IRS Wields the Two-Edged Sword of Code Sec. 956

by Howard J. Levine
Published: April 15, 2002
Source: Taxation of Global Transactions

Section 956 of the Internal Revenue Code of 1986 (the "Code")(1) is a double-edged sword in that can be wielded by both the Internal Revenue (the "IRS") and the taxpayer. As discussed in greater detail below, Section 956 provides the IRS with a tool for taxing a "United States shareholder" of a "controlled foreign corporation" ("CFC") on the CFC's "non-Subpart-F" earnings, even when the CFC does not distribute those earnings, if the CFC makes an impermissible investment in United States property; conversely, a United States shareholder may deliberately cause its CFC to make such an investment in order to trigger Section 956 and thereby claim a credit for foreign taxes paid by the CFC.

Taxpayers attempting either to avoid or to invoke Section 956 face many pitfalls, including the risk of an IRS challenge on the ground that the substance of their transaction is inconsistent with its form. As illustrated by a Field Service Advice and a Legal Memorandum recently issued by the IRS, taxpayers that fail to carefully structure their transactions to minimize these risks do so at great peril.(2)

Statutory Overview

Before describing the FSA and the ILM, a brief overview of Subpart F in general, and Section 956 in particular, may be helpful.

United States persons generally are not subject to U.S. tax on income earned through foreign corporations until they sell their shares or receive dividends. Pursuant to Subpart F, however, special rules apply to "United States shareholders" of a CFC.

For this purpose, a United States shareholder is a United States person that owns (within the meaning of Section 958(a)) or is considered to own (pursuant to certain constructive ownership rules set forth in Section 958(b)), stock possessing at least 10 percent of the voting power of the foreign corporation.(3) A CFC is a foreign corporation a majority of the shares of which (measured by either vote or value) are owned (within the meaning of Section 958(a)) or considered owned (pursuant to certain constructive ownership rules set forth in Section 958(b)) by United States shareholders.(4)

Pursuant to Section 951(a), each United States shareholder who owns (within the meaning of Section 958(a)) stock of the CFC generally must include in his income each year the CFC's "Subpart F income" attributable to such shares.(5) In the case of a corporation, the United States shareholder generally is entitled to claim a foreign tax credit for the foreign taxes paid by the CFC to the extent attributable to such Subpart F income.(6)

With respect to non-Subpart F income, United States shareholders generally may continue to defer U.S. tax until they receive dividends or sell their shares.(7) Such deferral is considered appropriate in the case of CFC earnings that remain offshore, but is considered inappropriate for CFC earnings that are invested in U.S. property. Accordingly, Sections 951(a)(1)(B) and 956 generally provide that a United States shareholder's proportionate share of any increased investment by the CFC in U.S. property will be taxed in the same manner as Subpart F income. A corporate United States shareholder is also entitled to a foreign tax credit for any foreign taxes paid by the CFC to the extent attributable to such Section 956 inclusion.

The presence or absence of an increased investment in U.S. property is determined based upon the average of measurements made at the each of each quarter of the CFC's taxable year.(8) U.S. property generally includes, among other things, stock of domestic corporations and obligations of United States persons.(9) Exceptions to this general rule apply in the case of, among other things, certain ordinary trade receivables, bank deposits, and stock or obligations of domestic corporations that do not possess a specified relationship to the CFC.(10)



A domestic corporation ("Corp. A") and numerous foreign subsidiaries (collectively, the "Group") manufactured and distributed goods throughout the world. Corp. A directly owned 100% of the shares of a Singaporean holding company ("FC1"), a domestic distributor ("DC1"), numerous European marketing subsidiaries (collectively referred to as "EMS") and numerous non-European marketing subsidiaries (collectively referred to as "NEMS"). FC1 directly owned 100% of the shares of a Singaporean subsidiary that actually manufactured the product ("FC2") and of another foreign distributor ("FC3").

