1. Purpose
This memorandum assembles the documentary support for Ednaldo Silva’s contributions to the Section 482 transfer pricing regulations proposed in 1992, issued in temporary form in 1993, and finalized in 1994. Its principal exhibit is an Internal Revenue Service memorandum dated March 28, 1991, from Christine Halphen, Special Counsel (International), to Charles S. Triplett, Deputy Associate Chief Counsel (International), and John T. Lyons, Assistant Chief Counsel (International), with copies to Steve Lainoff, Associate Chief Counsel (International), and four other Counsel officials.
The Halphen memorandum establishes three facts from a source independent of Silva and contemporaneous with the events. First, on March 26, 1991, Silva, then an economist with the Manhattan District, presented to Counsel and to the Manhattan and Newark examination districts a proposal to determine transfer pricing adjustments from a statistical database of profit ratios (gross margin, net margin, return on assets) computed for groups of comparable companies. Second, the Manhattan District was already applying this statistical method and reporting settlements. Third, Counsel’s own conclusion was that the method merited recognition in the regulations. The proposed regulations of January 24, 1992 introduced the comparable profits method and the comparable profit interval ten months later.
The record therefore places the profit-ratio comparables method inside the IRS, under Silva’s authorship, before the regulatory drafting cycle that codified it. The sections below describe the document, reconstruct the proposal as Counsel recorded it, map its elements onto the regulatory text of 1992, 1993 and 1994, distinguish it from the 1988 White Paper’s basic arm’s length return method, and identify what further corroboration would add.
2. The document
The exhibit is a three-page internal memorandum on Internal Revenue Service letterhead, subject line “Small Inbound Pricing Cases Manhattan Meeting,” date-stamped March 28, 1991. Its provenance is the Office of the Chief Counsel (International), the office that drafted the 1992 proposed regulations.
| Attribute | Content |
|---|---|
| Author | Christine Halphen, Special Counsel (International) |
| Addressees | Charles S. Triplett, Deputy ACC(I); John T. Lyons, Assistant Chief Counsel (International) |
| Copies | Steve Lainoff, Associate Chief Counsel (International); Bob Culbertson, Assistant Chief Counsel (International); Carol Doran-Klein, Acting Assistant Chief Counsel (International); George Sellinger, Branch Chief; Ken Wood, Senior Tax Reviewer; meeting participants |
| Meeting reported | Tuesday, March 26, 1991, Manhattan |
| Attendees | Manhattan examination (Seymour Zwick, Mike Shulman); Ednaldo Silva, economist, Manhattan District; Manhattan counsel (Jill Frisch, Sharon Katz-Pearlman); Newark examination (Neil McNulty, John Kaffenberg, Ted Feindt, Lou Milano, Richard Wojkowski); Newark District Counsel (Pat Taylor and one other attorney); Graham Allan, AC(I) |
| Antecedent | Counsel memorandum of August 20, 1990 to the Manhattan District conditioning support for statistical pricing adjustments on sufficient similarity of the sample companies to the taxpayer and on an audit effort to obtain resale price data first |
| Case context | Manhattan and Newark non-CEP inbound pricing inventories, with instructions to increase inbound audit activity; pricing described as the dominant inbound international issue |
Two features give the document its evidentiary weight. It was written by Counsel, not by Silva, so it is not self-attestation. And it circulated to the senior officials of the International division, so the proposal reached the office that drafted the regulations at the level where drafting decisions were made.
3. The March 26, 1991 proposal as recorded by Counsel
The memorandum devotes its second page to Silva’s proposal and records six elements.
- Profit-ratio indicators. The database would derive gross profit margins, net profit margins, return on assets and other relevant ratios. These are the profit level indicators of the later comparable profits method: §1.482-5(b)(4) of the 1994 regulations names the rate of return on capital employed, the ratio of operating profit to sales, and the ratio of gross profit to operating expenses.
- Comparables selected by industry and size. Companies were to be grouped by Principal Business Activity code (the Form 1120 classification) and by asset size. Selection by functional class and by scale is the comparability screen that §1.482-5(c) later formalized.
- A purified sample. Foreign-controlled companies were to be excluded, on the stated premise that most are themselves controlled and would contaminate an arm’s length benchmark; U.S. consolidated filers were to be excluded because consolidation mixes non-distribution activities. This is the independence requirement and the functional-segmentation requirement of the 1992 and later regulations.
- Statistical inference from aggregation. The memorandum states the purpose of the study as comfort derived from the aggregation of data, even though no single company in the sample is exactly comparable to the taxpayer. This is the conceptual kernel of the comparable profit interval of 1992 and the arm’s length range of 1993 and 1994: a range derived from a set of inexact comparables replaces a single matched transaction.
- Disclosure to the taxpayer and the court. The participants agreed the database must be shareable, with identifying data removed, so that the sample and its assumptions could be tested. This anticipates the reliance on public financial data that the final regulations require of a comparables set.
