§ 521(b)(2)(D)(xi)

APAs require taxpayers to file timely and complete annual reports describing their operations and demonstrating compliance with the APA's terms and conditions. Not every annual report will include each of the items listed in the following table ¹³; they are required where the facts demonstrate a need for such documentation.

Approaches for Sharing of Currency or Other Risks
§ 521(b)(2)(D)(xii)

In appropriate cases, APAs may provide specific approaches for dealing with risks, including currency risk, such as adjustment mechanisms and/or critical assumptions.