Goods manufactured by FC2 were in all events sold directly to Corp. A. Thereafter, the goods intended for European destinations were sold by Corp. A to DC1, by DC1 to FC3, by FC3 to EMS, and by EMS to customers. The goods bound for non-European destinations were sold by Corp. A to FC3, by FC3 to NEMS, and by NEMS to customers.

FC3, which was a Netherlands corporation, purchased and sold goods at the same price, and therefore earned no profit on the back-to-back sales. For Netherlands tax purposes, FC3 characterized the back-to-back sales as financing transactions. In order to support such characterization, FC3 needed to show net income. FC3 earned a modest profit from leasing a warehouse distribution center (pursuant to a sale-leaseback arrangement with DC1) and from the provision, on a cost-plus basis, of certain "logistical" services related to the warehouse distribution center.

In general, the back-to-back sales among the members of the Group were not accompanied by actual cash transfers. Instead, the goods were purchased on account by creating intercompany accounts payable at each stage of the distribution chain. In accordance with industry practice, the Group's formal policy was to satisfy intercompany payables within 60 days on a first-in-first-out ("FIFO") basis. Thus, for example, Corp. A had 60 days to pay for goods acquired from FC2 under the terms of this policy.

All payments by customers were collected by EMS and NEMS, since these were the entities that ultimately sold the Group's goods to customers worldwide. EMS and NEMS then transferred the amounts received directly to Corp. A.(11) For accounting purposes, such amounts were treated as if paid up the distribution chain (ending with Corp. A), and thus were credited against existing intercompany payables on a FIFO basis. Since the cash was retained by Corp. A, however, payments by EMS and NEMS to Corp. A did not reduce Corp. A's intercompany payables to FC2.

As noted above, certain ordinary trade payables are not considered U.S. property (the "ordinary trade payable exception").(12) Accordingly, FC2's extension of credit on Corp. A's purchases of goods in accordance with normal commercial practices would not itself have created a significant risk of FC2's being considered to have made an investment in U.S. property under Section 956. Payables that remain unsatisfied after an unusually long period, however, would be unlikely to qualify for the ordinary trade payable exception and would likely result in a taxable investment in U.S. property if still outstanding at the close of any quarter.

In order to avoid current inclusion under Section 956 with respect to certain "overaged" payables, Corp. A assigned several of its receivables due from FC3 to FC2. These assignments took place on February 28, August 31 and November 30 of "Year 4." Later, presumably at a time when Corp. A had run out of its own FC3 receivables, it began assigning DC1's FC3 receivables. These assignments took place on February 28 and May 31 of "Year 5" and on May 31 of "Year 6." Apparently, there was no formal corporate action on the part of DC1 to first distribute those receivables to Corp. A. In each case, the FC3 receivable assigned was the oldest then outstanding (and not previously assigned).

In accordance with Corp. A's assignment of the oldest outstanding FC3 receivables to FC2, and the Group's policy of making intercompany payments on a FIFO basis, the next amounts received by EMS and NEMS, and credited to the account of FC3, should have been applied for the benefit of FC2. Nevertheless, all of the amounts received by EMS and NEMS, and credited to the account of FC3, were credited (and paid) to Corp. A.

In April, Year 4, FC2 distributed the first assigned receivable (at that time, the only assigned receivable) to its parent, FC1. Pursuant to a pre-existing master loan agreement between FC1 and FC3, that receivable was converted to long-term debt with a 30-year term, providing for interest payable on a quarterly basis at LIBOR (the London Interbank Offered Rate).

As noted above, FC3 needed to earn net income in order to support its characterization of the back-to-back sales as financing transactions for Netherlands tax purposes. Once FC3 began accruing interest expense to FC1, however, this was no longer possible under the arrangements then in effect. To permit FC3 to earn net income, Corp. A made a year-end adjustment, reducing the price at which it had sold goods to FC3 during the year by an amount equal to the accrued interest expense for Year 4.