- Expert validation. The next step agreed was to retain an outside statistician to opine on the validity of the samples.
The memorandum further records that the Manhattan District had already applied the statistical method in a number of small cases and was reporting settlements, that Counsel’s litigation strategy would be to defend the statistical determination of the adjustment rather than to retrofit a CUP or resale price analysis, and, in its closing sentence, that recognition of the method in the regulations would enhance the government’s litigation stance.
One qualification is Counsel’s, not Silva’s: the attending attorneys supported the effort provided statistical studies were used as a last resort after conventional methods. The 1992 proposed regulations preserved this ordering, and the 1993 temporary regulations removed it in favor of the best method rule.
4. Priority
The memorandum dates the method, its author, its operational use and Counsel’s recommendation for codification to March 1991, ten months before the proposed regulations of January 24, 1992 (57 Fed. Reg. 3571) and under the 1968 regulations’ method hierarchy of comparable uncontrolled price, resale price, cost plus, and “some other method” (§1.482-2(e)(1)(iii)).
Three points of priority follow.
- The method preceded the regulation. The proposal was presented, and the Manhattan District was settling cases on it, before any regulatory text recognized a profit-based method applied to a set of comparable companies. The memorandum’s closing recommendation shows the direction of causation: field practice and Silva’s proposal came first; regulatory recognition was proposed as a consequence.
- The author was identified by Counsel. The memorandum attributes the proposal to Silva by name and by position. No other economist is credited, and no other source for the method is cited.
- The audience was the drafting office. The addressees and copy recipients include the Associate Chief Counsel (International), the Deputy Associate Chief Counsel, two Assistant Chief Counsel (International), and the Branch Chief. The 1992 proposed regulations issued from that office.
Silva’s prior account, that he introduced firm-level gross and operating profit margins from Compustat at the IRS in the autumn of 1988 in place of industry averages read from Robert Morris Associates’ Annual Statement Studies, and that the comparable profits interval was first regarded in the field as a “fourth method” under the 1968 ordering, is consistent with the memorandum: by March 1991 the method had matured from a Compustat screen to a proposed SOI-based database and was in use on live cases. The memorandum does not itself document the 1988 origin; it documents the 1991 state.
5. Mapping to the 1992, 1993 and 1994 regulations
Each element Counsel recorded in March 1991 has a counterpart in the regulatory text that followed.
| Element of the 1991 proposal | 1992 proposed regulations (Jan. 24, 1992) | 1993 temporary regulations, T.D. 8470 (Jan. 21, 1993) | 1994 final regulations, T.D. 8552 (July 1, 1994) |
|---|---|---|---|
| Profit ratios of comparable companies as the measure of an arm’s length result | Comparable profits method, with the comparable profit interval as a mandatory check on the transactional methods | §1.482-5T comparable profits method, with the check removed | §1.482-5 comparable profits method |
| Gross margin, net margin, return on assets as indicators | Profit level indicators | §1.482-5T(b)(4) profit level indicators | §1.482-5(b)(4): return on capital employed; operating profit to sales; gross profit to operating expenses |
| Companies grouped by industry code and asset size | Comparability of the tested party and the comparables | §1.482-5T(c) comparability, with functional and resource criteria | §1.482-5(c) comparability; §1.482-1(d) comparability factors |
| Exclusion of foreign-controlled and consolidated filers | Uncontrolled comparables; segmented data | §1.482-1T(d), comparability of uncontrolled transactions | §1.482-1(d)(3)(ii) and §1.482-5(c)(2)(ii), consistency and segmentation of financial data |
| Inference from aggregation of inexact comparables | Comparable profit interval | §1.482-1T(e) arm’s length range, with the interquartile range | §1.482-1(e) arm’s length range; §1.482-1(e)(2)(iii)(C) interquartile range |
| Statistical study as last resort after conventional methods | Priority of transactional methods retained | §1.482-1T(c) best method rule, no priority of methods | §1.482-1(c) best method rule |
| Shareable, anonymized database testable by taxpayer and court | Reliance on data of uncontrolled taxpayers | §1.482-1T(c)(2) reliability of data and assumptions | §1.482-1(c)(2)(ii) data and assumptions; §1.482-1(e)(2)(ii) adjustment of the range by reliability |
The first five rows are direct descendants of the 1991 proposal. The best method rule in the sixth row reverses the one qualification Counsel attached in 1991; Silva’s own account attributes his participation in the best method rule and the comparable profits interval to his subsequent position at Chief Counsel, addressed in Section 7. The seventh row records the principle that a comparables set must rest on data the taxpayer and a court can examine, which the final regulations express through the reliability provisions.
Two features of the 1991 proposal did not survive into the regulations. The SOI tape as data source gave way to public company financial data, since Section 6103 confidentiality barred disclosure of return information even in aggregated form to the extent of identification. And the Principal Business Activity code as the classification key gave way to functional comparability without a prescribed industry code. Neither departure affects the method; both concern its data.