FC2 also transferred the remaining receivables to FC1, but as non-interest-bearing loans, rather than dividends.(13) Corp. A represented to the IRS that a dividend could not be distributed without permission from the local authorities, that FC1 had no intention of repaying those loans, and that the loans were to be forgiven once such permission was received. Thus, Corp. A treated the interest-free loans as disguised dividends, and the FSA does not appear to challenge this treatment. In each case, the disguised dividends to FC1 took place more than two months (and, in some cases, a significantly longer period) after the assignment to FC2.

The remaining receivables were also converted into long-term debt pursuant to the master loan agreement. Presumably, Corp. A made further year-end adjustments to the price at which it had sold goods to FC3 at the end of Years 5 and 6, but this is not entirely clear from the FSA.

On or about February, Year 6, Corp. A transferred a an amount to FC1 as payment against the principal amount owed by FC3 under the master loan agreement.

IRS Analysis

The FSA does not dispute the general proposition that an assignment of a third-party receivable can constitute payment, but concludes, on the facts presented, that the substance-over-form doctrine supports disregarding Corp. A's formal assignment of receivables to FC2.

"In this case, the assignment lacks economic substance because as the parent company of its wholly-owned subsidiaries, Corporation A controlled such entities and designed a series of individual accounting journal entries that had to be recorded by each subsidiary to give the appearance that Corporation A paid its overaged payables to FC2. In structuring these series of transactions, it is evident that Corporation A sought to reduce its overaged payables to FC2 to avoid Section 956, but did not intend to give any cash, or other property, to the benefit of FC2 at the time the assignments were made."

In support of this conclusion, the FSA relies on, among others, the following factors: (i) FC3's failure, when credited with amounts received by EMS and NEMS from customers, to credit amounts to FC2 in accordance with the assignments of the receivables; (ii) FC2's willingness to accept the assignments of overaged receivables due from FC3 in full satisfaction of the amounts owed by Corp. A, notwithstanding FC3's continued failure to credit amounts to FC2 in accordance with the assignments; (iii) FC1's willingness both to accept overaged assigned receivables as dividends (or disguised dividends) and to convert them into long-term debt under the master loan agreement; (iv) FC1's willingness to allow FC3 to use all of FC3's incoming cash to prepay underaged payables due to Corp. A, leaving FC3 with no ability to accumulate funds to repay the master loan; and (v) the fact that Corp. A economically bore the interest expense and repaid a portion of the principal under the master loan agreement.

Having determined that Corp. A's assignment of FC3 receivables lacked economic substance, and should therefore be disregarded for U.S. tax purposes, the FSA concludes that "FC2 continued to hold United States property for a period of time beyond that which would be considered ordinary and necessary to carry on the trade or business of both Corporation A and FC2." Thus, the FSA concludes that the ordinary trade payable exception did not apply and that FC2 had therefore made an investment in U.S. property taxable to Corp. A under Section 956.

The FSA offers an alternative theory that would also support a Section 956 inclusion by Corp. A even if the Corp. A payables were still considered to qualify for the ordinary trade payable exception through the respective dates distributed to FC1. Under the alternative theory, FC2's distributions to FC1 of FC3 receivables would be recharacterized as distributions of the receivables due from Corp. A.(14) Upon the conversion to long-term debt under the master loan agreement, however, the Corp. A receivables would no longer qualify for the ordinary trade payable exception. Each such conversion, then, would constitute an investment in U.S. property by FC1 taxable to Corp. A under Section 956.(15)

Comments on the FSA

On the facts presented in the FSA, the IRS's conclusion that the formal assignment of FC3 receivables lacked economic significance appears to be correct. Corp. A's argument that it intended a true assignment seems to be belied by, among other things, (i) FC3's failure to make a single payment to FC2 with respect to an assigned receivable; (ii) FC2's failure to enforce its rights under the assigned receivables; (iii) FC2's continued acceptance of FC3 receivables in full payment of overaged Corp. A receivables; and (iv) FC2's continued willingness to supply goods to Corp. A.

Once the assignments are disregarded for tax purposes, the IRS's conclusion that FC2 continued to hold the Corp. A receivables for a period of time sufficient to preclude application of the ordinary trade payable exception appears entirely justifiable.(16) Under this approach, the subsequent distributions and conversions of the receivables to long-term debt, however characterized, took place too late to avoid Section 956.(17)

Conceivably, the Corp. A receivables might continue to qualify under the ordinary trade payable exception through the respective dates on which the FC3 receivables purportedly were distributed to FC1, but this appears to be unlikely. Furthermore, the IRS would probably be successful, under its alternative theory, in characterizing the long-term debt as Corp. A debt and in treating FC1 has having therefore made an investment in U.S. property to which Section 956 applies.(18)

Corp. A's goal of effectively eliminating its payables to FC2 without triggering Section 956 was sufficiently aggressive that success would be doubtful under even the best of circumstances. Nevertheless, a critical starting point for a serious attempt to achieve that objective would have been the realization that each party to the transaction must actually comply with its obligations under the form chosen by the parties.

If, for example, Corp. A had itself converted the oldest FC3 receivables to long-term debt, reduced the price at which it sold goods to FC3 on a going-forward basis (providing FC3 prospectively with a source of cash flow for making interest payments), and then contributed the long-term debt to FC2, it might plausibly have been able to argue that actually transferred property to FC2 in satisfaction of its overaged payables. Certainly, such a transaction would still have been vulnerable to challenge on a number of grounds (e.g., on transfer-pricing grounds or on the basis that Corp. A was the true debtor under the long-term loan), but it would have been difficult for the IRS to simply dismiss the assignment as illusory.



A group of corporations had a U.S. parent ("Corp. A") and two U.S. subsidiaries ("DC1" and "DC2"). DC2 in turn wholly-owned a foreign subsidiary ("FC1") which in turn wholly-owned another foreign subsidiary ("FC2"). Although not expressly stated in the ILM, it appears that FC2 paid relatively high foreign taxes and FC1 paid relatively low foreign taxes.

DC2 wished to repatriate funds from FC2 to the U.S. The normal mechanism for accomplishing this objective would have for FC2 to pay a dividend to FC1, and for FC1 to then pay a dividend to DC2. If the repatriation had been effected in that manner, FC1 would have been treated as having paid a portion of the foreign taxes paid by FC2, and DC2 in turn would have been treated as having paid a portion of the foreign taxes paid (and deemed paid) by FC1. In that event, the high foreign taxes deemed paid by FC1 on the dividend from FC2 and the low foreign taxes actually paid by FC1 on its directly-earned income would have been blended for purposes of determining the foreign taxes deemed paid by DC2 upon the receipt of the dividend from FC1.(19)

In contrast, the Section 956 inclusion resulting from a direct investment by FC2 in U.S. property would in effect "hopscotch" directly to DC2, so that the high foreign taxes paid by FC2 (to the extent attributable to the amount invested in U.S. property) would be attributed directly to DC2; such high foreign taxes would not be pooled with, or diluted by, the low foreign taxes paid by FC1.

A simple loan by FC2 to DC2 thus would have avoided dilution of the foreign tax credit, but arguably would have accomplished only a temporary repatriation of funds, since DC2 subsequently would have been required to repay the loan. Accordingly, Corp. A and its subsidiaries (the "Group") devised a back-to-back loan transaction designed to create an investment by FC2 in U.S. property and to effect a more permanent repatriation of funds to the U.S.

A document dated December 22 of "Year 1" described a proposed back-to-back loan by DC2 to FC1, by FC1 to FC2, and by FC2 to DC1. The document indicated that the loans would be in effect for approximately three months but did not specify the interest rate (if any) payable on the loans. A workpaper of FC2 indicated that the purpose of the FC2 loan to DC1 was to trigger Section 956.

On December 29, Year 1, DC2 made a loan to FC1, which in turn loaned the same amount to FC2. It is not clear from the ILM whether these loans bore interest, had a specified term, or were formally documented. On or about December 29, Year 1, FC2 loaned the amount borrowed from FC1 to DC1. DC1's indebtedness was represented by a note (the "Note"), which provided for a three-month term commencing December 30, Year 1, interest at an "X%" rate, and a maturity date of March 30 of "Year 2." The Note was repaid, along with interest, at maturity.