6. Distinction from the 1988 White Paper
The 1991 proposal is not a derivative of the basic arm’s length return method (BALRM) of the 1988 White Paper (Notice 88-123, 1988-2 C.B. 458), and the distinction matters to any claim of originality.
BALRM assigned a return to each factor of production of the tested party (a return on assets employed, a return on functions performed) using market rates for those factors, and summed the factor returns to an arm’s length profit. It was a factor-pricing construction, criticized at the time as formulary and as detached from observed profits of independent enterprises. The 1992 proposed regulations abandoned it.
The method Counsel recorded in March 1991 proceeds the other way. It takes the observed profit ratios of a set of independent companies in the same line of business and of comparable size as the benchmark, and derives the arm’s length result from the distribution of those ratios. It prices the enterprise’s outcome from comparable enterprises, not its inputs from factor markets. This is the structure of the comparable profits method as enacted, and of the transactional net margin method the OECD adopted in 1995.
The sequence is therefore: BALRM (1988) proposed and rejected; the comparables profit-ratio method presented by Silva and in field use (1991); the comparable profits method and comparable profit interval proposed (1992); the best method rule and arm’s length range issued (1993); the final text (1994). The Halphen memorandum is the document that fixes the 1991 link in that chain.
7. Subsequent role at Chief Counsel
The Halphen memorandum documents Silva’s position in March 1991 as an economist with the Manhattan District. His participation in the regulatory drafting itself rests on his subsequent appointment as the first Senior Economic Advisor to the IRS Office of the Chief Counsel and as the first Senior Economist in the Advance Pricing Agreement unit, positions from which he took part in drafting the 1992 to 1994 regulations, the comparable profits interval, and the best method rule.
Silva’s account of the transition is that IRS counsel Kim Palmerino, who characterized the comparable profits interval as a statistical method and insisted it rest on comparables to survive litigation, campaigned for him to join Chief Counsel and take part in the drafting after he proposed amending the regulations without knowing a revision was already underway. The Halphen memorandum corroborates the premise of that account from Counsel’s side: in March 1991 Counsel had already judged the method worth recognizing in the regulations and had identified Silva as its proponent.
The memorandum does not itself record the appointment. The appointment documents, the Chief Counsel organizational records for 1991 and 1992, and the drafting files of the proposed regulations are the sources that would complete the chain from the 1991 proposal to the 1992 text.
8. Limits of the record and further corroboration
The Halphen memorandum proves what it records and no more. It establishes the proposal, its author, its field application, its audience, and Counsel’s recommendation as of March 1991. It does not establish the 1988 Compustat origin, the appointment to Chief Counsel, or the authorship of particular regulatory sentences. The following documents, in order of probative value, would close those gaps.
| Document | What it would establish |
|---|---|
| Counsel memorandum of August 20, 1990 to the Manhattan District | The earliest Counsel acknowledgment of statistical pricing adjustments; whether Silva is named |
| Manhattan District case files and settlement memoranda, 1989–1991 | Application of profit-ratio comparables on live cases before the 1992 proposal |
| Silva’s 1991 proposal document presented on March 26, 1991 | The method in his own words and notation at the time |
| Appointment papers as Senior Economic Advisor, Office of the Chief Counsel, and as Senior Economist, APA unit | Position and date of entry into the drafting office |
| Drafting files and circulation drafts of the January 1992 proposed regulations | Direct participation in the text |
| Preamble to the 1992 proposed regulations, 57 Fed. Reg. 3571 | Treasury’s own description of the comparable profit interval and its rationale, to be compared with the 1991 proposal |
| Opinion of the outside statistician retained after the March 1991 meeting | Independent validation of the sampling design |
Two cautions on use. Internal IRS memoranda may carry disclosure restrictions; the Halphen memorandum has no classification marking visible on its face, but a Freedom of Information Act release or an authorized copy should be cited as the source when the document is used publicly. And claims of authorship should be stated at the level the documents support: the 1991 record supports authorship of the method and its proposal for codification; participation in drafting supports contribution to the text; neither supports sole authorship of the regulations, which no participant could claim.
Sources
Halphen, Christine (1991). Memorandum to Charles S. Triplett and John T. Lyons, “Small Inbound Pricing Cases Manhattan Meeting,” Internal Revenue Service, Office of the Chief Counsel (International), March 28, 1991, 3 pp.
Internal Revenue Service (1988). Notice 88-123, “A Study of Intercompany Pricing,” 1988-2 C.B. 458 (the White Paper).
U.S. Treasury (1992). Proposed regulations under Section 482, 57 Fed. Reg. 3571, January 24, 1992.
U.S. Treasury (1993). T.D. 8470, temporary regulations under Section 482, 58 Fed. Reg. 5263, January 21, 1993.
U.S. Treasury (1994). T.D. 8552, final regulations under Section 482, 59 Fed. Reg. 34971, July 1, 1994; Treas. Reg. §§1.482-1 and 1.482-5.