DC2 claimed on its Year 1 return that FC2's loan to DC1 represented an investment in U.S. property resulting in the inclusion of income pursuant to Section 956. DC2 also claimed that, as a result of the inclusion, it was deemed to have paid a portion of the (high) foreign taxes paid by FC2.

In Year 1, FC2 also made a distribution to FC1, which in turn made a distribution to DC2. DC2 did not include an additional amount in income, however, since it took the position that the amount distributed had been previously taxed under Section 956.(20)

IRS Analysis

Relying on the step-transaction doctrine, the IRS recharacterized the back-to-back loan transaction as a loan by DC2 directly to DC1, and disregarded the participation of FC1 and FC2 on the ground that those CFCs were interposed for purposes of claiming a foreign tax credit.

Under the step-transaction doctrine, an analysis is made of the separate steps of a transaction to determine whether each step should be accorded independent legal significance or whether the steps should be treated as related steps in one unified transaction, and "stepped together" to produce the actual result.(21) Application of the step-transaction doctrine is highly unpredictable, in part because there are three alternative tests that might be applied: (i) the "end result test" (which considers whether the end result was intended by the parties from the outset), (ii) the "interdependence test" (which considers whether the intermediate steps would have been fruitless without completing all of the steps), and (iii) the "binding commitment test" (which considers whether, following the completion of the first step, there was a binding commitment to complete the remaining steps).(22) No comprehensive guidance is available as to which test should apply in any given situation.

The ILM offers no guidance on the appropriate test to apply, but opines that, on the facts presented, the IRS can argue that the step-transaction doctrine applies under any of the three alternative tests. With respect to the end result and interdependence tests, the ILM emphasizes that the first two loans were undertaken to provide FC2 with sufficient funds to make the loan to DC1 and that the first two loans would have been fruitless because neither FC1 nor FC2 needed to borrow funds (other than to loan them to DC1).

With respect to the binding commitment test, the ILM emphasizes that the document dated December 22, Year 1, "indicated that DC2, FC1, FC2 and DC1 planned to enter into a series of simultaneous back-to-back intercompany loans" and states that, at the time the first loan was made "the parties had already committed" to the other two loans as well. The ILM adds that the second and third loans were carried out on or about the same day as the first loan. "Therefore, there appears to be the existence of a binding commitment among the parties prior to the execution of the back-to-back loans."

Comments on the ILM

The ILM reflects the IRS's long-held hostility towards tax-structured back-to-back financing transactions.(23) The most notable feature of the IRS's view is its single-minded focus on the degree of integration between the back-to-back loans without regard to whether the intermediate financing entity earns a profit (a "spread"), and therefore has a legitimate business purpose for participating in the transaction, if there is no additional business purpose for the intermediate entity's participation.

The ILM's approach finds some support in Del Commercial Properties, Inc. v. Commissioner ("Del Commercial"),(24) a recent case in which the Court of Appeals for the District of Columbia Circuit affirmed a decision by the Tax Court that disregarded the participation of an intermediate finance entity in a back-to-back loan transaction, notwithstanding that the finance entity apparently earned a significant profit during some of the years at issue.(25) Not surprisingly, the ILM cites the Tax Court's opinion in Del Commercial.(26)

Notwithstanding the support found in Del Commercial, the ILM appears to be at odds with prior case law, which it declines to address. In Northern Indiana Public Service Co. v. Commissioner,(27) the taxpayer, a U.S. corporation, formed a Netherlands Antilles finance subsidiary in order to issue notes on the Eurobond market free of the obligation to pay U.S. withholding tax. The subsidiary then remitted the borrowed funds to the taxpayer in return for the taxpayer's note, which provided the subsidiary with a 1% spread on the financing. Relying principally upon the 1% spread as demonstrating recognizable business activity, and a profit motive, the appellate court in NIPSCO upheld the Tax Court's determination that the subsidiary's participation in the financing transaction should be respected for tax purposes.(28)

Moreover, even assuming that the step-transaction doctrine can be applied to disregard the participation of an intermediate financing entity that earns a meaningful spread, the ILM's contention that the binding-commitment test is satisfied on the facts presented strains credulity.(29) A far better argument would be that the binding commitment test should not apply to transactions between affiliates having a complete identity of economic interest.

For the reasons described above, the validity of the ILM's approach to back-to-back financing transactions is unclear.(30) In any event, the back-to-back loan arrangement implemented in the ILM was ill-conceived because it was wholly unnecessary.

Section 956(d) provides that, if a CFC guarantees an obligation of a United States person, the CFC will be treated as holding that obligation for purposes of Section 956. Accordingly, if DC2 had lent funds directly to DC1, and FC2 had guaranteed the loan, FC2 would have been treated as having made an investment in U.S. property subject to Section 956. The IRS would have been hard pressed to disregard the guarantee merely because the taxpayer wished Section 956 to apply.

Alternatively, FC2 could have simply made a loan to DC2 for a term of, say, 18 months. Presumably, FC2 would have expected to distribute the amount received upon repayment (or to forgive the loan shortly before maturity), but plans can change significantly in 18 months and FC2 would have had a strong argument that the transactions should be respected as separate.(31)


The issues raised, and the persuasiveness of the IRS's analysis, in the FSA and the ILM differ considerably. The lesson that each presents, however, is the same. Taxpayers wishing either to avoid or to access Section 956 should be forewarned that technical compliance with the applicable tax rules is insufficient; transactions will also be closely scrutinized to determine whether form and substance are consistent. Evidently, the taxpayers in the FSA and the ILM did not receive, or did not heed, the warning.


1. Except as otherwise indicated, all Section references are to the Code.

2. FSA 200127005 (July 6, 2001) (the "FSA"); ILM 200137005 (May 31, 2001) (the "ILM").

3. Sec. 951(b).

4. Sec. 957(a).

5. Sec. 951(a). The term "Subpart F income" includes, among other things, most passive income (e.g., interest, dividends, and capital gains). See Secs. 952(a)(2) and 954(c).

6. Secs. 902 and 960(a)(1).

7. This conclusion is premised on the assumption that other "anti-deferral" regimes, such as that applicable to U.S. persons owning shares of a "passive foreign investment company," do not apply. These anti-deferral regimes are not addressed herein.

8. Sec. 956(a)(1)(A).

9. Sec. 956(c)(1).

10. Sec. 956(c)(2)(A), (C), and (F).

11. Presumably, EMS and NEMS should have earned a profit for their marketing efforts, so that the amounts received from customers should exceed the amounts payable to FC3, but this is not expressly indicated in the FSA.

12. Section 956(c)(2)(C) provides an exception for: "any obligation of a United States person arising in connection with the sale or processing of property if the amount of such obligation outstanding at no time during the taxable year exceeds the amount which would be ordinary and necessary to carry on the trade or business of both the other party to the sale or processing transaction and the United States person had the sale or processing transaction been made between unrelated persons."

13. What it means to loan an account receivable is not crystal clear.

14. Apparently, even this alternative theory assumes that the assignments of FC3 receivables will be disregarded, since it would not be possible for FC2 to distribute Corp. A receivables if they had been satisfied through the assignment of FC3 receivables.

15. This analysis assumes that the distribution of Corp. A receivables to FC1 would not initially constitute Subpart F income, includible in the income of Corp. A. See Sec. 954(c)(3)(A) (certain dividends from a subsidiary formed in the same country as the CFC not classified as Subpart F income).

16. Alternatively, the IRS might have held that each putative assignment by Corp. of an FC3 receivable in satisfaction of a payable to FC2 in effect constituted a cancellation of such payable for no consideration. Presumably, such cancellation would be treated as a back-to-back dividend by FC2 to FC1, and by FC1 to Corp. A, which also would have resulted in the realization of income by Corp. A.

17. As noted above, each distribution of a receivable took place more than two months after it had been assigned to FC2. Moreover, each Corp. A receivable was outstanding for more than 60 days prior to assignment of an FC3 receivable.

18. Notably, FC2's distribution of the receivables to FC1 would provide FC1 with the earnings and profits required for an investment in U.S. property by FC1 to be taxable to Corp. A. Depending upon the relative rates at which foreign taxes were imposed on FC1 and FC2, the amount of foreign taxes deemed to have been paid by Corp. A may differ depending upon whether the taxable investment in U.S. property is considered to have been made by FC1 or by FC2.

19. See Sec. 902. This analysis assumes that the dividend paid by FC2 would not constitute Subpart F income to FC1, see Sec. 954(c)(3)(A) ("same country" exception for dividends, described above), but the foreign tax credit consequences would be the same either way.

20. See Sec. 959(b).

21. E.g., King Enterprises, Inc. v. United States, 418 F.2d 511 (Ct. Cl. 1969).

22. E.g., Weikel v. Commissioner, T.C. Memo. 1986-58.

23. See, e.g., Rev. Rul. 84-152, 1984-2 C.B. 381, obsoleted, Rev. Rul. 95-56, 1995-2 C.B. 322; Rev. Rul. 84-153, 1984-2 C.B. 383, obsoleted, Rev. Rul. 95-56, 1995-2 C.B. 322; Rev. Rul. 87-89 (situation (3)), 1987-2 C.B. 195.

24. 251 F.3d 210 (2001), aff'g 78 TCM 1183 (1999).

25. In Del Commercial, a series of back-to-back loans (and equity contributions) was structured so that funds borrowed by a Canadian parent corporation from a Canadian bank were ultimately loaned to a U.S. affiliate ("Del US") through a Netherlands finance subsidiary ("Del Netherlands"). Relying on the U.S.-Netherlands treaty, which provided for a complete exemption from withholding tax on interest, Del US did not withhold on interest payments to Del Netherlands.

Relying heavily on the step-transaction doctrine, and its determination that there was no business purpose for the loan by Del Netherlands to Del US, the court of appeals upheld the Tax Court's holding that interest paid by Del US was, in substance, paid to the group's Canadian parent, so that the U.S.-Netherlands treaty did not apply. Notably, the appellate court's statement of facts does not even reveal that Del Netherlands earned a profit; this can only be learned by reviewing the Tax Court's opinion.

See Peter A. Glicklich and Michael J. Miller, "Appeals Court Invalidates U.S. Netherlands 'Double-Dip' Financing Structure," Selected US Tax Developments (2001), vol. 49, no. 4 Canadian Tax Journal, 1076-1086.

26. The appellate court's opinion was not released until shortly after the ILM was released.

27. 115 F.3d 506 (7th Cir. 1997) (hereinafter, "NIPSCO"), aff'g, 105 T.C. 341 (1995).

28. See NIPSCO, 115 F.3d at 513-514.

29. According to the ILM (emphasis added), the "document" dated December 22, Year 1, merely "indicated that DC2, FC1, FC2 and DC1 planned to enter into a series of simultaneous back-to-back intercompany loans[.]"

30. The ILM's approach is consistent with its published rulings cited above (84-152 and 84-153), but revenue rulings merely represent the position of the IRS, and are not binding on the courts. Moreover, these rulings have been widely criticized. See, e.g., Robert T. Cole and Steven A. Musher, "Rev. Ruls. 84-152 and 84-153 and GCM 37940 Depart from U.S. Treaty Obligations" (1985), vol. 14, no 8. Tax Management International Journal 165-272. Notably, these revenue rulings purport to rely on Aiken Industries Inc. v. Commissioner, 56 T.C. 925 (1971), the seminal case on the use of back-to-back financing structures to gain treaty access, but in Aiken the Tax Court relied heavily on the absence of any spread earned by the finance subsidiary.

31. But cf. FSA 200135020 (Apr. 16, 2001) (capital contribution and equivalent dividend shortly thereafter recharacterized as a nontaxable stock dividend under Section 305; taxpayer's attempt to claim foreign tax credits on the purported dividend failed).